What is the Post-Cookie Era?
How Do Cookies Work in Browsers?
EU Regulations on User Privacy
Online data in Business – how to use them?
How to Collect Data in the Post-Cookie Era?
Privacy Sandbox
Enhanced Conversions
Consent Mode v2
Behavioral modeling
Conclusion

What is the Post-Cookie Era?

Imagine a perfect world of online data, where the user takes center stage. In this world, we can know everything about the online user: where they come from, what they do in our digital products, what their shopping preferences are, and their purchase history on our site. This knowledge is limited only by our imagination in its application. We can optimize ads, creatives, texts, service, checkout, and even manage inventory.

The online data world is available here and now, allowing us to continuously use the information it provides. However, the quality of available data is becoming blurred by regulations concerning the collection of users’ personal data on the site.

Before discussing ways to maintain high-quality online data, we should define why the cookie era is so crucial. Cookies play a key role in the online world and data quality. The online world is stateless, meaning each user interaction with a website is independent of previous ones. Cookies allow for the tracking of these interactions, enabling data collection and analysis. How, then, can we imagine digital analytics without them?

How Do Cookies Work in Browsers?

In the context of e-commerce, data quality is extremely important. With precise data, companies can better understand their customers’ behaviors, optimize marketing campaigns, and make better business decisions.

Comparing this to the automotive industry can be helpful in understanding the role of cookies in online analytics. Imagine coming home and getting into your car, only to find that all settings – mirrors, seat, steering wheel, driving mode, and even the home address in the navigation – are reset. Every time you get into the car, you would need to adjust these settings again.

In practice, if you have one car that you share with a partner, each of you has your own key associated with individual car settings. This key is analogous to a cookie in the online data world. When you use a website, add an item to your cart, log in, and then return to that site, the products remain in the cart.

If you move from a product page to the cart and then to checkout to complete your purchase, a cookie is essential in the stateless world to transfer this information between subpages. One might ask, if cookies are so useful, why are we starting to talk about a post-cookie era in which cookies will be limited?

In practice, technologically we divide cookies into first-party and third-party categories. More important is understanding how the European Union and legislation view cookies. They are seen as essential, meaning the service cannot function without them. For example, without cookies, adding a product to the cart and then completing the purchase would be impossible, as the service would not remember that the products were added.

Unfortunately, from the perspective of data professionals, most cookies are considered non-essential. Analytical and marketing cookies play a crucial role in understanding how users interact with websites and delivering personalized ads.

EU Regulations on User Privacy

The European Union takes a strict approach to privacy issues because companies like Google collect vast amounts of data about users. They track what we watch, read, buy, and even store our financial data if we leave them there. Logging into Gmail, we provide our personal information, giving Google access to a wide range of details.

The EU aims to better protect data of internet users in the future. This process began in 2002 with the e-privacy directive aimed at safeguarding data. However, it was mostly a declaration on paper. Real changes began in 2009, about 10 years after the internet revolution. Websites started displaying banners informing users about cookie usage, which introduced little change in terms of e-privacy and merely reduced site usability.

Significant changes started occurring in the last 4-5 years, starting from 2019. The EU Court of Justice’s ruling turned into guidelines from the European Data Protection Board. Websites must now seek user consent for cookie usage, not just inform them about it. Consequently, cookie management platforms began to emerge, managing and obtaining user consent.

Among these platforms, decisions often arise about what types of cookies can be used. From a statistical and marketing data perspective, all cookies are important. Moreover, GDPR requires active consent, meaning users must opt into cookie usage. There must be an easy way to withdraw consent, and dark patterns, such as hiding the “decline all” button or manipulating button colors, cannot be used.

In 2023, the Digital Markets Act introduced significant changes in the context of cookie usage, which has had a major impact on the world of data. The lack of an ID makes data less contextual. Without an ID, a visit cannot be linked between sub-pages or different time periods, making it difficult to recognize a user. Business, especially online marketing, badly needs this data. The lack of cookies leads to a problem referred to as “Garbage In, Garbage Out.”

Online data in Business – how to use them?

Collaborating with e-commerce and online businesses, we help leverage data in three main contexts. Every business should utilize data regarding:

1. User Acquisition to the Website – It’s essential to know how many times an ad was displayed, how many clicks it received, how many people visited the site, their actions on the site, and whether they made a purchase. This allows linking marketing spend to revenue, visible in analytics tools.
2. User Behavior on the Site – Analyzing how users navigate the site, what they click, and what subpages they visit gives insight into their preferences and needs.
3. Conversions and Sales Results – Monitoring whether a user made a purchase allows evaluating the effectiveness of marketing campaigns and optimizing expenditures.

In practice, the main source of truth is the data warehouse that many companies manage in the context of web analytics. One of the key tools is Google Analytics, which, properly configured, should show about 85% of market reality. This 15% is due to the fact that some people have blocked cookies, especially tech-conscious users. In technology sites, the differences can be much greater.

In e-commerce, especially in fashion, these differences should not exceed 10%, which means that 90% of the data should be available. However, with the changes that began around 2019, we are losing more and more data in online marketing. These losses can be as high as 50%.

How to Collect Data in the Post-Cookie Era?

So what are the solutions? How to act in a post-cookie world to preserve data quality and effectively support marketing and business?

Privacy Sandbox

The first solution preparing the market, specifically Google, is the so-called Privacy Sandbox. This is a bunch of various solutions, including the Protected Audience API. Instead of third-party cookies that transfer user activity data between sites (e.g., someone saw shoes on Modivo, then sees the same shoes in an ad on Onet), the browser will now decide on displaying the ad based on the user’s browsing history. The quality of matching may be lower as competitors like CCC, Answear, or other fashion shops might also appear alongside Modivo.

Currently, all browsers except Chrome block third-party cookies, preventing remarketing on these browsers.

Enhanced Conversions

In the Post-Cookie Era, we must prepare for these changes and hope that solutions will be refined enough that our marketing efforts do not suffer significantly. One solution to implement internally within companies is Enhanced Conversions from Google and Conversion API from Meta.

Enhanced Conversions work by having Google Marketing Platform recognize that user X saw a particular ad if they were logged in to Gmail while viewing it. When cookies are absent, implementing Enhanced Conversions involves sending a hashed email along with the transaction. During the transaction, an email must be provided for confirmation and status purposes. Google, upon receiving the hashed email, can match data and improve conversion attribution accuracy.

The improvement in conversion attribution quality with Enhanced Conversions averages 17% for video ads and 5% for Google Ads. This means that the actual revenue from a transaction is properly attributed, rather than erroneously, to another traffic source. Enhanced Conversions thus rely on sending hashed personal data, which significantly improves the accuracy of ad performance analysis.

Consent Mode v2

Another solution for data gaps is Consent Mode. Consent Mode fills in data based on information without user IDs. Implementing this solution requires the initial element—a cookie management platform—to know what consent the user has given.

The way it works is that if the user agrees to cookies (state granted), everything works as in the cookie world and there is no problem. In the cookie versus post-cookie era, nothing changes here. In the post-cookie era, when the user does not give consent, every interaction sends so-called pings. A ping is a packet of information that does not contain the user ID. There is also a third state – lack of consent. This is a key change that has significantly affected online marketing. Publishers cannot process data if the user has not made any decision regarding the use of cookies.

Behavioral modeling

Digital Market Act (DMA) introduced new regulations, with Google noting the GDPR directives of 2020 requiring companies to effectively manage cookie usage. Top companies like Google, Meta, and Amazon face penalties of up to 10% of revenue for processing data without user consent.

Many companies that did not implement these changes saw a drop in traffic and sales. This was because marketing mechanisms such as Google Ads, which automatically optimize campaigns based on machine learning, need data to work. When the flow of data was interrupted, ads stopped displaying, resulting in a drop in traffic and revenue.

One solution to this problem is behavioral modeling, using so-called pings. Pings, packets of information without user IDs, are sent even when consent is not given, thanks to consent mode. Models created in Excel or Matlab in the past were fairly effective, but today, with machine learning, data is even more precise. In the Post-Cookie era, despite the lack of cookies, we can achieve greater data coverage in data warehouses, up to 85%. In data warehouses such as BigQuery, data is constantly updated.

