The Strange Story of Online Ad Estimates

Three weeks ago, eMarketer – an analyst firm known for collecting research from multiple sources and digesting it for people interested in online media – announced that it was breaking with its previous practice of considering all available research. In what some industry figures termed a strange development, eMarketer decided to endorse the numbers from Price Waterhouse Coopers (PwC) for the estimation of online spending. These are the same numbers that the Interactive Advertising Bureau (IAB) – a lobby group for very large sites – employs to promote their industry.

Why would a firm, whose main value comes from considering all available information, decide to single-source such an important number? When asked, eMarketer CEO Geoffrey Ramsey said, “After interviewing or researching every firm out there, we couldn’t find anyone with a better methodology.”

PwC’s methodology amounts to comparing revenues that the sites self-report to the numbers the sites report to the Securities and Exchange Commission (SEC). This process makes sure they’re not telling two different stories. For instance, Yahoo! garners traffic of about 26 billion impressions per quarter and claims advertising revenues of $150 million. Presuming that an average CPM of $5.76 seems reasonable, they would pass the PwC test.



But a question arises: what happens if a big site inflates its numbers both in its financial reporting and in what it reports to the IAB? Pete Petrusky, the director of PwC’s new media group had a long answer and a short answer. The short one is that there’s no incentive for a publicly traded company to exaggerate the numbers. Inflating revenues only deflates profit margins, and the earnings aren’t affected one way or another. In other words: why would they want to lie?

But there has been just such an incentive for sellers to exaggerate revenues. A company that makes $1 billion in revenue and loses a few million dollars looks a lot more attractive than a company that gets millions in revenue and loses millions of dollars. If an Internet stock is supposed to incorporate a story of inexorable growth and the long-term “top-line” financial figures being more important than the short-term “bottom-line,” then there is a potentially powerful motive to make those numbers look as large as they legally can.

The temptation to pad the numbers may be further exacerbated by the fact that some advertisers like to spend money on sites that already appear to receive the highest spending.

Petrusky noted that the numbers are necessarily fudgy because other factors come into play, like barter between sites – some of which is counted as cash revenue – and other types of media deals, making things hard to track with precision.

In reference to the wild early days before the SEC moved to curtail the practice, Petrusky said that inflating ad revenues wasn’t “a sustainable long-term strategy. Most smaller companies were doing it.” He added that, “The bigger companies weren’t doing more than 10 percent [in barter transactions accounted as revenue].”

Petrusky did note that the percent that was barter referred to the percentage of reported revenues, not the percentage of inventory sold. In other words, since Yahoo! reported barter transactions approaching seven percent, that means that this accounts for about seven percent of its $150 million in ad revenues ($10.5 million), not necessarily seven percent of its 26 billion impressions (about 1.8 billion impressions). These figures were taken from Yahoo!’s most recent annual report, as filed with the SEC.

That means that, assuming the sellers are reporting revenues accurately, these fudge factors still prevent the media buyers from relying on them to determine how much inventory is really sold and how much is available for bargaining.

The SEC’s Early Efforts to Step In

A few years ago, the SEC cracked down sites over-reporting revenues through contrived barter transactions. Their rule-making body, the Financial Accounting Standards Board (FASB), said that in order to book revenue on a barter transaction, it must be substantially similar to existing real transactions.

When asked about more modern shenanigans going on in the Internet media space, FASB spokesman Tim Lucas said, “It should be considered to be revenue only under certain criteria. Valuation must be supportable. As long as enterprise can show same type of advertising for cash as they are swapping, they can still book it as revenue.”

Lucas’s interpretation differed somewhat from Petrusky’s. Lucas suggested that Yahoo!’s reported barter figure of 7 percent of revenues should also roughly correlate to a 7 percent of inventory, because of the “same type of advertising” provision in the rules for barter.

Calling up the audit arm of PwC – the firm also audits Yahoo!’s financial statements – proved fruitless, as their audit group’s PR firm referred queries repeatedly back to their new media consulting group. Yahoo!’s investor relations department did not respond to messages.

Petrusky, when asked how much “real” money is reflected in the reported numbers and how much “real” inventory is put up against it, said that there were many factors confusing the figures.

“There are impressions that are off the books,” he sad. “It’s not a one-to-one ratio; there’s a lot of performance-based deals, ceilings and performance-based bonuses, volume discounts, etc… It’ll never be a one-to-one reporting of the inventory.”

