Why was Google Universal Analytics banned in Denmark?
Before digging deeper into this - Fjord is not in any shape or form lawyers or legal experts, however, even though we hoped it wouldn´t happen we saw this coming and wanted to prepare both us and our customers for a world without the #1 web-analytics tool.
The biggest reason why GA was dropped is this; Personal Identifiable Information (PII) of EU- and EEA-citizens are possibly being transferred to a location where the safety of this information-, and the privacy of aforementioned citizens cannot be guaranteed.
After the infamous SCHREMS-II ruling was passed all the way back in 2020, most people and companies understandably had other things to worry about due to the world wide pandemic. But during this time the data protection framework between the EU and US, also known as Privacy Shield was invalidated by the EU-court. This means that the EU no longer saw the US as a safe harbor for PII belonging to European citizens. Though we feel that we should mention that the successor of Privacy Shield, also known as the Trans-Atlantic Data Privacy Framework was signed between the two parties in 2022, meaning that there still could be a solution in the pipeline.
Google as we all know is a gargantuan American company, and even though they have data-centers located all across the world, including the EU, Google could not sufficiently guarantee that information gathered from websites in the EU would not end up on servers in the US.
But this was just the tip of the iceberg. As many of you might now, the EU has some of the most strict privacy laws in the world. In 2018 the EU General Data Protection Governance policy came into effect affecting the way we as marketers could- or could not use personal data in our marketing. Even before the SCHREMS-II ruling many companies started looking into alternatives to GA.
This subject deserves an entire series of articles in itself, but we will leave it at that for now.
Client-side vs. server-side tracking
Our list consists of several different technologies that work a little differently from each other. During the walkthrough of the platforms you will see us mention client- and server-side tracking.
The major difference between them is how the data is sent from point A to point B.
Client-side
Client side tracking works by leveraging the power of your device, or or more specifically browser to tell a server what you are doing on a website.
Imagine this - you type https://fjord.app in your browser and hit enter on your keyboard. What happens now is that your device is sending a request to our web-server requesting to download our page into your browser. Our server will supply the webpage for you and you open it in your browser. While you browse our site, reading articles or other pages your browser is continuously communicating with our tracking-pixels, telling us a little bit about what you are doing, and bits and pieces of information about not necessarily who you are, but what device you are using, which channel or source you used to visit us and to some extent where in the world your device is.
It is the most straightforward way of thinking about client side tracking. It is your devices, and by extension you that is picked up by our analytics tool.
However; client side tracking has one drawback for marketers. You see if you block any or all cookies in your browser the analytics tool would be more or less blind.
Server-side tracking
In contrast to the former but using the same example - when you now browse https://fjord.app our web server is directly communicating with the server of our analytics tool. This happens even though you block cookies, so now we know that someone definitely has read our blog-articles and converted on some of our forms even though our client-side analytics tool tells us otherwise. However; the server-side tracking does not contain much of the same information used to, for instance, attribute a visit or a conversion to a specific channel. This is the downside.
The best thing about server-side tracking, though, is that you can trust the numbers reported much more than on client-side tracking.
Analytics-tools vs. Customer Data Platforms
Our chosen tools come in two variants - platforms that aid in tracking, transformation- and movement of data to other visualization tools hereafter called CDPs, and visualization tools that also have development kits for tracking called analytics tools.
We highly favor the former mentioned ones, since you can now unify all your tracking client- and server-side and send and receive all this unified to many different tools in your stack. However, this still leaves you needing a tool to actually visualize and analyze the data.
The analytics tools are simple - they all have their own SDK (software development kits) that you implement where you want to track something, and the tool itself receives all of this data in a standardized format- so if having web-analytics data up and running quickly is all that matters look no further.
And here is our list - we will stay as objective as we possibly can to outweigh the pros and cons of all of them.
All of them can be run client- , server-side or a mix of both.
Google Analytics 4 - Analytics Tool
What? Why are we including the updated version of Google Analytics, right after telling you that UA has been banned in Denmark, and that the same thing might happen to GA4 as well?
We want to be as transparent as possible, and we would not forgive ourselves if we did not give you our honest opinion of Google Analytics 4 as a tool before moving on with the list. The most obvious replacement for Universal Analytics, if not for the privacy-issues it might pose, is of course its younger sibling GA4. It is not without reason that Google Analytics in all shapes and forms is the world's most used web-analytics tool.
Pros of Google Analytics 4
- It is free! Sure, it exists in paid versions as well - but for nine-point-nine-out-of-ten the free plan is more than enough.