With consent mode enabled, this data is available without a user ID, but can still be remodeled and used, for example by combining it with data from media and analytics platforms. We often act as an assistant, creating data warehouses for marketing or helping to manage them as the main source of truth. We supplement these warehouses by improving the quality of the data.Conclusion

The online data world is becoming increasingly complex. Mechanisms exist to sharpen the data picture, though they will never be as accurate as in the cookie era. It’s worth preparing for the worst by implementing at least consent mode version two and hoping for the best solutions the advertising industry can offer in the future.

The post Marketing in Post-Cookie Era first appeared on Conversion.

An analytics strategy is a crucial element in working with online data. In this article, I will describe how we approach digital analytics strategy in our work with clients.

The primary goal of digital analytics
What is an analytics strategy?
Key elements of the strategy
Analytics services
Summary

The primary goal of digital analytics

The primary goal of digital analytics is to support organizations in making decisions using data. In the case of digital analytics, these are mainly online data. Companies that base their decisions on data grow faster, achieve their KPIs more easily, and obtain higher returns on their online marketing investments.

The outcomes that companies achieve largely depend on their level of analytical maturity. Companies at a higher level of analytical maturity can better understand and utilize data, which translates into better business results.

I have written about data-driven organizations and levels of analytical maturity in separate articles where I thoroughly explore this concept. In today’s article, I will focus on the practical aspects of the digital analytics strategy, i.e., the path to becoming a data-driven organization. The Holy Grail for any organization that wants to operate based on data is to become a data-driven organization.

What is an analytics strategy?

An analytics strategy, in itself, is a broad concept and is interpreted differently by various companies.

In the context of digital analytics, the strategy is a long-term path leading the organization to become data-driven. In practice, as part of our analytics services, we implement what we internally call roadmaps. These define how a company will achieve the status of a data-driven organization with our help. Throughout the process, different competencies may be taken over by the company or supplemented by us.

This approach comes from many years of experience in helping companies build analytical maturity. An analytics strategy can be compared to a football team’s strategy. Some teams are more focused on attacking, others on playing in the midfield, and yet others on defense and scoring goals on counterattacks. The strategy for each football match should be detailed as tactics.

Key elements of the strategy

In the context of digital analytics, the strategy consists of three elements: people, processes, and tools. People are those who execute the analytics, processes are how the analytics are performed, and tools are the data used in these processes.

To build effective tactics, we determine who, how, and what will be executed within digital analytics. We then create roadmaps that lead to achieving the highest level of analytical maturity.

This process begins with a business workshop where we map the organization. We identify who is responsible for what and how information flows between different individuals in the organization.

The next step is the analysis of the tool stack. We check which tools are already being used and whether they need an audit or can be immediately used to execute the processes.

The third element of the business workshop is the processes, which must be appropriately defined and optimized. Analytics processes within a company cover various aspects such as reporting structure, communication, ad hoc analysis, and conversion rate optimization. Mapping these processes allows us to understand the current status, namely the state of people, processes, and tools within the organization.

We then develop a strategic picture of the ideal state and create a roadmap that outlines how to reach it.

The roadmap includes steps such as implementing a single source of truth, building the organization’s know-how in fundamental analyses, and implementing a conversion rate optimization process. These significant points are broken down into specific tasks, providing a clear action plan within the scope of our analytics services.

Analytics services

Analytics services are carried out in four areas:

• The first is data provision, which includes analytics tools.
• The second is data maintenance, ensuring their quality.
• Once the data is verified to be correct, we continue to maintain them to preserve high quality of analyses. Data activation involves implementing and assisting in organizing processes using data.
• A crucial element is building know-how so that employees can effectively use these data.

Such a digital analytics roadmap contains tactical points that lead to building a strategy. We typically break it down into a 6-9 month period, describing which initiatives we will undertake, when, and in what order we will execute specific tasks and mini-projects. The roadmap also includes the expected outcomes of individual initiatives.

At the end, we summarize everything so that those managing the delivery department have a clear picture of what elements are being implemented for each client, leading to a data-driven organization.

This way, we know how to manage the analytics stack. The roadmap is continuously updated, and at least once every six months, it is reviewed and updated.

Summary

Our approach to the analytics strategy consists of two main elements. The first is a resource map involving people, processes, and tools. The second is the tactics that will drive us towards becoming a data-driven organization.

We do not create an ideal model where the people are a team of analysts and the processes are specific activities because this would require the involvement of many departments in the organization. We implement the analytics strategy with the ideal picture of a data-driven organization in mind. Thanks to our roadmaps, we develop the organization in this direction.

During the collaboration, various needs emerge, such as hiring an in-house analyst, using our services as an alternative to an online analyst, or other solutions. These issues typically arise during the collaboration.

We base the execution of these tactics on four areas: data provision, data maintenance, data activation, and know-how building. Data provision includes building this element, i.e., the data themselves and their maintenance.

Know-how building is the WHO element, meaning people, because a data-driven organization is not only analysts but also other employees who can analyze data at a basic level. The HOW element refers to the processes organized around data and those that utilize data within the organization.

If you are curious about the level of analytical maturity of your organization, there is a survey linked below that will help determine your level of analytical maturity. If you find the result unsatisfactory, feel free to contact us to discuss how building analytical maturity in your company might look.

The post Analytics strategy: how to approach digital analytics? first appeared on Conversion.

Marketing is like a football team. The success of the team is not only dependent on the performance of the striker, who scores goals, but on the performance of the entire team, from the goalkeeper to the defenders, midfielders, and forwards. At the IX SGH Forum, I had the pleasure of discussing the challenges in online marketing related to how to attribute revenue generated by our website to the traffic sources bringing users to the site.

What challenges do ecommerce businesses face?
Marketing’s perspective on revenue
How to determine channel effectiveness based on generated transactions?
Models recording user data on the website
Attribution modeling
Summary

What challenges do ecommerce businesses face?

I wanted to present the challenges related to revenue from a business perspective, particularly concerning e-commerce. Speaking as someone responsible for marketing data analysis, I’d like to shed light on the analysis of revenue and the challenges e-commerce faces from this perspective.

Most finance professionals look at revenue through the basic lens of an income statement. From this point of view, examining the structure of a typical online store reveals several positions that will be considered from a financial reporting perspective, focusing on the income and expenses disclosed in a profit and loss statement.

This analysis often follows a more product-centric approach, meaning net sales revenue is broken down into product categories, possibly even subcategories, ultimately down to a detailed product level.

Marketing’s perspective on revenue

In e-commerce, the marketing department is responsible for bringing users to the online store and bears the costs of the marketing mix to attract these users.
The challenge is how to attribute generated revenue to specific marketing channels. Marketers view revenue through what is known as the revenue model.

The outcome of how much revenue e-commerce generates depends on how many users it brings to the online store, the site’s conversion rate, and the average order value.
The conversion rate is simply a measure showing what percentage of visitors to our online store convert into transactions.

This model can be perfectly represented by a funnel diagram. Those familiar with marketing, particularly online marketing, know this model well. It illustrates our website, especially the online store, as a place where a large number of users enter the top of the funnel. Ultimately, a small percentage of these users, usually a few percent, result in transactions.

Users come to the site from various places on the internet through marketing activities known as the marketing mix. Depending on the industry we operate in and our marketing strategies, this marketing mix will generate different user acquisition channels.

These include Google search, display ads on portals, social media, price comparison sites, or our own resources like an email database and newsletters.

How to determine channel effectiveness based on generated transactions?

The key challenge in marketing is determining the effectiveness of individual marketing channels based on generated transactions.

In reality, every user follows a specific purchasing process, which depends on the product being bought, whether it’s comparable, complex, and most notably, its price.

Although it might seem straightforward, imagine a situation where you’re reading an online portal and see a Media Expert ad for a PlayStation console. Clicking on it, you see that it costs 500 euros.

If you click the ad and buy the product, it’s easy to attribute this revenue to the display ad. However, if you don’t buy it immediately and instead discuss it later with friends, eventually searching Google with the query “PlayStation or Xbox,” you might find Media Expert again via organic search.