He said that if you take out barter, you can get at a “real” spending figure, but that it would also include several other types of revenues, including email, promotions and other types of marketing.

Does the IAB Have a Useful Methodology?

Rex Briggs, the person in charge of the IAB’s research and methodology issues, provided some clarity for the organizations reasoning behind sponsoring the PwC research.

“There are three different ways to skin this measurement cat,” he said, listing out the methods with their associated pro’s and con’s:

  • Look at reported revenues. The advantage is that it’s easy to audit and provides very stable numbers. The downside is that there’s a big time lag in reported numbers.
  • Measure ad placements directly, using spider technology. NetRatings, Evaliant, and others do this. They then use rate cards and discount assumptions to derive revenues. The advantage is that it’s not self-reported. The disadvantage is that it’s quite possible to miss a lot of ads or make incorrect assumptions.
  • And finally, a bottom-up perspective by surveying marketers as to what they’ve spent. This has the highest potential for accuracy, because it would eliminate the need for many assumptions, but it hasn’t proven very accurate in the past because of participation problems

    Briggs said the first method is likely the best one because it reports the most consistent data with the collection means currently available.

    “If you take the premise that [eMarketer’s] goal was to come up with a more complete picture of the revenue, then it’s a rational next step to focus in on what the ‘right’ number is,” said Rex when asked why eMarketer would endorse any one set of numbers. “In part it’s because they need to speak the same language as their clients.”

    Those clients are increasingly investment research firms rather than media buyers and sellers, and the financial people have difficulty rationalizing some market spending estimates to what the sites are reporting in their SEC forms. PwC solves that problem for them with the circular auditing process: what’s in the forms is what is used as the benchmark for adding up the market estimate.

    The “Tight Little Circle”

    Michael Kubin, Co-CEO of Evaliant Media Resources, another research firm, put eMarketer’s endorsement in a more stark perspective. “It’s so obviously strange,” said Kubin. “It’s at least incestuous. It really eliminates the ability to see what’s going on outside of this tight little circle.”

    Kubin said self-reported data is interesting, but not the most reliable. He said that having a third party observe the ads and estimate their scale eliminates some potential for a conflict of interest. He added that when companies like his create these estimates, they can often better determine what type of advertising is running – and perhaps whether or not it’s truly paid advertising, rather than barter or self-promotion.

    eMarketer’s Ramsey said that this was considered as part of the evaluation process. He admitted that “there is something to be said for getting around self-reported [methodologies]. A flag goes up immediately.”

    When asked if eMarketer has a financial relationship with either the IAB or PwC, Ramsey said there was none directly tied to his endorsement. He did say that these organizations “bought a lot of reports.”

    PwC’s Petrusky said there was no financial connection.

    Greg Stuart, CEO of the IAB, did not return several phone calls over the two weeks during which this story was researched.

    Different Numbers for Different Audiences

    An issue that seems unresolved for the PwC numbers – or anyone else’s for that matter – is the question of just whom they’re designed for. There are certain types of figures buyers would like to have. Others, sellers would like. Still others are necessary for investors.

    It appears that the IAB figures most closely aligned with the “financial” figures which were originally designed more for disclosure requirements than media intelligence. This could function as a good base for more refinement, if companies were willing to be a little bit more expansive about their revenue reporting policies. Unfortunately, while canvassing the top ten publicly traded web site companies, their respective investor relations departments weren’t very forthcoming. For instance, when asked to further explain what they booked as barter, Microsoft’s IR department gave a response only a lawyer could love: “Microsoft follows generally accepted accounting principles (GAAP) which have been set forth by the Accounting Principles Board (APB) of the American Institute of Certified Public Accountants (AICPA) and the Financial Accounting Standards Board (FASB). Media barter transactions are not a material portion of our financial results.”

    It seems that buyers prefer some of the more bottom-up methods of constructing spending estimates, as they avoid more of the noise included in financial disclosures. Kubin voiced this sentiment, saying that “if the answer is ‘if it’s legal, then report it,’ then that’s just plain insufficient for a media person.”

    In the past, the sellers in the online media business have tended to gravitate toward the numbers that provided the highest figures, making their offerings look attractive relative to other media. Currently, those are the figures of the IAB and PwC.

  • Next story loading loading..