In the first of a regular interview series with our very own ‘Leafers’, Will Chalk, Head of Google Performance & Customer Analytics, discusses his experiences with Google Analytics 4 (GA4) and how eCommerce companies can get the best out of Google’s next-generation analytics platform.
Hi Will! Let’s jump straight in. What’s your general take on GA4 when it comes to dealing with third-party cookie loss?
Hi! Well, GA4 really isn’t the solution. The way cookies work is simple: if the user doesn’t accept a cookie on your website, you can’t track that user. GA4 will not solve that.
But it will try to give you as much information as possible, so you can track users across pages and attribute a channel back to them.
It all still feels very up in the air, to be honest: there’s no definitive answer.
But the real benefit of moving to GA4 is that you’re adopting the platform early. So you can gather as much information on your traffic while you still have access to third-party data.
Do you get a sense of how many eCommerce companies are moving over to GA4?
I think it’s been a really slow process for companies to adopt GA4. Everyone’s got their own methodology and they’ve been using GA3 for the past 10 years.
Moving to a new reporting suite is such a big job. It needs tech support, it needs marketeers who understand the user journey, attribution and channel tagging.
Unfortunately the industry skill set just isn’t there.
There are quite a few companies that don’t even have an internal team to do the most basic admin jobs. Being able to pull data from GA and put it in a spreadsheet, for instance.
So I think the adoption rate is going to be really slow for GA4, and it opens up a massive potential for companies like ours – which has a tech resource to be able to implement these changes – because it’s hard to find people with the right skill sets.
Demand is huge. The UK just doesn’t have a big pool of marketers who know what they’re doing with this stuff, unfortunately. We see it all the time at Leaf.
How have you found working in GA4?
So I know GA3 like the back of my hand, and it’s still taken me a while to get used to the way GA4 measures and tracks users. There are going to be a lot of mistakes made in GA4 implementation across the industry, because it’s such a different process.
The interface is so customisable, to the point where it’s hard to know where to start.
It’s like having a box of Lego, and someone pours the entire box out onto the table and tells you, “OK, go and build this model from scratch”. You’ve got all the pieces there but it’s hard to know what to do first.
Obviously having the App + Web information under one reporting suite is a massive advantage. And it’s a lot more flexible now: you can track a hell of a lot more user interactions on the site.
You can track anything depending on how a data layer is set up.
Data layers are pretty important, right?
Yeah. A data layer will track any interaction on your site. So if you click a page view, that data layer will contain everything there is about that page view: what page they clicked on; what the previous page was; what the category was; what the browser was; what the device was, etc.
You can capture all kinds of information with the data layers.
This is going to be so important going forward, especially on GA4, because that’s how their events now work.
But third-party cookies are going to be a bit of a black box for a while; there’s no right or wrong answer to this. Google says GA4 can fill in the gaps in data using algorithms and AI.
Who knows what that actually means in practice?
Obviously the more data those algorithms have, the more accurate they’ll be. So most clients need to jump on GA4 now, because GA3 is going to be sunsetted – probably in the next 2-3 years.
It’s a big job to move companies over. It’s tech intensive and not just about putting a bit of Javascript code on the website anymore.
You’ve really got to map out the user journey, create a data layer with your developer and also input what information you want to track from that user.
We’re working with several of our clients on building that out right now.
And what about tracking and attribution?
Tracking interactions on your website shouldn’t be a problem, but it’s harder to track where that user actually came from. That’s going to be the black box that will really influence investment. We’re seeing it in Facebook right now with the iOS updates – we’re just getting less feedback from iOS devices.
On average we’ve experienced a 40% decrease in attributed sales back to Facebook.
So everyone’s going to feel the impact of that once Chrome decapitates cookies.
I don’t think it will hit Google as hard because Google has its own data sources. Everyone uses YouTube, Chrome and Google Search. There’ll still be a lot of accessible information, and Google is going to be able to supply that.
Facebook, however, relies heavily on third-party cookies – that’s their business model. I think they’ll try to keep the user within the Facebook ecosystem. Maybe they’ll develop a shopfront search engine so they can attribute back to ads and campaigns.
There are rumours about Apple developing a search engine. It’s obviously all going to get even more complex as we move further along.
Do you have any best practice advice for eCommerce brands adopting GA4?
For sure. The first thing I always ask clients is: what do you want to track? Then I map out what that actually looks like and develop a data layer solution – and a tagging solution – around it.
That way we know how we’re going to implement the tracking from a tech perspective – through Google Tag Manager, for instance.
And the second question I always ask is: how do you want your data to be visualised?