Still undecided, further discussions might convince you to buy a PlayStation. Next, you search “which PlayStation to buy” on Google, click on a sponsored link from Media Expert comparing different models, and finally, enter a price comparison site to find the best deal and make your purchase for 250 euros.

Summarizing this scenario, you encountered the display ad first, then organic search, followed by a paid search ad, and finally a price comparison site, resulting in a 250-euro purchase.

In terms of cost, marketing departments know the expense for each marketing channel. Assume these are the cost distributions across the channels.

Thus, to generate 250 euros in revenue, we spent 7.50 euros to attract this user to the site.
Assume a 10% gross margin on this product, giving us 25 euros on the 250-euro sale. The problem arises when attempting to split this revenue across individual marketing channels to calculate the return on investment (ROI) for each channel.

In practice, with hundreds of thousands of users, these conversion paths will vary widely—some shorter, some longer, with varying channels. Especially for complex products like booking international vacations online, there can be up to an average of 30 site visits before purchase, each potentially from different traffic sources.

Models recording user data on the website

Tools that record data on user navigation and origin typically assign 100% of the revenue, meaning in marketing, 100% credit for the conversion, to the last source. This is known as the last-click model.
This model is unfair to initial traffic sources since ROI appears higher for comparison sites and lower for initial traffic sources. Cutting off initial sources might prevent users from discovering our site, missing out entirely even through comparison sites.

There are other models, such as first-click attribution, but it also has flaws. Linear models, which are also not ideal, tend to distribute credit equally – akin to everyone at an Olympics receiving the same reward regardless of their performance.

Two other models pre-assign a distribution of revenue percentages across channels, which can be problematic because of the diverse paths and sources contributing to the revenue in the marketing mix.

Attribution modeling

In Conversion, we treat the Marketing Mix like a football team. We can’t solely credit Lewandowski for the team’s victories. Marketing Mix, like any team, functions cohesively. Evidence of an effective Marketing Mix is akin to shifts in public discussions like, “Robert Lewandowski hasn’t scored a goal in 670 minutes on the field,” underscoring the collective effort.

Therefore, in marketing, we use data from analytical systems for attribution modeling. Attribution modeling employs more statistical mechanisms, integrating science into business. The results of attribution modeling are known as data-driven models.

Data-driven models accurately state the percentage of the final conversion attributed to aggregated paths due to the multiple paths leading to revenue. This allows us to precisely calculate ROI, enabling financial controllers to decide which channels to invest in.

The effect of these models is decisions on reallocating the marketing mix budget to generate a higher impact with the same cost or achieve the same effect while saving costs. Marketing aims to be efficient, and these models serve this purpose.

Summary

Although the explanation might seem straightforward, a significant challenge lies in the quality of input data. The principle of GIGO (Garbage In, Garbage Out) states that using these mechanisms with low-quality data renders data-driven model calculations ineffective.

Nevertheless, proper attribution itself remains a considerable challenge.

The post Challenges in online marketing first appeared on Conversion.

Everything begins with the acquisition of users of our site. In this article, I will describe how to analyze user acquisition in Google Analytics 4, how to prepare for it, and which reports to pay attention to.

What to ensure before analyzing acquisitions from GA4?
Tagging marketing campaigns
Reports for analysis in Google Analytics 4
User acquisition report
Traffic acquisition report
Summary

What to ensure before analyzing acquisitions from GA4?

Before you start analyzing acquisitions from Google Analytics 4, you must first ensure the data quality. The well-known acronym GIGO (Garbage In, Garbage Out) means that if poor data is used for analysis, the conclusions and recommendations from this analysis will also be inaccurate. It is worth referring to materials dedicated to auditing Google Analytics 4 to ensure that data is correctly collected and processed.

Our GA360 clients usually have a department or dedicated person responsible for data analysis, drawing conclusions, and formulating recommendations. It is worth noting that proper preparation and analysis of data are key to effectively utilizing Google Analytics 4 in the process of acquiring users.

Check the tips below to ensure your data is correct. If you are confident that the data is accurate, the next step in preparing for GA4 analysis is to ensure proper tagging of all elements within the marketing mix.

Tagging marketing campaigns

The foundation is tagging marketing campaigns. For this purpose, we use basic parameters.

In Universal Analytics, we are familiar with five parameters:

• {utm_medium}
• {utm_source}
• {utm_campaign}
• {utm_content}
• {utm_term}

In the case of GA4, four additional parameters are added:

• {utm_id}
• {utm_source_platform}
• {utm_creative_format}
• {utm_marketing_tactic}

Marketing parameters should be collected and defined within the tagging scheme. When performing online activities, all elements that drive traffic to your site should be properly tagged. Links leading to your resources must contain the parameters used in the tagging scheme, which clearly specify their values.

This is crucial for accurate data analysis. Through this, you will know what each campaign means.

It’s worth using tools that can help with this. We created such a tool for Universal Analytics, but it can be easily extended to include the four additional parameters used in GA4. Despite the availability of these parameters, they are not always utilized. The five basic parameters I mentioned are sufficient to manage the entire Marketing Mix.

Reports for analysis in Google Analytics 4

In Google Analytics 4, in the context of user acquisition, we have two distinctions and two reports.

• The first concerns user acquisition, i.e., where the user first arrived on our site.
• The second concerns session acquisition, i.e., what source the next visit comes from. These distinctions are found in two different reports.

The reports worth starting the analysis from are found in the acquisition section. By entering the reports section, we find the acquisition subsection, which includes both traffic acquisition and user acquisition.

User acquisition report

User acquisition encompasses all channels, which can be modified in the context of UTM parameters. By default, this is the default channel grouping imposed by GA4.

Custom channel grouping is also available, but I won’t elaborate on this. If you need a different breakdown of channels, you can always ask the person responsible for implementing GA4 in your company.

In the user acquisition report, dimensions such as first user, new user medium, new user source, first source medium, and others are available.

The primary dimension is the first source medium, which shows where users were acquired for the first time or after clearing cookies. In web analytics, a new user is someone who arrives at the site from a new device and isn’t remembered as a returning user.

Customizing the report for analysis can be simplified with two useful tricks.

• Firstly, you can collapse the sidebar to see more columns. Continuous horizontal scrolling can be bothersome.
• Additionally, it is worth reducing the browser window. On a Mac, this can be done by selecting Command and Minus, and on Windows – Control and Minus. This way, the entire table will show our source medium through the lens of the most important metrics.

During data analysis, metrics regarding user acquisition and their engagement are particularly important. The “new users” metric informs us about the number of new visitors, while the engagement rate shows whether the user stayed on our website. The engagement metric is crucial because it tells us not only if the user arrived on our site but also if they spent time there. Bringing in empty users, those who enter and immediately leave the site is pointless.

A crucial metric to pay attention to is the engagement rate. It informs about the percentage of users who stayed on the site. When talking about user acquisition, it’s unlikely that someone who didn’t stay will return later.

Another important aspect is conversions, which are linked to conversion analysis. Conversions are now called key events. We can choose key events we wish to analyze. Although this report doesn’t have substantial analytical power because it merely shows what’s happening on our site, if more advanced analysis or data segmentation is needed, it’s worth using the comparison options.

To choose a comparison, the appropriate option is in the top right corner of the report. For example, we can compare mobile traffic with overall traffic. After selecting “apply,” these segments will be overlaid on the report, allowing for more detailed analysis.

While we have a basic comparison, additionally, in user acquisition, we can select an additional dimension. For instance, we can see from which parts of the country or region users were acquired and what devices they are using. If that’s not enough, in the more advanced use of reports, we can select the option in the top right (though in this case, it’s unavailable) to transfer the report and further edit it in explorations. This topic will be covered in a separate article in the future.

Traffic acquisition report

The second report in the acquisition section concerns traffic acquisition. To view it, go to reports and select traffic acquisition.

Analyzing traffic acquisition will look similar. By default, this is channel grouping which we can modify by choosing dimensions related to session acquisition. It’s worth noting that if a user first arrives from Google Organic and then comes from Facebook, the initial attribution will be to that traffic source. Each subsequent session will be separately attributed to different sources. This way, elements are broken down.

