Friday, September 25, 2009

ANA Agency/Client Forum: Agency Performance Isn't Based on Results

I spent most of yesterday at the Association of National Advertisers (ANA)’s Agency/Client Forum. The agenda covered a range of timely topics including digital advertising and social media. But I think it’s fair to say that most of energy was focused on the pocketbook issue of agency compensation.

In particular, the question du jour was value-based compensation or its cousin, pay-for-performance. I would have thought those were pretty much the same thing, but as Coca Cola’s Director of Worldwide Media & Communication Operations Sarah Armstrong set out in a detailed description of Coke’s own process, value-based compensation works largely by estimating a reasonable cost in advance, while pay-for-performance is based on after-the-fact assessments. (Coke’s process incorporates both – a base fee that is intended cover agency costs, plus up to 30% bonus based on performance.)

However, as several speakers made clear during the day, most pay-for-performance measures are based on agency behaviors such as innovation, strategic thought and execution, rather than business results such as sales, market share or even communications activities such as media cost savings. Specifically, an ANA survey that will be formally released in mid-October found that 56% of agency performance measures were based on qualitative metrics, vs just 19% on business results and 25% on communications metrics.

My initial reaction to this survey was pretty dismissive – either you pay for results or you don’t. But the fundamental rationale, mentioned by several speakers through the day, is that business results are affected by many factors beyond the agency’s control, so it really wouldn’t be fair to penalized or reward them purely on that basis. The analyst in me says it’s still worth trying to isolate the agency's contribution to results, but this is definitely a valid point. So including the subjective measures does make more sense than I initially thought.

A related question that ran through several presentations was whether agencies are commodities. One presenter flashed a survey that showed 80% of respondents thought they are (I didn’t capture the details of that survey, but think it was an informal online poll, so it’s probably not very meaningful). But both the clients and agencies among conference speakers felt strongly that they are not.

What was interesting, though, was the sorts of distinguishing features that speakers cited – strategic insights, brand stewardship, creative genius, etc. Those are based largely on the skills and chemistry of the individuals working on an account. As the cliche says, those assets “go down the elevators every night” – that is, they are individuals rather than property of the agency itself. So it’s possible that the agencies themselves are pretty much commodities (i.e., have about the same processes and technology) even if their people are different.

And even when it comes to people, I find it hard to believe that any one agency can really have people who are on average much better than any other agency. There are, in fact, plenty of smart and creative people in the world. Yes, there are occasional true geniuses, and clients lucky enough to find them working on their accounts may indeed gain a strategic advantage. Perhaps some of those geniuses are even so clever that they can build an entire culture around themselves to leverage their skills. But I'd say that level of genius is very much the exception.

In general, then, I suspect that once a quality agency comes up to speed, it would produce roughly similar results to another quality agency. This doesn't mean that you could immediately switch from one to another. But over the long term, agencies probably are something close to a commodity.

This relates back to the performance measurement questions. The value of an agency really does lie in its strategic, creative, and execution contributions, plus its ability to work closely with the client. In theory, most agencies should be able to do these equally well. But in fact, there will be variations based on the individual team members as well as (to a lesser degree, I think) differences in agency processes and culture. So it makes sense for performance evaluations to focus on those factors, even though they’re subjective. Marketers must measure those factors to identify areas needing improvement, either by changing performance of their current agency partners or switching to new ones.

No comments: