When two review articles disagree, stop averaging them
Teams under deadline often “split the difference” between two technical review articles. The resulting strategy sentence sounds balanced and is usually false: it implies a midpoint that neither paper measured.
In our claim audits we ask three questions instead:
- What population and outcome does each review actually cover?
- Where do inclusion criteria diverge enough to explain the conflict?
- Which decision branches remain open until primary evidence arrives?
Sometimes the honest strategy is a narrower claim plus a dated evidence gap. That outcome disappoints sponsors who wanted certainty from secondary literature alone. It also prevents product and research teams from discovering the conflict later, in public.
A Team Strategy Intensive we ran last dry season spent an entire afternoon on two incompatible dosing reviews. The group left with a soft-launch criterion and a primary study outline — not with a averaged dose that no author had tested. Averaging would have been faster. It would not have been review.