In GA4 you can build customer reports if you want to, but it can very quickly get confusing to the user, and some of the reports aren’t really useful for analysis. For clients’ board members – or even for an eCommerce manager – they generally want everything condensed into an instantly actionable format.
I always recommend either channelling the GA reporting into a Microsoft Power BI tool or through Data Studio. That’s what I’ve been doing with clients over the past few weeks.
At Leaf we condense all the information – all the data streams – into bespoke Actionable Insight reports so that someone at board level or management level can come in and have an overview of the most relevant, insightful metrics.
We also build bespoke interfaces for our clients to help them visualise the data. It’s helped us to have real-time data insights so we can deliver quick results and make changes and improvements to live campaigns.
What’s next?
It’s important to take a step back and develop a user journey plan, and identify exactly what you want to track and how to visualise it.
A lot of companies overcomplicate things quite quickly. They want to track everything, but at the end of the day they don’t look at 90% of the metrics they’ve created.
It should always come back to where the money and the traffic comes from. What’s driving what, essentially.
It does also depend on how advanced the company is and what resources they’ve got in terms of tech, UX and CRO. If companies have got those in-housed then GA4 can support all of those teams; but if they don’t, it does become a lot more complex.
Developing a plan that supports each team will become really important if companies want to get the best use out of GA4.
You can go down the rabbit hole with any analytics platform, but if you can identify what metrics will make a real difference to your conversion rates and to driving more revenue, you can get really good use out of GA4 and integrate it quite quickly.
But that’s in a perfect world, and not many companies have the resources to deliver that level of planning and roll-out.
What’s Leaf’s advice on prioritising GA3 over GA4?
At the moment our best practice recommendation is to run both GA3 and GA4 side by side.
We take all our learnings from GA3, extract what we want to track outside of GA3 and transfer it over to GA4.
But the way GA3 tracks users is not the same as GA4, so companies will have to make some adjustments to their website. For instance if you make a change in GA4 it can knock off some of the GA3 reports.
That’s partly why some companies have been hesitant to move over.
I think that hesitancy is also there mainly because the knowledge around GA4 is just not strong, and a lot of the features are still pretty buggy.
You can’t do attribution modelling as well as you can in GA3. You lose elements of the attribution because GA4 only registers first-click and last-click.
GA4 is pretty bog standard at the minute, to be honest.
Have you seen any discrepancies from running GA3 and GA4 in tandem?
For sure, especially in traffic. I like to think G4 is a bit more accurate when it comes to traffic, because it can tie multiple client IDs back to a user quite easily – whereas GA3 has multiple sessions per user.
Eventually GA3 will stop working, perhaps in a few years’ time. So running both side-by-side is important.
Even if you’re just tracking the basics in GA4 – user count, orders and revenue – then as long as you gather 6-12 months of useful data, I think that’s a good start.
But we’ve seen discrepancies of about 20% in terms of traffic levels between the data from each platform. GA4 is just better at matching people. Page views can get inflated quite quickly as well, and user counts too.
It just depends how users install GA3. Loads of companies have never installed it correctly. I still see that now.
Clients might have multiple domains, and where GA3 is concerned if a user moves between them to complete checkout, it creates two sessions and two unique users.
I can’t tell you the number of companies I’ve seen who don’t have cross-domain tracking. It’s really common to see that installed wrongly on the domain level, the sub-domain level and the sub-directory level as well.
At larger companies it’s also common not to have filters for internal IPs. They’ll have staff viewing the site every day and generating hundreds of sessions.
You can imagine how quickly traffic can get inflated and diluted in that situation. It’s so important to exclude internal IPs and have a uniformed UTM structure and filters that formalise and standardise UTM structures.
Hardly anyone uses that.
Are SMEs better positioned than larger companies to make the move to GA4?
With support, yes. If we recommend to an SME client that they go ahead and adopt GA4, they can usually give us the sign-off pretty quickly and we start the process. At larger companies it’s like turning a huge ship, because for them it would have a huge impact if they just switched overnight.
They’re spending millions of pounds each month on marketing and have loads of internal teams managing segments of campaigns. So a bigger company’s migration to GA4 will probably take a year or so, whereas SMEs are usually more interested in knowing where the money comes from. For them we can make those changes much more quickly.
And if you add in the issues surrounding Android and app tracking, well that’s a whole other kettle of fish for businesses of all sizes.
Tracking in apps is even harder.
It’s another skill set that’s becoming hugely competitive. App development has gone crazy right now: the average job retention rate for an iOS developer in London is about four months. They get poached by clients with more resources and bigger budgets who are willing just to spend the money on getting the app built.
That’s a big issue: staffing, and finding people with the right skills.