- GA4 has more granular event-tracking possibilities through custom properties when compared to her big-brother UA, and it is a welcome addition to any business wanting to take a more modern approach to web- and product-analytics.
- Very well integrated with the rest of Google's stack (duuh!), including Google Ads and Google Cloud.
- Since Google has one of the largest databases in the world when it comes to information about you and me, it lets us do more in-depth cohort-analytics on both geographic- and demographic dimensions - something other tools on this list just can´t match (if you don't find a way to gather this information from your users on your webpage (something we are not recommending you to do due to, you guessed it - GDPR).
Let's face it. Google Analytics 4 is a great tool - all things considered. Especially when you account for the price. For us as marketers, CROs and SEOs it is almost imprinted in our DNA, but the truth is that we as a an industry need to take a more responsible stance as to what data we are actually gathering, where we are sending it, who is reading it - and maybe most importantly; what they are using the data for.
Cons of Google Analytics 4
- While some people would argue that this statement might as well have come from a man with a tinfoil-hat - Google is still Google, and the interest Google is looking out for most is Googles own interest. Why are so many companies blindly trusting the attribution reporting from Google themselves? After all; if Google can convince you that their channels are driving most of your revenue, which channels do you think you would invest the most in? This is problematic when they own both the platform to spend money, and the platform that reports the results on it. In fact; in one of our earliest tests of the Fjord attribution model up against Google's last-click non-direct model we noticed a difference in between 10-15% over-reporting in GA vs. Fjord when it came to Google Ads performance on different KPIs. This was for a brand that in its nature should have short time-to-buys, few interactions and relatively low-threshold products - things that should point to a higher percentage of conversions given to traditional last-click channels. The difference here was that we pointed to the other last-click channels as having more impact in 10-15% of the cases. For the sake of transparency - this could have been a fluke and we are looking forward to further testing.
- The privacy issues, as mentioned at the top of this article.
- Not natively letting you integrate data from other channels into its reporting and calculations. While it is possible to feed GA with Facebook spending and interaction data, the platform isn't built for this.
Segment - Customer Data Platform
Segment.com is the market-leader of easy to use CDPs, helping tech-teams, business owners and - developers, security experts and marketers get the absolute best out of their data. In the case of companies not wanting to use our on-premise CDP sending data to the Fjord-Platform we usually end up suggesting Segment as an alternative. The reasons for this are simple, and are best summarized in our pros-list of Segment.
As Segment is not an analytics-tool you still need a visualization tool if you want to analyze the data it captures.
Pros of Segment
- They have a high number of integrations - over 200 of them. So you can send and receive lots of valuable data from practically all your existing tools in your stack.
- It is scalable. Their free-plan should be enough for any business just starting up, while their paid entry point is not more than 120$.
- Customers in all industries and of all sizes and compositions - meaning you probably won't be lacking in functionality no matter who you are.
- Helps you unify both server-side and client-side data.
- Built in consent-management that aids in GDPR-compliance.
- Even non-developers can easily learn how the architecture and client-side SDK works, and maintenance of a tracking-plan or your internal data governance can now involve more than a few stakeholders, if needed.
Cons of Segment
- Even though the client-side of it is easily maintained, we continually see that development of server-side events is usually under-prioritized by the development team - especially in SaaS where they usually have more important tasks to prioritize than tracking. And if all you are doing in the long run is tracking client-side events and sending them to an analytics-platform and a few marketing tools - is the increase in price when you grow really worth it? That is not up to us to decide.
- The price can become high, depending on who you are and as we mentioned in the previous point how much data you are actually shuffling through Segment to impact the cost-benefit. It starts at around 120$ and includes 10.000 MTUs (monthly tracked users) - every 1.000 users after that costs 12$ more until you reach certain levels where the price per 1000 MTUs gradually drops to 11$ and then 10$.
- Some of the coolest functions in Segment are addons that in some of our customers' experiences are pretty expensive. In particular we are talking about Personas-addon, which lets you build cohorts of interesting users in your echo-system.
- And to a point that for some could be very important, while for others might not matter at all. Segment has a replay-functionality which can be problematic in the EU. It is stated that Segment does not store your data. They simply receive it, and send it downstream to your chosen destinations. However; when the privacy-shield deal between the US and EU became void due to the Schrems II law being passed it was no longer allowed to send data of EU-citizens to the US. This point means one of two things for Segment. First off; Segment is an american company, and even if the data was sent from your website, CRM or application to Segment and was actually deleted the second it arrived it still does not change the fact that data on citizens of the EU are being sent to a cloud-service that might wholly or partially exist outside of the EU could still be problematic. The second point is the event replay - if the data is not stored by Segment, this means that there should be no way of replaying historical event- and user level data at all. So where is this data stored? According to Segments own webpage - if you are on the Teams-plan, this data is not guaranteed to stay wholly within the EU. You need to upgrade to the Business-tier for this. Just something to keep in mind.