The functionalities that can be used here include comparison, filtering, and adding additional dimensions. Basic functionalities allow you to get an overview of this section. A deeper analysis can be conducted in the explorations section.

However, it’s worth noting that the most valuable insights can be drawn using BigQuery. If you haven’t yet connected your GA4 to BigQuery, it’s worth doing so as soon as possible to start accumulating data there. Then, you can ask individuals with the appropriate competencies to analyze it.

Summary

At the e-commerce or marketing manager level, there is rarely the need to delve into advanced configurations, analyses, and segmentations. The reports section, user acquisition, or session acquisition should be sufficient. For more advanced analyses, it’s best to use BigQuery, for which the appropriate skills will be necessary.

Alternatives for an in-house web analyst vary. This entry aims to help better understand what’s happening in your marketing mix and how you are acquiring users.

I encourage following the “GA4 for Managers” series, where we present the skills necessary for every manager working in e-commerce and marketing and how to effectively use Google Analytics 4.

The post Acquisition analysis in Google Analytics 4 first appeared on Conversion.

GTM Server-Side – what it is, how it works, how to implement it, and common questions arising in Google Tag Manager Server-Side implementation projects. In today’s article, we will discuss GTM Server-Side, how to use this tool in online marketing, and answer some of the most frequently asked questions regarding its implementation.

Google Tag Manager as the foundation of marketing
Differences between GTM Server-Side and Client-Side
GTM Server-Side – advantages
How to implement GTM Server-Side?
How much does GTM Server-Side cost?
Does GTM Server-Side differ from measurement protocol?
What is needed to implement GTM Server-Side?
What tools can be implemented using GTM Server-Side?
How does GTM Server-Side affect data collection?
Can GTM Server-Side be implemented outside Google Cloud Platform?
How long does it take to implement ssGTM?
Does GTM Server-Side work for mobile applications?
Summary

Google Tag Manager as the foundation of marketing

Google Tag Manager is an extremely important tool in online marketing. Effective online marketing is based on data. Data consists of scripts and tools that we install and implement within online marketing.

These tools are most often implemented using scripts and tags embedded in the source code of our website. An alternative to implementing source code for individual tools is using GTM, which is a content management system for scripts. Effective marketing should be based on Google Tag Manager (GTM). After implementing scripts via GTM or directly in the site’s source code, when a user visits our site, its content loads. These scripts interact with external tools, gathering data from the browser, system, and user interactions on the site. This data is then sent to third-party servers.

By visiting any website, we can see what information is being gathered and sent to third parties. Using developer tools available in the browser menu, in the “Network” tab, we find a summary of all scripts, especially JavaScript scripts, that are loaded. As part of online analytics, we can monitor what scripts are loaded in the browser. For example, by using the Collect function, it is possible to see what information is sent from the browser to third-party tools. This data allows for generating metrics related to user behavior and optimizing marketing efforts.

Differences between GTM Server-Side and Client-Side

In the case of GTM Client-Side, all scripts implemented on the client-side send information directly from the browser to third-party tools.

On the other hand, the Server-Side container is a tool installed on a server. The browser communicates with this container in a way that instead of sending multiple requests to various tools, a single packet of information is sent to the server.

On the server, data is divided into portions, and the relevant information is forwarded to third-party tools. All the load, which was previously handled by the computer and the browser, is taken over by the server. Naturally, we will incur costs for maintaining the server, as it needs to perform operations to distribute the information to all implemented tools. Thanks to the server configured with GTM Server-Side taking over the interaction load, the performance of our site has improved.

GTM Server-Side – advantages

In the browser console, one can notice that many scripts are loaded, consuming the computer and the browser’s computational power. After implementing GTM Server-Side and sending a single packet of information from the browser to the server, dozens of other communications are eliminated. The entire load is taken over by the server, which then distributes the information further.

As a result, our website’s performance improves since the computer has more resources available to load content. The content is delivered to the user faster, positively affecting the Core Web Vitals metrics. This, in turn, has a favorable impact on rankings in organic searches.

Additional benefits of utilizing GTM Server-Side include the ability to enrich the data sent to the server with information that we may not want to reveal to the user at the browser level. This could include information about margins or the cost of purchasing a product. Enriched data, along with user information, is sent to tools like Google Ads, allowing for ad optimization criteria based on more advanced data, such as product margin.

Besides adding information at the server level, we can also remove certain data. Suppose we do not want to send personally identifiable information to third-party tools, following GDPR compliance. In that case, we can remove it from the data packet before distributing it to other tools.

How to implement GTM Server-Side?

Since the packet of information is sent to a server, usually configured on our subdomain, there is a lower level of blocking these requests by ad blockers and various tools. These tools operate by identifying where the individual signals are going. If the signal is on the list of things to block, it is not forwarded. Implementing GDPR increases the number of registered conversions, as ad blockers do not act on these types of information sent to our server.

The process can be divided into three steps:

• Server Configuration: Set up the server where the container will be hosted.
• Tag Configuration in the Server Container: Configure tags in the server container.
• Redirecting Information Transmission from the Client-Side Container: Ensure that information does not go directly to the tools but to our server. The server will know where to distribute this information based on the configuration from step two.

In practice, within implemented projects, we check the server load to optimize costs. Basic tools from Google Platform for setting up the server container include App Engine and Cloud Run. Before choosing the first step and configuring the server container, we analyze the traffic to estimate container handling costs and optimize them.

The second step, post-container implementation, involves assessing its resistance to ad blockers. GTM Server-Side is generally resistant to ad blocker operations, but this can still be refined. Therefore, we conduct analyses to evaluate this resistance. If it is at an appropriate level, we leave the configuration unchanged.

Additionally, we optimize client-side tracking, as not everything needs to be tracked through client-side solutions. Companies that implement GTM Server-Side often have duplicated tracking, meaning the same information is being sent from both containers. When implementing GTM Server-Side, we are always mindful of this. Along with configuration, consider the possibility of enriching the data. Clients often need this enrichment, adding information about the product manufacturing cost. This way, campaigns can be optimized based on actual margin and direct product profit, not just revenue. Different products can have varying margins, which is critical information.

For larger organizations with many containers, we conduct analyses to optimize their number. Often, it turns out that a large number of containers is unnecessary. This is especially true for organizations with many websites, where the Client-Side container structure was often not well thought out.

How much does GTM Server-Side cost?

One of the most common questions clients ask about GTM Server-Side is the cost of such a solution. This cost should be divided into two factors.

The first aspect is the implementation cost, and the second is the maintenance cost.

Maintenance costs include managing the implementation of new tools and operating the server hosting the container. With a request rate of about 4-5 requests per second, the cost is around 100 euros. For smaller services, these costs can be at most 250 euros and often close to 25 euros per month.

It is also essential to consider the cost of implementing and maintaining new tracking.

Does GTM Server-Side differ from measurement protocol?

Measurement Protocol is a feature of Google Analytics that allows sending data to GA4 without the user being present on the site. If we have devices that do not support browsers or are not mobile applications, such as kiosks in physical stores, we can send information directly to GA4 using Measurement Protocol. Measurement Protocol is often used to deliver transactions that would not normally be recorded by GA4.

Google Tag Manager (GTM) is not solely for GA4. GTM is used to send data to various tools. GTM acts like a pocket that gathers information from the browser and then distributes it to the appropriate tools. Whether the data needs to go to GA4, Google Ads, or another tool, GTM can send it to the chosen system. Measurement Protocol will not replace Google Tag Manager Server-Side and vice versa, as these are two completely different tools.

What is needed to implement GTM Server-Side?

First and foremost, a configured server is needed. This means having the space to install the server container. Clients typically choose space within the Google Cloud Platform, using tools such as App Engine or Cloud Run.

Cloud Run is increasingly chosen due to its greater scalability, allowing better cost control. With App Engine, we must determine the number of machines to handle traffic in advance. Cloud Run offers scaling options, meaning the machines will automatically adjust to demand.

Other environments like Azure or AWS can also be chosen. From our experience, the biggest challenge is creating the appropriate space. Other issues, such as server-side configuration and redirecting data from GTM Client-Side to Server-Side, can be handled by us or the web analyst responsible for the implementation.

