Message Optimization
Our client needed data to guide selection of the most compelling promotion of a cancer drug.
Armature helped to develop promotional messaging materials for our client's cancer product by finding most compelling "story" for presentation to health care providers. Their treatment was a second-to-market agent for a rare cancer type that requires special diagnostic tests that were not yet routinely adopted.
Our client needed robust, empirical data as rationale for selecting promotional messages and story flow.
The first problem: Armature's long experience told us that building stories through quantitative data generally yields low consensus and often over-weights efficacy.
The second problem: 28 messages, divided among 14 categories, was a lot of material to test. (There are over potential 1.2MM message-story combinations.)
Principals at Armature have done multiple message optimization studies using univariate and choice-model techniques, and are fluent in adaptive choice-based methodologies. Based on this long experience, we determined to field a self-administered online survey with a nationwide sample of health care provider specialists.
We chose an adaptive choice-based conjoint (ACBC) approach, where an individual's response history determines which specific message combinations they will react to. We knew that multiple message combinations tested via tasks that mirrors daily decisions (not rating, ranking, or MaxDiff) would reduce respondent fatigue, yielding better-quality data.
We also asked respondents to select messages they believed to be a necessary part of a sales call in which the rep could only discuss three (3) statements with them. We then correlated the stated and derived importance to further build support for which messages to include.
No other method would thoroughly test the message set given the practical limitations of time and the size of the specialist universe.
Our analysis showed where implicit and explicit decision-making converged. This gave our client the confidence that the recommended "stories" had robust support -- and our client also gained insight into what health care providers really need to know vs. what they say they need, which could inform all future evaluations of "willingness to Rx."