Limits of traditional TURF
Traditional TURF has three fundamental limitations: it maximizes reach instead of volume, the greedy algorithm can miss better combinations, and it ignores cannibalization between products.
The first limitation is about what TURF optimizes. Reach measures how many unique consumers are reached. But a consumer who accepts a product is not the same as one who buys it regularly. A niche product with small reach but loyal, high-frequency buyers can generate more volume than a broad product with high reach but low conversion.
The second limitation is algorithmic. The greedy algorithm chooses the locally best option at each step. That does not guarantee the globally best result. In practice, early choices lock in a direction — and sometimes choices that look good early lead to worse overall outcomes. The problem worsens with larger assortments and more candidates.
The third limitation is cannibalization. TURF counts a consumer as "reached" if they accept at least one product. But it does not model what happens when the consumer accepts several — which do they choose? And what happens with products that compete for the same segment? Without a cannibalization model you risk an assortment with high overlap.
Key takeaways
- Reach does not equal volume, high-frequency buyers matter more than broad acceptance
- The greedy algorithm does not guarantee global optimum
- Early choices can lock in suboptimal directions
- Cannibalization between products is ignored
- Large assortments with many candidates amplify the problems
Example
A snack producer used TURF and selected 8 variants. Four of them competed for the same segment (salty/crispy/adult). Reach looked good but volume came in lower than expected — consumers had 4 similar alternatives instead of the assortment covering different needs.
Discuss your assortment with us
Contact us