What tools can be implemented using GTM Server-Side?

At one client’s site, we have a list of about 36 tools, some of which will be implemented via GTM Server-Side. Google offers many solutions, such as Analytics, address tracking, Floodlight, and other Marketing Platform tools.

These include pixels for Criteo, Pinterest, TikTok, Twitter, LinkedIn, and others that rely on implementing scripts sending information directly to these tools. Everything we implement on the client-side can also be implemented server-side, i.e., Server-Side.

How does GTM Server-Side affect data collection?

This can be summarized in two aspects. Firstly, it increases data collection. Tracking performed using GTM Server-Side is more resistant to ad blockers, increasing the level of recorded data and conversions. Secondly, it modifies the data. This can mean both enriching data by adding additional information and removing information we do not want to pass on to third-party tools.

Can GTM Server-Side be implemented outside Google Cloud Platform?

Absolutely. An alternative in the case of Microsoft Azure is Azure App Service. With Amazon’s AWS, it’s Elastic Beanstalk, which is very similar to App Engine in Google Cloud Platform. Elastic Container Service corresponds to Cloud Run. This way, other clouds can be used to handle the server-side container.

How long does it take to implement ssGTM?

The process can be divided into two parts. The first concerns configuring the server and the container on that server. The second part relates to moving the track from the client-side to the server-side and depends on the current GTM Client-Side configuration.

The more complex the configuration and the more data you transmit, the longer this stage will take. On average, such implementations last from 3 to 6 months, including switching tracking from Client-Side to Server-Side.

Does GTM Server-Side work for mobile applications??

Currently, no, although this may change in the future. Within the next 1-2 years, ssGTM will become the standard in online marketing tracking.

The sooner you implement ssGTM, the faster you will benefit from its advantages, such as faster website performance and enhanced data transmission. ssGTM also allows for the registration of more data, avoiding, for example, ad blockers. Therefore, it is prudent to delve into this topic as soon as possible.

Summary

Implementing GTM Server-Side offers many advantages and positively impacts many aspects of online marketing, such as Core Web Vitals and GDPR-related issues. I am convinced that ssGTM is the future of tracking and data collection on the internet. Hence, it is worth ensuring proper implementation now and improving data collection activities on your site.

The post GTM Server-Side – how does it help in marketing? first appeared on Conversion.

What is Google Analytics 360?
Differences between GA4 and GA360
Key benefits of Google Analytics 360
Summary

What is Google Analytics 360?

Google Analytics 360 is a paid version of the popular digital analytics tool, offering advanced features and capabilities that can significantly support business development. For many organizations, Google Analytics 360 provides features that are essential to deep-dive data analysis and reliable metrics management, with the premium version characterized by the lack of data collection limits and guaranteed Service Level Agreement (SLA).

Differences between GA4 and GA360

Deciding whether your organization needs the paid version of Google Analytics is crucial to understanding the differences between the free and paid versions. In today’s world, where data plays a key role in the decision-making process in companies, choosing the right tool to analyze this data is essential. In the case of Google Analytics, users face a choice between the free version and the paid version, known as Google Analytics 360 (GA 360). Differences between the two can be divided into several categories, particularly noticeable in terms of support and performance.

Firstly, the Google Analytics 360 license offers an SLA, which ensures the reliability of data collection, updates, and report availability. Additionally, the paid version differs from the free one in terms of the availability of various features, such as the number of reports available in exploration, the number of audience groups, or the number of conversions, now referred to as key events.
Differences between GA4 and GA360, part 1

From my perspective, aside from the SLA, one of the biggest differences between these versions is the event limit, which in the free version is limited to one million events per day. This limit may be a significant barrier, especially for websites with high traffic and those with mobile applications.

Additionally, the free version offers access to historical data for only 14 months, while the paid version extends this period to 50 months. For companies that rely on long-term analysis of trends and patterns, this aspect may be decisive when choosing the paid version.
Differences between GA4 and GA360, part 2

Key benefits of Google Analytics 360

As you can see, Google Analytics 360 significantly differs from its free version in many aspects. When making the decision to purchase the extended version, consider several key aspects, such as:

• Data certainty and stability
• The desire to analyze long periods of time (beyond 14 months)
• The level of advancement of your company’s digital product tracking structure
• The amount of data collected

If your analyses cover broad time ranges and you make intensive use of the interface, GA 360 will undoubtedly facilitate your work. Its advanced features allow for more efficient data processing, especially for long-term analyses.

For companies with complex organizational structures, offering a variety of sites and digital products, GA 360 enables detailed analysis through subproperties and roll-up properties. These features, which are the equivalents of views from Universal Analytics, allow for more precise tracking and data analysis.

Summary

To those who manage large volumes of data, I definitely recommend considering investing in Google Analytics 360. In this regard, I invite you to check out our guide, and if you have additional questions, I encourage you to contact us directly.

The post Does your company need Google Analytics 360? first appeared on Conversion.

The first source of truth – a reference point
Data presentation in Google Analytics
Data in Google Analytics vs. RODO
What level of data discrepancy is acceptable?
How to check the level of data discrepancy in a service?
Data discrepancies depending on the type of service
Google vs. other advertising systems/a>
How to reduce data discrepancies with changes in transaction recording
Summary

The first source of truth – a reference point

At Conversion, we most often work with e-commerce services, and as a result, the main point of reference in the projects we carry out is the warehouse and accounting system. It is the one that most often provides the previously mentioned first source of truth, relative to which we compare other data such as the number of transactions or revenue from those transactions.

When we talk about comparing data from Google Analytics to that from a company’s internal system (transactional system or CRM), we need to be aware that we are talking about two areas of data comparison – relevance and accuracy. Let’s start by clarifying these two key terms.

Relevance vs accuracy in Google Analytics

We talk about relevance when some external tool (here Google Analytics) shows exactly the same data that we see in our reference point. For the purposes of this article, let’s assume that in ecommerc’s case it is a CRM system. For example, we can find data on orders placed in our store. If Google Analytics collects data characterized by accuracy, the number of transactions will be equal to that in the internal system.

When examining the accuracy of the collected data, we no longer pay attention to the exact representation of the data in quantitative terms. Here, trends are a much more important element. If the number of transactions in our CRM is growing at a given rate during the period under study, this should also be reflected in our Google Analytics.

Data presentation in Google Analytics

Google Analytics collects – and consequently – presents data based on a couple of foundations. The first of these is JavaScript, which is embedded in the site’s source code or inserted into the page using Google Tag Manager. It is triggered when the page is loaded. Its task is to create and read cookies, which contain a unique user ID within them. In this case, we say that Google Analytics operates on the basis of JavaScript. However, not all visitors to our site have JavaScript or cookies enabled. Users also often use plug-ins that intentionally block not only ads, but also Google Analytics scripts. In such a situation, the actions performed by the user will not be tracked.

Now let’s return to the concept of accuracy. As we mentioned before, the main function of Google Analytics as a tool of the Digital Analytics class is not to show exactly the same data as the internal system. Its main purpose is to link the source of a user’s traffic (the place from which they came to the site) with their behavior on the site once they got there. It gives website managers the information they need to assess how, depending on the traffic source and behavior on the site, the user performs the actions they want – that is, they make conversions. So we need to remember that web analytics tools exist to answer questions about how to achieve the goals we have set for our site, not to collect 100% accurate data. This, unfortunately, is not possible due to the blocking of some of the information shared by users.

Data in Google Analytics vs. RODO

We live in an era of increasing concern for user privacy (GDPR). For some time now, website owners have had to take a proactive approach to obtaining user consents for the creation and use of cookies. It is obvious, then, that the more users accessing our service do not give this consent, the greater the discrepancies in the data will be. For this reason, Google Analytics will never reflect 1-to-1 the data that is collected in the internal system. That’s why it’s so important to study the trends we observe in CRM and compare them with those noted in Google Analytics 4.

