Running Adobe Analytics and Google Analytics in parallel

Why would you run both Adobe Analytics and Google Analytics on your website? Won’t you end up with two parallel running tools reporting different figures?

A frequently asked question when it comes to Adobe Analytics and Google Analytics is, why would you run the two different analytics tools in parallel? Let’s take a look at some of the differences, the strengths and weaknesses, and the target audience for each tool.

If you have a team of analysts and a dedicated resource to throw at it, Adobe Analytics (AA) is great because it’s an analyst’s tool. You can dig into levels of granularity and set your specific level of customisation more so than with Google Analytics (GA). Most people cut their teeth with GA though because it’s free and it gives you exactly what you want.

Google Analytics is an entry-level tool for people who want to come in, see reports, and get an easy view of all of the metrics that they’re used to. Fundamentally though, both tools are web analytics platforms tracking conversions, page views, etc. just with slightly different approaches.

Analysts vs. marketers

A tool like GA sits within their marketing product suite (the Google Marketing Platform) as a marketer’s product, while AA is more of an analyst’s product. Google Analytics is very accessible and everybody who works either as an analyst or as a marketer has had some experience with GA. Marketing teams usually push for GA because most of the tools that they use are in the Google Marketing Platform, like Google Ads, Display & Video 360 and Search Ads 360.

Google Analytics is a no-brainer for the ease of connection and the level of ability in terms of things like attribution. GA is easy to plug and play, and Google favours their own product connections, whereas Adobe requires extra legwork.

In AA, you can do all the drill-downs and visualisations in one tool, which suits analysts who have the time to dig into the data. It’s not so much about what’s missing in Adobe as a lack of understanding about how to set up the connections. There are a few extra hoops to jump through.

Why are these figures different?

The problem with working with both GA and AA is that you might end up with two different versions of what should be the same number – two sources of truth in a sense. So, when you’ve got a tool that’s free and an enterprise-level tool, and two figures that say something different – which one is right?

It’s very difficult to have two tools that are set up in exactly the same way, saying the same thing. That would be the ideal, but it’s never going to happen. Whether it’s Adobe, Google, or any other platform, you’re always going to have slight nuances or variations.

Google is so dominant in the industry that people just assume it’s right, but human error does come into play – GA can be set up in a way to be wrong. Data very rarely lies, but it can tell a different story depending on the implementation and the specific metrics you’re using. There could be an underlying tracking issue, and only after an audit do you realise that you’ve been reporting on the wrong metrics.

The Google halo effect

This is the idea that Google is always right, or that it can’t be wrong. The perception is often, because it’s free, why not run it alongside another tool? You might be validating the session numbers across various different analytics product, but it’s hard to entertain the idea that GA might be wrong when the numbers don’t match.

At the end of the day, GA is a product owned by Google, who have a vested interest in making media look performative. Things like the Campaign Timeout in Universal Analytics have helped them do so. We have to be cognisant of the fact that Google (like any other ad platform) will want their own products to look good. Whether you go into Facebook Ads or Google Ads, they pick attribution models that favour their own ads – as you would expect.

Mirroring like-for-like

Web analytics tools are not accounting tools. Analytics tools are more for trend observation, and for observing those trends over time and seeing how your changes affect things. The numbers are probably going to be wrong (because of so many different factors), but it’s more about being able to spot the trends.

So, can these two tools run happily in parallel? You can run them both in tandem with strong governance in place and if the connections are set up properly. You can try to ensure – as much as possible – that the data that’s being sent to both platforms is consistent, and in the same format. Then it’s purely processing on the tool side that would change any data between them. The tricky part comes from ensuring that you’ve got the same data in different places.

However mirroring like-for-like is not necessarily the right thing to do, as a lot of work has to go into that to start with and it has to be maintained at the same level of integrity. If you don’t have that strong level of governance in place, and you aren’t clear about what you want to get out of each tool, then you are going to end up with two parallel running tools with different figures and applications.

When a use case comes up, there has to be a real discussion around that before enriching the dataset with that extra conversion or interaction point. If that’s something you want to do, discuss why and ask how it’s going to change your implementation. How is this going to change the analysis and the data that you’ve got?

Do you buy or build?

Purely from an analytics or data perspective, it’s also worth asking how far to go with a software-as-a-service product before you take that spend and build your own stack. In some verticals, you may have a very specific nuance for how you want the data to fit. To get off the ground quickly, tools like Adobe and GA get you running, getting reports out, and meeting business needs. 

Google’s data schema is pretty rigid – it’s great for entry-level analytics as there’s less to get wrong. Adobe is a stepping stone beyond that, more of an analyst-friendly product that enables you to do advanced data manipulations and customisations. When you’re trying to bend the restrictions that are put in place by analytics platforms though, at this point, you possible need to start looking at building versus buying.

The build option is the next level when a flexible tool becomes inflexible, so the next logical step is to build. It’s a maturity progression. If you’re not bending your current solution, then you probably don’t need the next level up because you’re not making full use of what’s already available. Each time you start bending to the point of breaking, you’ve got justification to move to that next level.

Adobe Analytics and Google Analytics

It’s not really a case of GA versus AA, or GA versus anything else to be honest. It’s more about understanding what the needs of the business are and then picking a tool, or multiple tools, that are going to meet those requirements. Whether you’re running with one analytics tool or multiple in parallel, governance is hugely important. Everyone using the data has to understand the differences between the different data sources, and the know-how to play to the strengths of each one.

If you’re interested in finding out more about why you’d have both Google Analytics and Adobe Analytics on your website, check out episode 30 of The Measure Pod.

Written by

Nick is a Senior Analytics Engineer at Measurelab. You'll usually find him poking around in GTM, rummaging through your site's source code or looking for new ways to use the Tag Manager API. Elsewhere, he collects obscure punk records and more film cameras than is reasonable.

Subscribe to our newsletter: