Attribution in Google Analytics: why we need it and how to understand it

Read what attribution is in digital marketing, why we need to know the impact of all digital channels on the success of our web, and what it has to do with football.

In my previous article on attribution Lesson on attribution: Why did you have web visits from Facebook that “dropped” on Monday? we could learn that reports following “Last Non-direct Click” (the attribution model offered by Universal Analytics) or “Cross-channel last click” (the model in GA4, which also ignores the last direct visits) will always favor the last time visitors arrived at our website, and ignore other arrivals and their channels. Because of this, marketers in performance marketing often wrongly focus only on those channels that “bring them conversions”, and ignore other ways of visitors arrival.

If you want to learn why we need to know the impact of the digital channels mix on the performance of your web, sit back, relax and start reading.

Not only Google Analytics, but also other analytical tools, such as Facebook or Google Ads, use their pixels to monitor the success of their traffic, but they mostly attribute conversions to the last visits to your website from “their” channel. For example, we can have a complex situation like this with 5 different visits to the web store:

Display -> Social -> Google Search -> Corporate web (Referral) -> Direct = Purchase

What will we get in the reports?

  • Google Analytics will say that the corporate website is responsible for the purchase (Referral)
  • Google Ads will say that Google Search is responsible for the purchase
  • Facebook will say that the second visit was responsible for the purchase

Therefore, we need to know well the attribution models of individual tools, so that we can compare their results, and so that we don’t think that “someone is always lying.”

However, in Universal Analytics it is possible to analyze the so-called “auxiliary” or “assisting” channels that lead to conversions. At the same time, Google Universal Analytics offers us several attribution models that can be used for analysis:

Google Analytics 4 offers similar models, which are called something different and unlike Universal Analytics, they can be set to apply to all Google Analytics 4 reports. These models are as follows:

  • Cross-Channel last click
  • Cross-Channel first click
  • Ads-prefered last click
  • Cross-Channel linear
  • Time Decay
  • Cross-Channel position-based
  • Data-Driven

In addition to the “last-click” model, marketers often like to look at “first-click” reports. Especially classic marketers who are primarily interested in creating brand awareness, not so much conversion. In such campaigns, it is important which channel first brought the user to the web, so “first-click” models or models offered by pixel-based platforms such as Facebook or even Google Ads are suitable for such campaigns.

“Data-driven” or a model based on your specific data, is part of the paid version of Google Analytics, for which you still have to allocate a considerable monthly sum of money. However, if you are investing millions in marketing annually, you might want to consider the paid version of Google Analytics (also known as Google Analytics 360) because it can practically give you live data on how you need to change your marketing strategy and tactics to be more successful.
The free version of Google Analytics gives you such information after the campaign is over, so that you know what to do next time.

If you are developing a digital marketing strategy through multiple channels, and what you should actually do regardless of whether you are doing performance marketing or just creating brand awareness, you can use some of the other models (Linear, Time Decay, Position Based) or create your own model, depending on what strategy you have in mind. In fact, all of the above models can serve you to evaluate your strategy and help you decide through which marketing channel you will create brand awareness, through which channels you will maintain that awareness, through which channels you will lead users to conversion, and through which channels you will conclude that conversion. At the same time, each industry or type of campaign may have its own rules and seasonality or recommended channels, but also each agency or team may have its own channels in which it does well, with which it will achieve better results than the channels in which it is not the best.

Why do we have these complicated models and why all this“, you ask? Well, primarily in order to responsibly manage the marketing budgets we have at our disposal and to achieve better results with less money. Each campaign can have its own strategy, and we should not focus only on one arrival channel or one attribution model, because this can mean that we miss opportunities for better results. How can we know that? Well, simply – by looking at the data provided by Google analytics and thinking about your strategy.

Football attribution

I often compare the marketing strategy to football, and the marketing manager to the coach of a team, which can be football, but can also be handball, basketball or any team. For example we all know that Leo Messi and Christiano Ronaldo are “goal machines”. What does this mean in marketing terms?
Well, a goal in football is a “Goal”, and that’s what conversion is called in analytics, which means that Messi and Ronaldo give the most conversions in football. More precisely, their foot is usually the one whose kick ends up in the opponent’s net (Goal). If we look at the “last non-direct” model of attribution (where “Direct” can be in the case when the ball bounces off the goalkeeper into the net), these two are most often responsible for conversions. This makes them expensive and everyone would want them on their team.