Often, when working with clients, it happens that when we don’t have 100% of transactions recorded in Google Analytics, they are “sent” to it, e.g. via measurement protocol. This is not an appropriate approach to the subject of data discrepancies, for the reason that was mentioned earlier – Analytics is used to evaluate the effectiveness of traffic according to its sources or site behavior – not to collect fully complete data. If we “send” transaction data from CRM to Analytics, which it did not record due to the user’s cookie blocking, we will not have information about the related traffic source or user behavior on the site. As a result, we will not be able to take a closer look at the transaction and will not get valuable information about it.

So let’s keep in mind when using web analytics tools about their main function and not require them to be fully accurate and relevant – such a situation does not happen in real life.

What level of data discrepancy is acceptable?

Having reached this point in the article, you’re bound to wonder what level of discrepancy you shouldn’t be concerned about. Let’s assume that the data you see in Google Analytics is characterized by accuracy – the trends correspond to those seen in the CRM system. So let’s consider what level of accuracy we can consider appropriate.

In the projects we carry out, we aim for a level of data convergence in Google Analytics with internal systems of 85%. This means that for every 100 transactions recorded in the CRM system (the actual number of transactions on the site), an average of 85 should be reflected in the data in Google Analytics.

How to check the level of data discrepancy in a service?

Thankfully, there is a simple way to do this on ecommerce sites. In the internal transaction system (the first source of truth), we have data on all transactions made on the service, along with the ID assigned to the users making them. With Google Analytics configured correctly, we will see the same transactions in it, with the same assigned ID.

So the simplest thing to do is to export the data from the internal system and compare it with the data available in Google Analytics, and then see how many IDs from the CRM system are missing in Analytics.

In order to make the most valuable comparison, it is also important to pay attention to the characteristics of users, by which we can segment the completed transactions visible in CRM and invisible in Google Analytics. This will provide additional hypotheses about what the discrepancies between the systems might be due to – and that’s the first step to reducing them.

Data discrepancies depending on the type of service

We mentioned before that in most cases 85% data convergence is the level we should strive for. After delving more deeply into the subject, however, the answer is not so zero-one. A satisfying level of divergence also depends on the type of service – or, to be more precise, on the characteristics of the service’s users.

We have to realize that, as marketers, we are characterized, in general, by a higher level of awareness of Internet use. However, this does not mean that “ordinary” users are homogeneous in this respect. This is also reflected in the level of discrepancy in the data.

On websites that collect more aware users, they are more likely to have disabled browser functionalities on which Google Analytics collects data, such as JavaScripts and cookies. They can also block the transmission of information about themselves to the tool with special add-ons, blocking not only ads, but also tracking of their online activities by Google Analytics. For such sites, the level of data convergence will be noticeably lower. This is mainly the case with specialized sites, especially in the IT industry, where the level of convergence will reach “only” 40-50%.

On the other hand, for websites visited by moderately less informed Internet users, such as clothing or electronics stores, we can expect a data convergence level of 85% or higher, as mentioned before.

Google vs. other advertising systems

Discrepancies between internal systems and Google Analytics, is not the only challenge analysts face. Differences in data will also be noticeable between different advertising systems (assigned to different online marketing tools) and Google Analytics 4. This is despite the fact that these tools are based on the same JavaScript technology. So why does this happen? We should look for the answer in the attribution model used.

As an example, let’s take Facebook’s advertising system, Facebook Ads. It will strive to show the highest possible number of conversions made through ads in this system in order to attract advertisers who are encouraged by the results. On the other hand, Google Analytics receiving this data no longer has such an interest. So we can assume that the data in Google Analytics 4 should be more objective.

To illustrate this sort of war of giants between companies, let’s take a look at ads in Facebook’s mobile app. Like most users, we probably have an in-app browser installed on our phones, running on what is known as WebView. In this case, when we switch from Facebook to a third-party service, it is not displayed in the new browser, as a result of which access to data from the Google Analytics perspective is blocked. This is why the data available in Facebook’s advertising system will notice and note this action – unlike Analytics. As a result of this action, as advertisers we are encouraged to use the data in Meta’s advertising system – because that is where we will see the data.

When analyzing the data, we need to keep these nuances in mind and choose the most appropriate (reliable) attribution model for us, and it is this model that will guide us in further analysis. Each of them has its advantages and disadvantages, so it is the decision on how we want to analyze the data that is crucial – after all, we don’t want to end up in a situation where the number of conversions coming from the advertising systems used is several times higher than the actual number of transactions visible in the CRM system.

How to reduce data discrepancies with changes in transaction recording

One of the most common problems in the field of ecommerce data discrepancies are those related to transaction registration. These arise when using third-party payment gateways. In this case, transactions are registered by default when the user returns to the site. However, users very often do not return to the service after making a payment, which causes a lot of discrepancies….

How to deal with this? We recommend counting transactions in Google Analytics just before going to the external payment itself. Our observations show that there is a much higher probability that a user will not return to the service after making a payment, than that he or she will go to make a payment and abandon it immediately afterwards. This is caused by two things:

• Call to action in ecommerce – or rather, how it is phrased. In most cases it reads “place order with payment” or the equivalent, suggesting that payment will be required in the next step. Since the user is aware of this, he or she is much more likely to give up before moving on than when they get to the point where they were informed of the payment obligation.
• reason for not paying for the order – not paying for the order after going to the payment gateway is usually the result of an unexpected error or random factors such as forgetting your bank login information. However, there are mechanisms to restore the previously lost shopping cart and return to the payment, which largely eliminates the problem.

Because of this, in order to mitigate data discrepancies, we recommend setting up conversion counts just before going to the payment gateway.

Summary

When using web analytics tools, let’s remember what their purpose is and not treat them as the only source of truth. The data collected by such tools should be characterized by accuracy (not relevance) and this is what we should strive for, and a satisfactory level of convergence in most cases is 85%. However, depending on the type of service, this will not always be possible, and we should keep this in mind as well. In analysis, let’s pay more attention to the consistency of trends between Google Analytics and CRM, rather than accurately reflecting the number of transactions. In this way, informed analysis will lead us to more valuable conclusions!

The post Data discrepancies in Google Analytics – what do they stem from and how to minimize them? first appeared on Conversion.

Have you failed to deliver KPIs in the last quarter, month or year? It happens, especially in a challenging and often changing market situation. However, there is a way to get back on track with your business goals! To do so, you need to go through the process of analyzing data, drawing conclusions and making hypotheses to avoid such a situation in the future. Web analytics data will be an invaluable help in this process!

Success in e-commerce is not only a matter of providing attractive products, but also the ability to manage and analyze data. Key Performance Indicators (KPIs) play a fundamental role here, allowing you to assess the effectiveness of your strategy and identify areas for improvement. In this article, we will discuss how to improve the KPIs of an e-commerce site using online data analysis.

Revenue model in E-commerce services
Analyzing data and understanding causes
Using A/B Testing
Useful reports in Google Analytics
Summary

Revenue model in E-commerce services

Before we start any data analysis, it’s important to understand what drives our revenue. In e-commerce services, there are 3 main metrics: number of users, average order value and conversion rate.

1. Number of Users: The first element is the number of users who come to the site. In marketing efforts, we should aim not only to increase their number, but also to bring users who are most likely to make a purchase.
2. Average Order Value: Another factor is the average order value. In purpose of increasing the value of this indicator, it is worthwhile to use promotions when buying more products, or at least to recommend to users products that match those already added to the cart.
3. Conversion Rate: The third element is the conversion rate. It represents the ratio of the number of transactions to the number of users. In our activities, we should strive to increase this ratio.

The revenue that our service generates here is the result of the number of users and the conversion rate (which tells us the number of transactions) and the average order value. By multiplying these values, we get the total revenue of our store.

For example, let’s assume that our store was visited by 10 thousand users per month, generating an average order value of PLN 200, with a conversion rate of 5%. The revenue of the store will be PLN 100 thousand.

This formula, with minor modifications, can be applied not only to e-commerce, but to any service, converting average order value to average customer value over time. Similarly, by defining conversion as the desired action performed by a user on the site, we can study its coefficient and use it for analysis. In this article, however, let’s focus on e-commerce services.

