Most marketers today – outside of the e-commerce space – understand that a large share of our buyers actually did not click on an ad or an email. Instead, after seeing our ad or other communication, they found our brand via a search engine. And then they visited our site … or dropped by our store … or called our 800#.
And when we analyze “buy-throughs” after a campaign, we see that the majority of people who clicked are not the ones who purchased. Here’s an old post on that topic. It notes a study showing almost zero correlation between clicks and conversions.
At the heart of this is of course having the ability to measure.
For many companies, this all culminates in site visitors and conversions tracked via GA or another analytics tool. But what if somebody saw our ad, didn’t click, but later entered our name into their browser? Many might assume they came from either our SEO or Paid Search efforts. This is an example of “last-click attribution” – and it tends to bias results (and future spend) toward Search.
We have seen this turn a bit dicey when an ad agency shares a client with an SEM agency – and the latter gets credit for many of the former’s site visits and conversions.
So how to more accurately attribute conversions?
Here’s a great article by Casey Carey from a few months back. Casey covers this topic well – and how the media has covered this topic. It’s on everybody’s mind, but there are relatively few answers. He also poses a number of problems with last-click and simple rules-based attribution models. They tend to pre-define attribution channels, use those proportions to weight results, and end up with “exactly what we thought it’d be” – but with no empirical evidence to prove it. (And few real insights.)
These and other data-driven attribution models are many and varied. And they can be fairly challenging to grasp. A great view of all the options is in this post by Avinash Kaushik – author of two best-selling books on Web Analytics.
Casey also referred to this case study from Adometry’s automotive client that spells out how “advanced attribution” works in a real campaign environment. Yes, it’s a subtle promotion for Adometry’s SaaS-based attribution platform – but it’s some great info. (Thank you, Casey and Avinash, for some thought-provoking discussion.)
Back to the convergence of measurement and ethics…
Years ago, we could look like a hero simply by measuring marketing campaigns. And as data scientists, we have constantly sought more accurate measurement.
But now when a client spends $100,000’s on a complex campaign that is ostensibly “entirely measurable” – the stakes are much higher. Especially when these campaigns’ results are so easy for the layperson to misinterpret.
So it’s never been more important than now for all of us to act as trustees of our clients’ budgets – and the analytical data that drive them. It’s our fiduciary duty to help clients understand which tactics and budgets are truly producing the best results.