They also admitted that most of the time, the reason they do hardcore quantitative optimization is for organizational buy-in reasons, rather than true quantitative rigour. I'd concur--I've seen similar things come up in my clients over the years. However, using AHP to optimize mix does not mean that all the good ROMI work can get tossed--far from it. In fact, it's critical to understand the outputs of each marketing tactic quantitatively. What AHP does is function as a team-based tradeoff / optimization model--but the individuals involved still need good information on how different tactics are likely to pay off.
This is why building out the marketing metrics and the response curves is still critical. So, don't get excited, you still have to measure marketing. Before you start AHP, it's helpful to have a robust list of metrics on a per tactic or per segment basis, such as:
- Rough ROI on a gross margin basis
- Diminishing returns levels--where does the curve get steep?
- Non-financial impacts due to different types of marketing
- Segment-specific "sweet spots"
- Current attitudinal levels by segment, country
There are some good software providers out there for AHP--one of my favorites is a DC-based company called DecisionLens.