Developers are asking for a premium so it’s harder to keep everything under one roof. A lot of companies will start outsourcing all this stuff again, whereas a few years ago it was the other way around and they were taking it all in-house.
Everything’s been flipped on its head because the tech side has become so much more complex.
You see all these boutique agencies and freelancers being phased out now, because they don’t have the skill set to support the client through the iOS 14 change, the GA4 change, etc. Internal teams can also become lazy for sure, because they’re looking at the same data and the same campaigns day in, day out. They get less sharp and less attuned to the changes going on around them.
Do you have any other recommendations on adapting to a cookieless future?
Collect as much information as you can on the back end.
Email is the only channel that isn’t going to be affected, so make sure your mailing lists are up to date and your data warehousing infrastructure is sound. Basically anything that isn’t breaking GDPR and helps you understand the end user.
Your first-party data needs to be up-to-date and tracking correctly. I recommend getting it all into a visualisation tool – either Data Studio or Power BI.
Email is going to be a core channel for a lot of businesses going forward, so being able to send targeted campaigns out to that user base will be key.
Look at lifetime value and customer acquisition cost, so you’ve got a good understanding of what channel is being profitable for you while you have access to cookies. That will give you confidence in spending the money until the tracking gets corrected or an alternative comes out to help with conversion tracking.
It must be pretty hard to plan for a lot of this stuff though? So much is changing.
It’s so hard because, really, nobody knows. You see all these articles on LinkedIn saying ‘do this, do that’, and in the end it’s just their opinion.
We don’t know yet what will work.
Businesses should find multiple opinions from different sources and then come up with their own planning that works for their specific company.
And that’s always our approach with clients, because we’re learning at the same time. It’s about saying, ‘come with us, we’ll go on the journey together’.
And we’ll work it out together, too.
It might sound cheesy to say it, but we always try to stay on the bleeding edge of this kind of stuff. If a new feature does come out, we’re among the first to adopt it and get it integrated because we’ve got the full service here. We can do the performance marketing but we’ve also got the tech expertise.
If a client asks us overnight to implement server-side tracking and update their GA codes and pixels, we can.
We’re flexible that way.
There aren’t many agencies that have tech and performance built together. Some of the bigger agencies do, of course, but usually when we audit their clients, what we see is absolutely shocking.
And these are agencies considered to be the best performance marketeers in the UK. Half of them talk a load of shit. There’s no actual data analysis, and nobody with real-world experience looking after the data accounts.
Leaf can afford to build an experienced team and keep it small but dedicated.
Big agencies now can’t run or manage big media spends – because they don’t have the team to be able to support and optimise them. Smaller clients are pitched to by the ‘big dog’ account managers and then get passed on to the juniors as soon as they sign.
We’re going to see bigger companies breaking up into smaller agencies, for sure.
How do you see the immediate future?
Firstly there will be a lot more outsourcing, especially in the next 12 months as cookies (and possibly GA3) get phased out altogether.
And GA4 will get even more complicated as Google looks to integrate the whole attribution model (not just first- and last-click as it is now) into the platform.
GA4 is a development tool in beta. So we just have to roll with it and keep up with Google on this one. Same with FLoC: we keep up to date and roll with the changes.
It’s exactly how we’ve managed the introduction of iOS 14.5 – we just take things day-by-day, see what’s different and come up with ways to ride out the changes. That might be through implementing different campaign structures or integrating the GA data with Facebook data, for example.
What’s crucial is making the client aware of it all.
Because attribution is another subject altogether: not many clients understand the process of attribution, and not many agencies talk about it with their clients.
But now it’s a talking point because of the changes to the attribution window in iOS 14.5.
A lot of these changes are going to kill off a fair few analytics agencies that just do a bog-standard job. The ones who add a bit of code on the client’s website and set up basic tracking.
It’s going to hurt them because everything has become so much more complex.
And the developments surrounding server-side tracking too: that’s something we’re doing a lot of work on right now. That’s totally new and may help with cookieless tracking, because if the cookie isn’t getting sent to Google but is instead sent to your server and then to Google: that may get around the whole first-party restrictions on cookies.
But it’s all up in the air, and the last 18 months have turned the whole world of digital upside down.
It’s going to get a lot more bumpy over the next few years, especially as Chrome phases out cookies altogether and as GA4 gets pushed onto more people.
I’d love to give you a definitive answer, but who the hell knows where we’ll be next year?
We’re all learning together though, and some of us are better positioned than others to keep delivering results – despite the twists and turns.
Want to scale your eCommerce revenue profitably? Learn more about Leaf’s end-to-end performance marketing services here.