Explained: how we should treat GroupM, IPG and Zenith adspend forecasts

WPP’s GroupM, Publicis Groupe’s Zenith, and IPG Mediabrands all regularly provide updated forecasts about the state of global and country-specific adspend.

Earlier this month, The Media Leader reported on each one, all of which were quite similar in their overall estimates. Generally, all three sources predicted between 7% to 9% growth in adspend in 2022, each a slight downgrade from their earlier forecasts made in late 2021.

While the reports offer various potential reasons for why adspend growth is slowing down (ranging from the war in Ukraine affecting Eastern Europe to China’s ongoing pandemic lockdowns stifling advertising opportunities), the overall conclusion is practically the same. As it should be, arguably, if a forecast is meant to be accurate.

Does that mean we should be looking at these agencies’ spend forecasts as a part of a larger view? Or are there meaningful differences between the way these three companies report their numbers that mean we should treat them differently in our reporting?

“The three sources use essentially the same methodology: estimating net media owners ad revenues and forecasting based on year-to-date adsales and macro-economic outlook,” Vincent Letang, executive VP at MAGNA and author of the Mediabrands report, tells The Media Leader.

Letang says that small differences between the reports’ estimates can typically be attributed to the fact that some countries, such as China, are not fully transparent with their media owners’ ad revenues, meaning that gaps need to be filled with educated guesses.

But, Letang believes “there’s very little difference in the 2022 growth forecast” despite potential differences in market-by-market estimates.

Local market estimates ‘frequently disregarded’

Brian Wieser, global president for business intelligence at GroupM struck a slightly different tone, arguing that to parse out the differences between the reports, one would have to be very granular.

“Regarding methodology, there will be many, many differences. To properly assess that you need to do it data point by data point.

“In general, we are trying to estimate media owner ad revenues. Focusing on historical data points, we frequently disregard commonly relied upon local market estimates because we can see at a global level that collectively they are wrong, and so are fairly involved in helping create new historical estimates.”

Wieser suggests that the best way to assess the data is to look at the relative change in forecasts for any one forecast provider over time, though he hedges that even that doesn’t always work, as GroupM regularly makes historical revisions to its data whenever Wieser’s team identifies it can be improved.

Why US election spend is treated differently

One of the few noticeable differences between the Mediabrands and GroupM reports were the forecast of US political ad spending –  GroupM estimated $13bn in political adspend this year compared to Mediabrands’ $7bn estimate.

Letang explains that this is due to using different definitions for what counts as political advertising –  Mediabrands’ report only counts political ads that are explicitly focused on campaigns for elected office, whereas GroupM’s lumps those political campaign ads with ‘issues’ campaigns ahead of ballot initiatives (e.g. from the energy or pro-gun lobbies). Magna files those under more general adspend instead, arguing that because they run non-cyclically (i.e., not just during election season), they should be counted under a different category.

Despite such relatively minor differences in methodology, the multitude of similar reports does beg the question – do we need this many forecasts? And if they’re all useful, what is the best way to release them?

Rob Laurence, head of Mediatel Connected (the sister company to The Media Leader) and former Kantar director, looks at the reports this way: “Having multiple reports […] helps to give confidence in the general direction of travel that is being forecast.

“If they were wildly different, I wouldn’t know which one to believe”.

So, while critics may rightfully suggest that the various reports be coordinated for release, the value of multiple forecasts is great, and greater still when readers take the time to understand the methodology that makes them similar – and different.