Twenty months after Coca-Cola’s Stan Sthanunathan promised in a conference speech that agencies would have to get used to the idea of “value-based compensation”, the evidence of it being used in practice is slim.

According to the ANA study, 49% of global advertisers, and 46% in the US, report using performance-based incentives. Yet the figures don’t look half so impressive against a survey by the American Association of Advertising Agencies, which found that performance incentives accounted for only 3% of the revenues of agency holding companies (as reported by AdAge).

There’s no hard data on how this might compare in the research industry, but anecdotal evidence suggest the figures would be significantly lower.

Awareness of the pay-for-performance concept is high, reports David Timberlake of The Marketing Information Company, which has set up a service to make sure that research agencies deliver on the contractual promises made to clients. Yet his firm focuses on measuring performance at the point of delivery – to verify that research quality and service targets are hit.

Much more elusive is the ability to link a research project directly to a successful business outcome and reward the agency accordingly. Research suppliers often point to the huge number of steps between research delivery and the execution of a strategy, meaning there might well be factors beyond their control that will determine whether the research lives up to its potential.

Others suggest that pay-for-performance is largely unnecessary: an agency is unlikely to be re-hired if the client-commissioner doesn’t feel that they have received sufficient value from a piece of work. This, they say, should be incentive enough, especially if the bulk of an agency’s work is on ad hoc projects.

• Click here to join the Research-live.com Debate Hub on Linkedin in to discuss this topic – and more – with 300 other research professionals