Media measurement has long been a bone of contention.

John Wanamaker once famously said,

Half the money I spend on advertising is wasted; the trouble is I don’t know which half.

Yet with digital advertising and social media marketing, we thought we could finally have an exact science.

But over the past few years, plenty of issues have been exposed. The world has changed.

From fake media profiles and click fraud from automated bots, to human farms repeatedly clicking, ad stacking and more.

This has put the spotlight on marketing and advertising technology investment and whether it’s actually delivering more effective outcomes.

It has been great to see more and more industry leaders come out in agreement, that the model is broken.

In particular when it comes to the one critical measure: does digital advertising drive incremental profit.

Not profit from a one off conversion. Not the first month of an ongoing relationship. Not custom from people who were going to spend anyway. But true incremental profit and how the acquisition was attributed.

Was it last click? Was it a multi-channel attribution? And is the data being analysis actually correct?

Well I’m here to break the news to you.

It’s not correct nor accurate

Data science in advertising measurement is badly flawed.

Vendors will have you believe that you need a tonne of tech in your martech stack.

To house the data, to query the data, to glean insights and trends, to target, to personalise, to push out messaging, to update, to optimise, to manage assets, to comment, to listen, to measure, to communicate between divisions and external vendors efficiently. And the list goes on.

All sounds fantastic. In theory.

If you scratch below the surface of the spiels and buzzwords, then you get to illogic and bias.

How?

Well let’s take a look at some scenarios.

Google analytics typically shows you a first and last click view. Or as they would have you believe with GA4 a multi-channel view if set up correctly with the “data-driven” model. Which “takes all interactions in a conversion path and calculates an attribution weight to each interaction based on the influence it has on a conversion path.”

But if conversions are the focus, then what about conversions that are not completed due to credit risk, and some other reason for knocking back a customer?

And what about conversions that are priced at a certain point, but in reality the retail store offered a discounted rate?

In both examples, if the data isn’t being passed back to the model in time to reanalyse, then the initial attribution model is optimising to a bias.

Plus what’s the lookback window period? The lookback window determines how far back in time a touchpoint is eligible for attribution credit. 30 days, 60, 90 days? Depending what you’ve set will impact the accuracy of the model. 

Social media walled gardens are great within their walled garden. But for sales beyond the walls, then the conversion results don’t always stack up to Google Analytics.

Take ‘view-through’ conversions for example. Where someone may have seen an ad but not clicked on it. And then at a later stage gone directly to a website and bought a product. Facebook can attribute the conversion as theirs if they’ve tracked the customer to where the ad was displayed. Even though the person may never have clicked or engaged with the ad when in Facebook.

And as mentioned earlier, what about people that were already going to buy anyway.

It’s important to ask your analytics teams, are you measuring incremental value only?

The bottom line

Media agencies are now having to adapt. Getting off the optimisation drug to revenue conversion.

And refocus to understand how real bottom line value can be identified and optimised against.

But it’s important to understand that data models for media measurement are never 100% accurate.

Getting too micro can take your eyes off the real game. Building a healthy brand, acquiring good quality customers, and building long term valuable relationships and custom.

And getting too macro can waste a lot of budget.

So it’s a delicate balance.

It’s also important to ensure that your media teams and agencies understand marginal contribution in determining the return on marketing investment (ROI).

Focus must be given to the audiences that drive exponential incremental marginal contribution (MC). As well as advertise the products that deliver incremental MC.

Don’t fish in the low value tiers.

My tip

Don’t believe all the bells and whistles of technology vendors or media managers.

Accept a degree of bias with media measurement.

But focus on identifying real bottom line value.

And seek agreement across your business as to the level of inaccuracy that you’re happy to have when looking at marketing performance reports.

Indicative trends will always win. And they’ll end up saving you in terms of time invested, team member numbers, set up costs and the variable costs of technology, data management, and creative variation.

If you’d like to discuss your current state of media management, and to evolve to the metrics that matter most, then please touch base with us here

We’d love to hear from you.