Matomo - Analytics Tool
Marketing themselves as the “Google Analytics alternative that protects your data and your customers' privacy”, Matomo is a great alternative that is used by organizations like the United Nations and NASA. Just those two names should tell us something. In addition to letting you run Matomo in their own cloud-hosted environment, they also let you set everything up on-premise. This means that you can host your own instance of Matomo on any server you want, further increasing the data-protection for your visitors (of course dependant on how you protect it).
Pros of Matomo Analytics
- Both the cloud- or self-hosted versions are without any data-samling whatsoever.
- No matter which version you choose, both are compliant with EU privacy regulations, of course as long as your self-hosted version of
Matomo is hosted in an environment that follows these regulations in the first place. The cloud-version is running in a EU-cloud.
- For someone that is used to Universal Analytics the learning-curve to get up to speed in Matomo is not that steep, sans some of the personas-functionality mentioned under Google Analytics 4.
- Someone once told us that “almost whatever I could do in Universal Analytics, I can now do in Matomo”. So there you have it.
So Matomo is fixing a lot of the problems that made Denmark ban Universal Analytics, while still retaining a lot of the same functionality.
Cons of Matomo Analytics
- While Matomo does a great job of replacing Universal Analytics, it offers little new when it comes to the more advanced functionality in Google Analytics 4. So one could argue that Matomo is not a true replacement for GA4 at all, only the older UA. On the other hand; for many businesses this might be more than enough.
- No data-driven attribution, unfortunately. Even though Matomo advertises that they offer “Multi Channel Conversion Attribution”, this only contains the same rule-based attribution models that you are used to from universal analytics: linear, Last Click, Last Click non-direct, first-click, position based and time decay, though the names might vary.
Link - https://matomo.org/faq/multi-channel-conversion-attribution/faq_25500/
Rudderstack - Customer Data Platform
The main competitor to Segment.com - at least in our eyes. Rudderstack started as an open-source initiative that over time has transitioned to be more commercialized. If you are using Segment but want to transition to another CDP, migrating to Rudderstack is easy. The pricing model is also different from Segment, where instead of being billed for the amount of “monthly tracked users” you get billed by number of events. The free plan starts at 5.000.000 events monthly. The main difference in price here is the lowest paid tier. Segment Teams-plan starts at around 120$, and Rudderstacks Pro-plan starts at 750$. Oh, just like Matomo-analytics Rudderstack also lets you run either in-cloud or on premise.
As Rudderstack is not an analytics-tool you still need a visualization tool if you want to analyze the data it captures.
Pros of Rudderstack
- Flexibility of choice in where to host your instance of Rudderstack. Choose between cloud or self-hosting.
- Easy tracking setup, for those used with Segment.com
- Roughly the same amount of integrations, compared to Segment (180+).
- The possibility of using the open-source variant in your own environment simplifies tailoring and custom solutions for more experienced teams.
Cons of Rudderstack
- The same top point made about Segment also applies here - to unlock the full capability of a tool like this you should implement it in your entire stack to maximize the amount of first party data you can leverage. In our experience, this type of tracking-setup will take valuable time away from your dev-team, which might need to prioritize other projects.
- This could as easily be a pro depending on who you ask, but since you are paying for the number of events fired rather than tracked users the price could become rather high under some circumstances. Especially in ecosystems where you want to track a lot of pages and have many events fired both on your webpage, in-app and server-side. However; we think that the right usage of a CDP has a great ROI and in the right circumstances the price is almost always worth it.
Mixpanel / Amplitude - Analytics Tool
As previously mentioned none of our top picks for CDPs lets you analyze the data you capture in-tool. Both of them depend on you picking up this data at one location, sending it downstream to another tool where you can visualize it. All of the analytics-tools in this list can connect to both Segment and Rudderstack (except for Matomo that right now only works with Segment).
We bundled Mixpanel and Amplitude in one section, since our experience is that the two are rather similar in functionality and usage. The main difference is that Amplitude has more customization than Mixpanel, for better or worse, but in return the entrypoint is a little steeper price-wise.