Analyzing data and understanding causes

In most cases, in E-commerce services, the main KPI faced by store managers is revenue. While we already know what metrics it consists of, in order to properly analyze the reasons for lower-than-expected revenue, it is necessary to make a deeper decomposition of the component factors. As a reminder – here we will look at the number of users of the service, the average order value and the conversion rate.

The first step is to diagnose which (or which) of these 3 values is at a lower level than we assumed.

Number of users

In the case of a lower-than-expected number of users, we should take a look at the advertising campaigns being run. Based on the historical data of running campaigns, we can diagnose drops in their effectiveness. Have the current campaigns brought us the same number of users with the same budget? If not, this is a signal to decompose this area and study what is happening to the users targeted by the campaigns.

Important metrics here will be the number of page views and the CTR of each campaign. It may just be that some of them are ineffective, and these are the ones worth working on!

Average order value

With a lower-than-expected average order value, it’s worth looking at discount policies and up-selling strategies used in the store. Offering the user suggestions in an accessible way for products that match the ones they have already added to their shopping cart can prove to be a hit. It may also be a good idea to introduce promotions and discounts applicable to a specific order value, which should encourage the buyer to increase the value of the shopping cart.

Conversion rate

When we get to the point where we determine that the number of users and the average value of their transactions are at the right level, we should look at the conversion rate. This is a very capacious metric, so we should go into it in a bit more detail.

Each user, between entering our site and making a transaction, performs a huge number of intermediate actions. We can call them micro-conversions and arrange them into a kind of purchase funnel. For example – before making the final conversion (purchase), the user first had to reach the check-out. To get there, before that he probably already visited the order summary page, which he got to from the shopping cart, which he went to from the product card, and so on… Namely – the action that we should perform at this stage is the decomposition of the customer’s purchase path (that is, the previously mentioned funnel) and its in-depth analysis.

It is good practice to start this analysis “from the end”. So let’s start with the check-out of our service. To standardize, if the check-out closing ratio (understood as the ratio of the number of transactions to the number of users/sessions in the shopping cart) is below 40%, this is definitely an area in need of improvement. If it is above 70%, we can safely conclude that everything is fine at this stage of the customer path.

The next step should be to look at the ratio of users in the shopping cart to those who reached the product card. Here it is worthwhile to refer to historical campaign data and see what actions on the product card are performed by users who did not go to the shopping cart. This will allow us to diagnose what changed their decision. It may also turn out that users acquired from current campaigns have a visibly higher rate of this particular micro-conversion. This may be an indication of their inadequate customization.

In analyzing the customer path, it is worth going into as much detail as possible to effectively diagnose problems. On each subpage visited during the buying process, we can decompose separate funnels of this kind. For example, when filling out a form with personal information, it is worth looking at at which stage (after interacting with which field) the largest portion of users drop out.

A meticulous analysis will help us diagnose the problem areas on our site. But… what’s next?

Using A/B Testing

The first step of conducting an analysis based on online data is already behind us. But how to use the information gained from it? After mapping the areas for improvement, it’s time to make hypotheses about what should be done or changed to improve conversion rates at different stages of the funnel.

Once the hypotheses are set, it’s time to verify them. This is where A/B testing comes to our aid. Let’s assume that in the course of analysis we noticed that in some of the campaigns we run the creatives have changed. Coincidentally, there was also a decrease in the engagement of users who were on the product card. So the negative impact of the creative change here will be a hypothesis that we will verify.

It is A/B testing that we will use to verify this hypothesis. We will display both version A (the original) and version B (which has changes in the creation resulting from the hypothesis we are analyzing) to the recipients we are targeting with our ads. Over the course of the test, with two variants of the ad creation, one of them will go to one half of the users, and the alternative to the other. Similarly, when the tested creatives will be more. With the data from these campaigns, we will be able to determine which of the creatives results in higher audience engagement – that is, the rate of transition from the product card to further stages of the purchase process. Ultimately, this will allow us to assess the validity of our hypothesis.

If the question popped into your head whether the changes in the campaign resulting from the hypothesis can’t simply be implemented to it and compare the results of the next period with the previous one – not the best idea. We need to keep in mind that campaigns are also affected by external factors, which can differ significantly from period to period. These factors include, for example: the economic climate, the actions of competitors, seasonality, etc. The obvious conclusion, then, is that A/B testing will give us the most reliable information when verifying our hypotheses.

Useful reports in Google Analytics

When looking for the causes of under-reported KPIs in the data in Google Analytics, it is much easier to find them when we have a clearly defined problem. A completely different, but also very important topic, is analyzing data when the results are at the right level. Often there is then no clear motivation to work with the data, although there are always smaller or larger areas for improvement. It’s a good idea to start by monitoring the 3 reports described below, which definitely make it easier to analyze online data in E-commerce on a regular basis.

Funnel report

The first report we recommend is the funnel report, which will show us at what stages, after hitting the product card, users do not make further conversions. In e-commerce services it is worth paying special attention to the tightness of the checkout – if we note a value below 40% there is probably room for improvement.

To get to the report in Google Analytics 4, go to reports -> revenue generation -> path to purchase (the name may vary, depending on when the report appeared and when we set up the account). Compared to UA, in GA4 we have an important change – this report shows the flow of users who went through the next stage of the funnel, rather than the number of events (i.e., the number of additions to cart vs. start of checkout vs. purchase). We can analyze the funnel per device category, country, region, city, language and browser as standard.

Landing pages report

The second valuable report is the landing page report. Thanks to it, we will learn how users get to the site, as well as check at what level the engagement rate on specific subpages of the site is (the inverse of the rejection rate). In a situation where the mentioned coefficient is not at the right level, it is worth working on the consistency of the advertising creation with what we present on the product card.

To get to the landing page report, go to reports -> engagement -> landing page. This is one of the newest reports to appear in Google Analytics 4. If you don’t see the report here, it should be available in the library, from which you can add it to a set of reports anywhere you choose. The report differs strongly from the one available in UA, which was basically a copy of the source/medium report. In the case of the report in GA4, we have information about all users (including new users) who started a session from a particular subpage of the site or app screen. What’s missing, however, is information on rejections or engagement and the number of subsites per session, which was available in the UA report. However, nothing prevents you from adding such data by editing the report using the pencil icon available in the upper right corner. When analyzing, it is worth paying attention to the “Conversions” column and marking only those conversion events that you want to analyze.

Service effectiveness report towards technology

The third valuable report in Google Analytics is the report of the site’s effectiveness towards technology (e.g. browser or screen resolution). Sometimes at the development stage of a website, important technical elements are overlooked. This could be, for example, a call to action button placed under a page wrap, or a service element not displaying correctly in a particular browser or its version.

To access the site’s effectiveness report against technology, go to reports -> technology -> technology related details. This report will show us potential issues related to the devices and browsers the user is using.

When analyzing the web version, it is worth paying special attention to such dimensions as browser, device category, screen resolution and device make or model. It is worth remembering that in the case of Apple-branded devices, we will get very limited information about the device itself. In the case of browsers – it’s worth looking at the report more closely if any of them has recently rolled out an update (this can cause problems with the operation of our site, especially in the case of Firefox and stores on PrestaShop). New is an additional device category, smart tv, which includes TVs with built-in browsers, as well as gaming consoles such as PlayStation and Xbox.

When analyzing traffic from an app, be sure to check the version users are using – an outdated app can generate errors. Also take a peek at the Overview report in the Technology folder – there you will find additional information about the app’s stability and potential bugs in its operation.

Monitoring the previously mentioned reports will be a good start to conducting regular analysis of the site. It is also a good practice to pay special attention to the pages that generate the most traffic. When running E-commerce advertising campaigns, these will mostly be product cards. We should also pay attention to the check-out page from our store, where we can often easily diagnose a simple-to-solve problem affecting a low conversion rate. Let’s not fool ourselves – when a user reaches the very end of the purchase path, it means that he is determined to buy, and a low conversion rate at the last stage of the funnel is a clear signal that we should improve something here.

Summary

In conclusion – improving KPIs in e-commerce requires conducting meticulous data analysis and focusing on key components: number of users, average order value and conversion rate. Once the problematic component affecting revenue has been diagnosed, it should be decomposed as accurately as possible, hypotheses should be set and testing should begin. Using A/B testing to verify hypotheses and considering external factors are key aspects on the road to success.

