Archive for April, 2009

Photo: Laineys Repertoire http://www.flickr.com/photos/76283671@N00

Measurement is a chicken and egg argument. Some will argue that unmeasured media has no value (usually the media company) others will argue measurement should follow innovation (the measurement companies). I’ve spoken about this subject before and I still believe what I said to be true: measurement companies are a rightfully risk adverse crowd, and companies looking to open new doors need to create their own measures of success. That means for publishers to move more brand dollars online they not only need to revise creative formats, focus on brand buying solutions, but they also need to develop new measures of success.

As an industry everyone’s wanted more brand marketer dollars even as far back as 1999 when the IAB did a press release titled “Abolish Clickthrough Now!” thinking that the metric didn’t help the cause with brand marketers. The idea of clickthrough as “the” metric for measuring performance took off early and to this day remains an accepted form of ad measurement. As you would expect brand advertisers didn’t see any value in clickthrough – why would P&G care if someone clicked on a Pantene ad? On the other hand direct marketers, those looking for the immediate satisfaction of a click loved it. As such we see the top categories advertising online (according to Nielsen) being the more conversion oriented folks in financial services (think credit card applications), retail, travel, and telecom. Right at the bottom of the list are the consumer packaged goods advertisers, automakers, and pharmaceutical companies – practically the inverse of spend on TV.

Okay, I’ll admit we’ve seen an evolution in the click-through metric to encompass: actions, performance, conversions and view-through, but come on people, it’s all an evolution of the same direct marketing measure. What the industry needs are brand metrics. Hard direct marketing style numbers than can be used to gauge brand success. Well that could be tough. As I pointed in the last post, the data that you’d want to base those metrics on are not available online. So what have we done to measure the effectiveness of online media for brand marketers? Well we’ve defaulted to what they already know, brand awareness studies. Now in full disclosure, brand awareness studies are my heritage. I got my start in Internet research in 1997 at Millward Brown where we tested the first Java ad, the first Flash ad, the first sponsorship, Unicast, etc. We were the only game in town and had plenty of demand for proof that brand marketing works online. We delivered stacks of reports that proved people see and are aware (at least subconsciously) of online brand advertising. We used strict experimental design methodologies, and had extremely convincing data. Everyone was happy. Or were they? It’s been over ten year since I left Millward Brown and we’ve see the online brand awareness research market grow considerably and the number of studies probably now amount to over 10,000. Yet here I sit at my computer eying the dollar spend of brand marketers online and wondering where are the big numbers? They obviously didn’t believe the studies or we would have seen more than the slow trickle of growth from brand advertisers, there should have been an inflection point in the growth rate of brand dollars – but it doesn’t exist.

It’s hard to pin all of the blame on ad formats, and buying options, those are simply impediments. Most if not all brand marketers have dabbled online, but none have come back with their checkbook. It appears they weren’t thrilled with what they saw, and that’s the rub. One could argue that the blame falls on the research companies and there may be some truth to that. Brand awareness research has a reputation of always returning great results, of course they have typically been underwritten by publishers or agencies who are not necessarily the most unbiased bunch. Let’s not forget to mention the way researchers presented the data made the numbers look entirely suspect. Results like “1239% increase in purchase intent” just don’t seem believable and are a pure manipulation of the data for better optics. But even if those negatives adversely impacted a clients view of the data it’s hard to argue with the aggregate data that shows brand marketing works online, right? Wrong. I mean yes of course the aggregate results looked good but the reality was even though the studies were designed to breed a level of familiarity to TV brand awareness studies, the methodologies were incompatible. As a result marketers had an out with online advertising: I don’t trust the publisher who funded the study, I think the numbers as portrayed are unrealistic and I have no idea how this compares to TV. If there’s one lesson to be learned from this exercise it’s not to give marketers measurement that claims to be the same as what they use in other media but is fundamentally different. Had online publishers opted for a more familiar measurement standard, say continuous tracking, then maybe they would have seen the benefit, then again I’m not sure online advertising would have looked all that good next to TV. The other option to take, and the more attractive option in my opinion, is to offer the measurement that was the promise of the internet. To provide a clickthrough type measurement for brand marketers.

There has been some success along those lines. Perhaps the best example comes from Yahoo! and their Consumer Direct product. Designed in conjunction with Nielsen, it allows consumer packaged goods advertisers to buy their specific target within the Yahoo! audience (e.g. diet cola drinkers) and measure the sales lift by using Nielsen data to monitor product purchases in grocery stores. It’s some of the coolest measurement I’ve seen but suffers from a couple challenges: it’s expensive and so it only works for large campaigns (no one wants to spend half their media budget on measurement), it relies on panels (small campaigns are hard to measure), and it’s doesn’t scale well (lots of manual labor). This is the closest we’ve come to offering compelling, data rich measurement of online ad performance. If only the measurement were automated, available for multiple industries, and as ubiquitous as the clickthrough, well, then we’re talking business.

The key takeaway here is that the measurement innovation came not from research companies but from the publisher looking to woo brand marketers. It’s the reality of how the market operates and more collaborations between sellers and researchers could lead to more online brand advertising. I also feel strongly that replicating offline measures online was an exercise designed to fail. It’s downright near impossible to control for all the differences between how media is consumed and brand marketers (like consumers) expect more from the internet. It’s now up to us to deliver.