Assessing Dueling Mobile Ad Forecasts: $27B vs. $16.8B

Two mobile ad forecasts were released last week almost on top of one another: one from eMarketer and another from BIA/Kelsey (BIA). (IDC is slated to come out with theirs very soon.) The figures they project for mobile advertising in four years are $10 billion apart. 

In one way or another most forecasts turn out to be wrong. Forecasts typically either fail to anticipate technology shifts or they have the opposite problem. They are often aggressive in assuming how quickly technology adoption will happen or change the market. Think about past predictions regarding the rise of digial advertising and the erosion of traditional media. It's happening but years after many thought it would.

I've certainly been guilty of incorrect predictions and aggressive forecasts in the past. So I now generally prefer the IAB's methodology, which reports on actual revenues after the fact.

Mobile ad spending emarketer

Let's take a look at and compare the eMarketer and BIA mobile forecasts, which are strikingly different. BIA says that US mobile advertising in 2013 will be worth $5.4 billion and $16.8 billion by 2017. By comparison eMarketer is much more bullish, saying that US mobile advertising will be $7.3 billion this year and $27 billion by 2017. The 2017 number is almost certainly way too aggressive.

New York-based eMarketer pegged 2012 mobile ad revenues at $4.1 billion. However my view is that when the IAB numbers come out we'll see something closer to $3.5 - $3.8 billion. However it's possible that eMarketer has it right. Google told financial analysts several months ago that the company's mobile "run rate" was $8 billion globally (including non-ad revenue). 

Local vs National mobile ad spend BAI Kelsey

BIA has raised the amount of its overall forecast from last year considerably but dialed back somewhat the portion allocated to local. That's because the firm began to recognize marketers weren't buying local fast enough. SMBs aren't buying mobile ads directly and brands have only recently started to explore local targeting in earnest. Depending on several variables that may accelerate in the next 12 - 24 months. 

YP said that it had $350 million in mobile-ad revenue today. It's not selling mobile advertising directly to the company's mostly small business advertiser base. Rather this is how the company is allocating or attributing revenue from ads that appear on mobile devices but are originally sold as part of a broader package. 

The local portion of BIA's forecast is dominated by search advertising, which has been the major contributor to local-mobile ad revenues. BIA maintains the assumption that search will continue to dominate local advertising throughout the forecast period. And mobile paid search is consistently expected to have more than 2X local ad revenues vs. mobile display in the BIA forecast.

Share of Mobile Local Ad spend by Format BAI Kelsey

Yet there are many more display impressions (in apps for example) than search queries. I don't know ratio off the top of my head but it's quite significant. If we're going to see billions in local-moble ad revenue it can't all come from search queries on Google. (Almost all paid search revenue in mobile [95%+] will go to Google for at least the foreseeable future, if not indefinitely.)  

Today paid search represents just under half of all PC-based ad revenue. It's likely that will track with mobile over time. 

I do believe that location will increasingly be used by mobile display advertisers, networks and exchanges. But it will also be used together with other variables as a way to reach particular audiences. Location will be both simplified for advertisers and incorporated into larger mobile ad-targeting concepts ("context"). Thus location will be a layer, among other variables, in mobile display and probably not remain a single targeting methodology -- except in geofencing and related "conquesting" scenarios.

Emarketer projects that the majority of mobile revenues will be controlled by a small number of companies: Google, Facebook, Twitter, Apple, Pandora, Millennial Media and "other." Other includes a large number of companies, including Microsoft, Yahoo, mobile exchanges/DSPs and still others. 

mobile ad revenues by company emarketer

The collective "other" category above is probably much too small. However I do agree that a relatively small group of large companies with significant scale will control and collect the lion's share of mobile advertising in the US, just as on the PC. Google, Facebook and Twitter will certainly be among them. 

Google is accelerating the growth of mobile ad revenue with its recent introduction of Enhanced Campaigns, which will push more AdWords advertisers into mobile at higher CPC rates. And by bundling PC and tablet advertising together paid-search on tablets will also grow much more quickly. Location-based ad targeting on tablets is a bit of a wild card: location matters somewhat less on tablets than smartphones but the "ad canvas" is much richer on tablets.

Facebook has also been dialing up the amount of ad revenue it generates from mobile, simply by showing more ads in its apps. Home is a wild card and may or may not favorably impact mobile ad revenue for the company. 

What qualifies as a "mobile" ad may become murkier and more of an attribution question as in the YP example -- such as combined tablet and PC ads in search or Facebook ads that appear equally in mobile and on the PC. And what qualifies as "local" ad in mobile is also a bit of an issue. I would argue a local ad in mobile is one that includes an explicit location mention in the ad creative (or landing page). A "local ad" can also be one that has no location mention but uses explicit location targeting at a DMA level or "below." Ads that target by state, province or region should probably not be considered "local."

Google's Enhanced Campaigns and related simplification of media buying and location targeting will significantly boost ad spending attributable to mobile. I think however the eMarketer numbers for 2016 and 2017 are still too high. I also believe the BIA position that most mobile ads will be localized is also incorrect -- unless the definition of local is radically enlarged.