Do you ever wonder that perhaps marketing isn’t quite as complicated as it appears to be from the outside? DAC’s Luke Regan suggests that marketers need to take back control from third parties with vested interests and those gaming the system.

Many marketers are keen to flaunt the scientific credentials of their methods and of the technology platforms they use. However, characterising marketing as a discipline steeped in scientific rigour would be something of a stretch.

In my experience, most of this ‘science’ is far from robust. At best it’s pseudo-science deployed to make marketers’ claims seem credible.

How do we know this?

True science is transparent – it declares its interests and welcomes being challenged. Peer review can be fierce in medicine and serves an important purpose. But where is this in marketing? I certainly can’t remember seeing the full source data or code released to qualify the results of any recent industry studies.

Fundamentally, our industry is just too cosy. If anyone challenges a finding they’re seen as a contrarian or troublemaker. But more of us should, because how much of the received wisdom would stand up to forensic examination?

Distribution in the dark

When it comes to budget distribution and attribution, marketers themselves are too often kept in the dark. We’re expected to trust tools such as programmatic DSPs or vendor-controlled, AI-driven black-box solutions to deliver optimal results, but without any real insights into what’s going on behind the curtain. 

As with industry studies, many choose not to question this because it suits them. There are many reasons why, with ‘goal hanging’ being one example – this is a strategy by which agencies and vendors aim to get as many cookies as possible served across an audience base to make a particular channel look good and justify investments.

However, this willingness to turn a blind eye or to actively game the system will eventually come back and bite agencies and vendors alike.

Conflicts of interest and moral hazards

The reason why it has become so hard for the industry to play fairly comes back to the status quo.

In the current marketing landscape, it’s in the financial interests of consultancies, agencies and technology platforms to deal in half-truths. The sector is filled with claims and counterclaims on which parts of the ecosystem are most effective and where the wastage lies – and it’s not just a minority of bad actors, the tech giants are as culpable as anyone else.

Not-so-coincidentally, the studies that individual agencies choose to reference tend to be those which align most closely with their own publisher volume discounts and overall business model. Few have anything positive to say about research that would call their model into question.

For instance, we see media owners funding studies into their own platforms and heads of effectiveness analysing effectiveness. Elsewhere, the ad fraud benchmarks used by the UK government come directly from the ad fraud vendors themselves!

In medicine any such conflicts of interest would be considered red flags, hence the randomised, double-blind studies that are seen as the gold standard in that sector.

Take back control

Despite the pitfalls around industry studies and third-party measurement/modelling, it would be impractical and unhelpful to ignore the data they provide altogether. Instead, marketers should be conducting their own experiments to validate previous findings within the context of their own brand.

We advocate a four-pronged approach:

  1. Do look at third-party research from others, albeit with a critical eye.
  2. Examine your own website and attribution reporting.
  3. Conduct media mix modelling.
  4. Find true incrementality in your performance media by running experiments with different channel mixes (with a statistically significant volume for each) to see what’s truly driving results.

For example, instead of running performance OOH nationwide, a brand could focus instead on a set group of locations, then find another group with similar historical sales performance and run no OOH in these. Over an (again) statistically significant time period, you would then measure to see if the locations with OOH in place experienced an incremental uplift.

With all four of these methodologies in place, marketers can be much more confident in identifying what’s working within their media mix – never 100% perhaps, but close enough to sleep soundly on our investments.

Simply put, one of the best ways to avoid bias, lies and pseudoscience is to do your own research.