OOH and econometrics – shouting the loudest or the longest?

Learnings from our Econometrics Week at Talon

The meteoric rise of digital advertising has had the effect of placing more emphasis on short term spend ahead of the values of long-term brand equity as a crucial part of the advertising offer.

Although we are now seeing some correction to this shift, this has still perpetuated a shorter-term vision for many brands, who cite the proof of performance solely in measures such as last click attribution and, yes, econometrics.

Econometrics’ definition as the modelled measurement of economics suggests it is not necessarily set up to measure advertising effect, let alone down to individual channels and products; but rather to oversee the effect of sales or other outcomes from various inputs that are economic measures.

It is not an exact science but has proved a very successful technique over the years when faced with numerous data points, data variation and predicting outcomes.

But businesses are using econometrics and market mix modelling to look for rationale for explaining every eventuality, even beyond how advertising drives sales effects. Most advertisers will start with a view on advertising’s overall effect, which sits well alongside traditional inputs like price, product, market size, promotion and seasonality.

But the main drivers of an advertising effect – and by some distance – are market size and the creative message. It informs the power of persuasion and the likelihood of consumers reacting to credible and influential messages.

Beyond this, many models then drive straight into two areas that are less robust and predictable. Channel and product performance.
The media mix may be a perfect one, but modelled data – which looks for variation not collaboration – will frequently recognise channels that shout the loudest or last, not necessarily those most persuasive or complementary.

Out of Home has suffered in this way for a long time. Effectiveness evidence from planning inputs like Route and device data which genuinely measure the complexity of OOH has improved beyond recognition. Measuring the specific OOH KPIs – which can include driving sales and footfall but will more likely extend to other key metrics like brand consideration and perception shifts – does not always extend as far as the econometric model.

Furthermore, inconsistencies in the quality and accuracy of ad spend data – still used by most models – creates an immediate issue. Granular data available through agency plans or Route is often not used.

Our econometrics week also highlighted some real concerns those modelling the data have when looking at product and channel performance to an individual level. And models will take variable approaches to longer term branding, halo effects and multimedia effects that we know play a significant part, whilst focusing on more measurable short-term outcomes.

Multimedia effects are more difficult to measure in today’s personalised world of advertising. Branding is fighting back – but 2020 plans will be analysed in 2021, affecting decisions in 2022. Meanwhile, location and device data can look at the here and now.
Creativity is certainly an additional elephant in the room. Our eye-tracking programme Talon Canvas identifies a huge range of attention outcomes to different ads. Compatibility with other brand content – for example across TV, video, digital and digital OOH – remains inconsistent despite the huge opportunity to converge screen messages.

All this points to using econometrics and MMM to inform but not in isolation. We fully understand and value drilling down to the contribution and effect of channels. But with the media and advertising industry never moving more quickly, a more rounded view of evidence factors and how individual campaigns or channels contribute to the whole is never more relevant.