The post Undelivered business KPIs – how to look for causes in the data? first appeared on Conversion.

How much does Google Analytics 360 cost? In this article today, we will take a look at the very topic of the price of the extended version of the most popular online data analytics tool – Google Analytics 360. Many companies considering purchasing a license of the paid version of this tool are afraid of high, unreasonable costs. However, it is worth taking a closer look and understanding how much it is and how to calculate the real price of the license. To begin with, however, you should consider whether, in your particular case, this particular investment will be the right solution.

When should you consider investing in Google Analytics 360?
How to calculate your Google Analytics 360 price?
What is included in the price of a Google Analytics 360 license?
Summary

When should you consider investing in Google Analytics 360?

Before answering the question of the price of Google Analytics 360, it is important to answer whether you really need an upgraded version of Google Analytics. Looking through the lens of our current clients, we note that companies using the paid version of this tool generate at least 5 million in net revenue per month. This immediately suggests that it’s an option aimed more at large organizations, operating with large amounts of data (high levels of website and/or mobile app traffic and a high number of events).

#1 Big number of events

As a first step, you need to analyze the limitations that Google Analytics 4 has. One of them is the one related to Google BigQuery. The free version of Google Analytics 4 is capable of sending a maximum of one million events per day to Google BigQuery. When determining the number of events, consider the number of users and the number of pages they visit. Going further, we come to the number of interactions they perform, such as clicking on page elements, scrolling through the page or playing videos.

In addition, we should add to this the events from the mobile app, where users are definitely more engaged and mostly generate more activity than in the case of the web app. After a thorough analysis of the number of events, ,it may turn out that the limit of one million events per day is insufficient, which is a clear signal to consider investing in a paid version.

#2 Internal analysis teams’ needs

As mentioned at the beginning, most often it is the larger companies that invest in Google Analytics 360. They often have specialized teams or at least dedicated people responsible for analytics. These experts are responsible for analyzing the data, presenting it, drawing conclusions and making recommendations based on the information gathered.

The presence of such people or teams indicates that there are probably defined processes in place, which include campaign optimization or reporting and conversion rate optimization activities. This makes analysis systematic and allows for effective data-driven action.

The presence of specialists makes the limitations of the free version of Google Analytics quickly become apparent in the course of analysis. This happens especially when we want to track user activity in more detail and segment data at a more detailed level than we currently do. Limitations related to data sampling, thresholding or other mechanisms that do not show the full picture of the data then become apparent and troublesome.

This is the second of the signals prompting companies to invest in Google Analytics 360 – the moment when a company needs to conduct advanced analysis on large data sets. To achieve the desired effect, the paid version becomes crucial, as it provides much more opportunities for detailed analysis of the collected data.

#3 An extensive ecosystem of sites and services

Another important reason to consider purchasing a Google Analytics 360 license is the expanded ecosystem of sites and services. In the previous version of the tool, Universal Analytics, there was a clear account structure that included a service with customizable views to help organize data. Unfortunately, in Google Analytics 4, the structure of the time has been abolished. This does not mean, however, that Google Analytics 360, does not offer facilities for managing complex data ecosystems.

The implementation of functionalities such as subproperties or roll up properties is becoming crucial for the effective management of a large ecosystem of sites and services in the new version of Google Analytics. Thanks to them, we can isolate a specific part of traffic within a single service, (e.g. for a specific product category or a given graphical market), which significantly facilitates data analysis. Subproperties thus allow better segmentation, while roll up properties allow aggregation of data into a single aggregate account.

An extensive analytics ecosystem is the third reason why a company should consider investing in Google Analytics 360. For extensive ecosystems of sites and services, the paid version of the tool is a key component for effective data management and valuable insights.

#4 Service Level Agreement (SLA)

The fourth key factor for companies to consider investing in Google Analytics 360 is the requirement for a service level agreement (SLA). If you haven’t encountered the previously mentioned limitations related to traffic, data analysis or ecosystem structure, it’s worth considering your current SLA. This is especially important for large corporations with international reach that require certain security standards to be met.

A service level agreement ensures that data is collected and processed in a way that guarantees security, which is especially important for businesses that operate on a large scale. It is worth noting that only the paid version of Google Analytics 360, meets these requirements one hundred or almost one hundred percent.

How to calculate your Google Analytics 360 price?

When considering the purchase of a Google Analytics 360 license, one of the first questions that arise is the price of the license itself. There are various myths circulating on the market, especially since we are currently dealing with a period of changes in analytical tools (migration to Google Analytics 4 and its appropriate configuration), which also affects changes in the price list. Compared to the earlier version (Universal Analytics), the current pricing depends on several factors that I discussed in detail earlier. As a reminder, these are:

1. number of events, 
2. internal analysis teams’ needs,
3. level of extensiveness ecosystem of sites and services

The actual cost of the license is the result of these three elements, with the number of events generated being a key factor.

It is worth noting that using subproperties and roll-up properties, we define categories of services, the main one of which we will call the source service. By creating subproperties (equivalent to views in Universal Analytics), the number of events from the source property is increased by 50% for each individual subproperty. For example, if we have 100 million events in the source service and four subproperties generating 25 million events each, then we have a total of 100 million from the source service and 50% of 100 million from subproperties (4 * 25 million), giving a total of 150 million events. We should be guided by this value when determining the price level.

The case is similar in terms of the dependence of the price on the roll-up properties used. They are especially useful when the company has more than one source service (where each of them corresponds, for example, to individual products, brands or geographical regions). By creating a roll-up property, we will obtain an aggregated view of data from selected source properties. Its (or their) use will also affect our Google Analytics 360 license price – events generated in the roll-up property are defined as 0.5 of the event generated in the services. Using an example analogous to the above: if we have 100 million events in the roll-up property and four source services generating 25 million events each, then in total we have 100 million from the source service and 50% of 100 million from source services (4 * 25 million ), giving a total of 150 million events. We should be guided by this value when determining the price level.

Roll-up property –  – calculating sum to be billed

The price, or rather the level of the price list we enter, will depend on the number of generated events, changing every next million. The more traffic we generate, the lower the increases in price thresholds for each additional million events. The price increase is therefore inversely proportional to the level of generated events (as the number of events increases, the price increase rate decreases).

Based on our experience, we can say that its average monthly cost ranges from 3-4 thousand euros. Of course, this is an average, and the price in specific cases depends on the factors discussed earlier.

What is included in the price of a Google Analytics 360 license?

What value, apart from the tool itself, will we receive in the price of the Google Analytics 360 license?

1. Lack of many restrictions imposed on us by Google Analytics 4, resulting in the possibility of more precise analysis of collected data, and support of a dedicated license manager.
2. At Conversion, we have noticed that companies often have problems with the effective use of the possibilities offered by Google Analytics 360. That is why we have created a GA360 license management process in which the customer, in addition to the license itself, receives:
   – product webinars once a quarter, thanks to which the company is up to date with new tools and cases of their use,
   – expert consultations with a person with at least 4 years of experience in the industry, thanks to which the company is able to obtain solutions to most data-related problems in one meeting,
   – feedback sessions, thanks to which the company has suggestions on how it can better use GA360 data,
   – care of an experienced internet analyst who carries out analyzes commissioned by the client’s business departments or supports the analytical department in providing data.

This is crucial for us, because when purchasing this type of tool, the question often arises about the return on investment. The answer is simple – it will be zero if you do not take action based on the collected data and do not use it to make decisions. In this area, we also provide support from an experienced team.

Summary

To sum up, the question about the price of Google Analytics 360 can only be clearly answered as follows: it depends. It depends mainly on three key factors: the number of generated events, account structure and the use of subproperties and roll-up properties. Additionally, the region in which you want to purchase the license also affects the final price, although it is usually the least variable factor. In our experience, the monthly cost of a license is usually in the range of 3-4 thousand euros. If you have more questions about purchasing a Google Analytics 360 license, contact us and arrange a free consultation with our specialist.

The post How much does Google Analytics 360 cost? first appeared on Conversion.