What is Advanced Media Attribution, and How Do We Get There?

Media and marketing attribution is at the best of times hard to solve in all analytics and ad platforms, even with a range of built in metrics such as customer life time value, and return on investment. The problem with this however is that it doesn’t take into account how different media channels interact with each other.

What even is media attribution?

In short, attribution works out what path a user took, before they became a customer, a lead or whatever key action your business has. Think of a soccer player, all the credit generally gets given to the goal scorer, but what happened before then that led to that goal? How many touchpoints were there, what touchpoint had a material impact, and what combinations worked the best? Much like a soccer game, the goal is to figure out what formula works best to score goals, or in our case drive business. The issue with modern day measurement tools like Google Analytics, Adwords, Facebook ads and the lot is that they largely measure last click attribution - meaning they give the goal to the person who kicked it in last. And if they don’t use last click they use a form of attribution called ‘7 day view or 30 day click’ which means if a person sees your ad, literally scrolls past it on a news website or social newsfeed it will give major credit to that channel. As a result, every channel looks like it’s doing great, and your ad agency will tell you that because well, it looks like it is. But the reality is, to truly find out, and if you have enough data, you need to get manual with it, scientific with it and be a mathematician with it. 

So how do we solve it to get a true reflection on what is working?

Two words, data & statistics. This is where maths can actually be really impactful in marketing (who would have thought maths and marketing are closer siblings than you’d  think.) Marketing data such as clicks, reach and impressions is a highly valuable commodity. And other than being used to tell if a media campaign has been good or bad, is rarely used in any other way. Downloading that data and using statistical methods like regression analysis and bayesian theory is where you start to use the power of maths to build media mix models to find out the impact all media channels had on scoring that goal, purchasing that product or whatever key action you use. Using statistical analysis shows the relationship that different datasets are having, and because you are using marketing datasets we can then see what relationship your marketing or media channels are having. As a result, you will be able to see the attribution of all your channels in a more statistically significant and accurate way. 

So, what’s the point of trying to figure this all out? Well it’s quite simple. Better results, and better impact which will lead to more marketshare and more business. If you can have 20% more effectiveness and 20% more efficiency on spend over a 12 month period, over time it is inevitable that you are going to outdo any competitor while also looking good and having reporting that outdoes the finance department. Marketing is about growth and media mix models, and analysis is a big part of finding out how to grow faster and more sustainable. 

Is there anything I can do to get a more accurate picture before getting into techniques like regression to solve media mix models?

If you don’t have the time or resources to get into advanced media mix modelling, you can pick one source of truth to help make your measurement and reporting more accurate. Don’t report on key sales and conversions separately from Adwords, Social, Display etc. Use one source of truth to do your reporting from, and this should probably be G4 or Google Analytics. Dig into sections such as ‘User Explorer’ and ‘Multi-Channel Funnels’ and make them your second home. Make sure all accounts are linked into analytics and you have audience segments built everywhere. There is a lot you can do in digital analytics, and you can always take things further. Overall it’s always going to be hard for a marketing team to solve attribution when there  are agencies of all kinds trying to tell them to do more, spend more and create more. But, having a focus on finding out what formula works is also just as important because in the medium term if you can find out how to be more efficient and more effective. This is where you will get real cut-through and a methodology of what media mix to use when entering new markets and industries. Embrace mathematics with your marketing, as together you will get a 1+1 = 3 effect.