Both platforms are marketed as a so called product-analytics tool, and we agree that this is what fits the best as. However; some companies that now migrate away from Google Analytics use Mixpanel / Amplitude instead. So with the right setup, you could use these product analytics-tools as web analytics-tools as well.
Pros of Mixpanel / Amplitude
- Integrates very well with the CDPs on our list. If you have a web page and one or more applications these tools let you visualize and analyze data from both ends.
- As mentioned before; they are very customizable while we favor Amplitude a little more in this area.
- Aids in future proofing analytics. Even though you have created no reports on a set of data, you can in the future as long as the data is being captured. What this means is that you can use a CDP, or their own SDKs to capture as much interesting data as you could possibly want, and then step-by-step build reports on them.
Cons of Mixpanel / Amplitude
- As with most tools that are not Google Analytics, if you are auto-tagging your ads (specifically those that are Google in this scenario), the analytics tool won't natively understand which campaign, ad-set or ad that drove this conversion or visit. It remembers the Google click-ID, hence letting you understand it is Google Ads, but without tweaking it is hard to go further than that.
- Even though you can create really in-depth reports, it can be slightly overwhelming for some users. With practice everyone can learn to find- and set up any in depth analysis they wish, but some of our clients think that the learning curve for these types of tools is steep.
- These platforms are not attribution systems. They give you the choice to use simpler rules-based attribution models such as linear, last-click or first-click. There is no innate datadriven-attribution algorithms within this type of tool.
Fjord.app - Customer Journey Analytics
Fjord is an analytics-platform with data driven attribution at its core. What does Customer Journey Analytics mean? We look at the entire journey from start to finish. No more data-silos between marketing and sales , for instance. We follow the conversion as far as you want to, from the website events and all the way to a closed deal stage in your CRM, feeding this data back to not only the sources or campaigns that contributed the most - but all the way down to the ad- or keyword level. However this is not a tool just for sales-organizations; we are agnostic to what industries we help attribute marketing contribution, where we can tailor your Fjord-instance to your specific KPIs and business model.
In addition our algorithm also runs on-page, meaning that we can attribute contribution-values to your blog-articles, product-categories, FAQs, micro-conversions etc.
We started out as a supplement to other analytics tools, but soon realized that we were sitting on (almost) all the data that Google Analytics did, except for the demographics (which is the most problematic feature in GA, according to EU-privacy laws. For every week that passes we add more, and more features to let our customers replace GA - step by step, while still remaining a neutral part as to what marketing actually contributed the most.
Pros of Fjord.app
- Datadriven attribution is our forte. We don't want you to have to analyze a chunk of rules-driven attribution models by yourself. They wont even tell the full picture even if you had the time.
- Our attribution is three-dimensional. Looking at dynamic attribution windows, number of interactions and placement in the chain of trends to attribute value, whether it be revenue, net-revenue or binary conversion-numbers, we help you find the touch-points that contribute to each of your KPIs.
- Your data is your own. And we will host your data-layer well within your national-, continental- or constitutional privacy laws. As long as you are privacy-compliant - so are we. We won´t need any PII, and we don't want any.
- Fjord is neutral. You don´t buy any placements from our platform, and we won't favorize anyone. We are not getting paid by any third party wanting you to spend more money to enrich their business. The story we tell is on your behalf.
- We let you attribute actual net-revenue, or actual closed deals. Not just web-site leads.
- We will use our algorithms to not only attribute value to your channels, campaigns, ad-groups, ads and keywords - we will also help you attribute
We love all the other tools we have mentioned in this article. Either as users, resellers or consultants to implement them. We are humbled by the experience we have had with all of these technologies, and the brands - they are simply great. Fjord is the culmination of all our experiences with attribution and analytics, and we believe that we have been able to take pieces of the experiences that made the other technologies great to inspire our own platform.
Even though we love our tool, in the name of transparency and honesty; we also have flaws and cons.
Cons of Fjord.app
- Datadriven attribution is not for everyone. Even though our layers of algorithms will help you get value on lower volumes of data, as well as higher, not all products or brands need this.
- To get the most out of the “customer journey” part of our platform, you need good tracking. We will of course help you set it up - but in this instance it is not just plug and play.
- We don't have as many integrations as the other platforms - we are new, and think we have covered the most important ones. But still. We are always adding new integrations to our platform.
- We don't have as many features as many of the others - yet. As previously mentioned, we wanted to live beside many of these tools, but have realized that our customers want us to help them replace them. We are launching new features weekly, and want to meet that demand.