2018 was the year when another footballer dominated the world’s turf – our own Luka Modrić. Regardless of the fact that he did not score the most goals that year, he won many polls and was even named the best player of the World Cup in Russia. Why?
I’m an analytical geek and I like to believe that it’s because according to the Time Decay model he was the most responsible for conversions :)))

Let’s ask ourselves the following: if the ball started from the goalkeeper, who threw it to the defense player, the latter passed it to the midfielder who spotted Ronaldo and threw him a perfect ball that Ronaldo “just had to push” into the net, who is responsible for the goal?

  • Is Ronaldo, as it is written in the statistics?
  • Is it the midfielder (Modrić), who is officially credited with an “assistance”?
  • Is he a goalkeeper?
  • Who is more and who is less deserving of the goal?
  • Would Ronaldo have scored at all if the goalkeeper had not started that action?

These are all questions that can be asked in the marketing strategy.
Is Facebook responsible for the conversion or is it the Newsletter?
Do we even need a newsletter?

But let’s also imagine the following: we have all the same players at our disposal, but it is:

  • Ronaldo (striker) in goal,
  • the goalkeeper is on the attack,
  • the midfielder is in defense,
  • defensive player in the middle.

What do you think, would such an organized team score more or less goals in a match than one in which the goalkeeper is in goal and the striker is in front of the opponent’s goal?

I believe we can agree that such a poorly organized team would have some results, but certainly not as good as a team that has a goalkeeper in goal and Ronaldo in attack. But be careful, in both cases we have the same team at our disposal, we just arranged the players differently on the field and the results are already different!

I give such examples when clients or colleagues ask me: “And which channel should we use” or “Which channel is the best”? Best for what? First we need to find which channels are good to start the attack, to figure out which channels to use to score goals, and which to spread around the rest of the pitch. We can do all this with the help of data provided by Google Analytics. If we incorrectly distribute the players on the field, or if we incorrectly distribute the channels available to us, the result we get may not be satisfactory.

If you have kids who go to elementary school, you’ve probably been to one of their soccer practices or games. Have you noticed the main characteristic of their game? Regardless of where they are on the field, EVERYONE (except the goalkeeper maybe) runs for the ball! :)))

OK, we know that little Perica runs the fastest and often scores a goal, but everyone on the team will still run and dunk the ball as soon as they get it. Only when the kids mature and grow up a bit, we can develop a strategy with them, according to which some will be in charge of stopping the opposing players, some of them to pass to the attackers, some of them to score goals and the like.

Unfortunately, most digital marketers use a strategy similar to first-year students, or rather the absence of a strategy:

“Let’s take all the available channels and try to shoot at the goal with all the channels and we will definitely score somehow”.

OK, if on the other hand we have similar competition, maybe we will have some memorable result, but imagine what would happen if that elementary school team played against high school graduates? High school graduates, who already have worked out tactics and know which of the players should stand in which position so that they are all effective together? I believe that the first-year kids would only accidentally score a goal.

Most of the strategies in digital marketing, not only in our country but also in the world, look like kids playing football in the first grade.

Everyone is shooting at the goal and everyone is running in the same direction. Successful campaigns and successful teams are those that, using attribution models, managed to find out for their team and their client who is the goalkeeper, who is the defender, who is Luka Modrić, and who is Christano Ronaldo. Or Mandžukić, so it doesn’t appear that I only know a few football names.
In one team Facebook acts as a goalkeeper, in another team Facebook is used as a midfielder, and neither the first nor the second team is necessarily wrong. Just as a player in his career with the help of different coaches can change several positions in order to find out in which position he is most effective with which coach, so also one coach (in our case the marketing manager) plays against a different opponent (competition) in every next game and with other players available (budget and available channels). Knowing the strategy he developed with the team, the coach develops tactics for the next game, just as a good marketing manager knows for each next campaign what to do and how to allocate the available budget to the players (channels) at his disposal, in order to achieve what better result. Whether everyone will run after the ball or everyone will hold their position is a matter of the coach’s knowledge and strategy.

Google Analytics can provide you with quality data on the basis of which you can conclude who in your marketing team is the goalkeeper, who is the defender, and who is in attack.

If you don’t know how to find attribution data, a good analyst will help you get that data, analyze it, and use it to improve your team, because that’s what analysts are for.

Do you want to lead a team of freshmen who just happily run after the ball, or do you want to be the leader of a successful team in which you achieved the best results with the help of experience and quality numbers? The tools and knowledge already exist, and the decision is up to you.