Market research method
TURF analysis explained: reach and frequency for line optimization
TURF analysis (total unduplicated reach and frequency) chooses the combination of products, flavors or features that appeals to the largest share of people. Reach is the share of respondents who like at least one item in the line-up; frequency is how many of its items each respondent likes on average. The best line-up is rarely the most popular items on their own.
How results are checked
Named statistical review pending. Selected automated reference cases are published on the validation page.Run a TURF analysis — free, no signup
Open a respondent-level file with one rating or 0/1 column per item, choose the reach rule and the line-up size, and get the best combinations and the incremental reach ladder. Free and browser-local.
Run a TURF analysis →Why the most popular items are not the best line-up
Ranking items by popularity counts the same people again and again. If the second-favourite flavor is loved by the same respondents as the first, adding it to the shelf wins almost nobody new. TURF counts each respondent once — unduplicated — and asks which set of items reaches the most of them.
The input is one row per respondent and one column per candidate item: a purchase-interest rating, a “would buy” flag, or any answer that can be turned into reached / not reached with a rule such as top-2 box (4 or 5 on a five-point scale).
The method
Formula
Reach(S) = Σᵢ wᵢ · 1[ any item in S reaches i ] / Σᵢ wᵢ Frequency(S) = Σᵢ wᵢ · (items in S reaching i) / Σᵢ wᵢ
where
- S
- the line-up being evaluated (k items)
- i
- a respondent in the base; wᵢ is their weight (1 when unweighted)
- reaches
- the reach rule — top box, top-2 box, a value threshold, or an existing 0/1 flag
- search
- Crosstabs evaluates every combination of k items when there are up to 50,000 of them; larger problems fall back to a greedy search
Worked example
The synthetic snack tracker that the TURF tool loads as sample data asks 600 respondents for purchase interest (1–5) in eight new flavors. With top-2 box as the reach rule, the best three-flavor line-up out of 56 possible combinations is Sea salt + Chili lime + Salt & vinegar: it reaches 80.3% of respondents (482 of 600) with a frequency of 1.26.
| Line-up | Reach | Respondents reached | Frequency |
|---|---|---|---|
| Sea salt + Chili lime + Salt & vinegar | 80.3% | 482 | 1.26 |
| Sea salt + Chili lime + Cheddar | 78.8% | 473 | 1.28 |
| Sea salt + BBQ + Chili lime | 78.3% | 470 | 1.25 |
| Sea salt + Cheddar + Salt & vinegar | 77.5% | 465 | 1.27 |
| Sea salt + BBQ + Salt & vinegar | 77.5% | 465 | 1.24 |
The three most popular flavors on their own are Sea salt (48.7%), Cheddar (40.0%), Chili lime (39.3%). Together they reach 78.8% — less than the TURF winner, because their fans overlap. The runner-up line-up, Sea salt + Chili lime + Cheddar, trails the winner by 1.5 points.
| Step | Cumulative reach | Added reach |
|---|---|---|
| 1. Sea salt | 48.7% | +48.7 pts |
| 2. Chili lime | 69.2% | +20.5 pts |
| 3. Salt & vinegar | 80.3% | +11.2 pts |
| 4. BBQ | 86.3% | +6.0 pts |
| 5. Cheddar | 89.5% | +3.2 pts |
| 6. Sour cream | 91.7% | +2.2 pts |
| 7. Truffle | 93.2% | +1.5 pts |
| 8. Honey mustard | 93.7% | +0.5 pts |
Synthetic snack-brand tracker · 600 respondents · the same data as “Use sample data”.
The ladder shows diminishing returns: a fourth flavor adds 6.0 points of reach (to 86.3%) and a fifth 3.2 more. Where the curve flattens is where extra shelf space stops paying for itself.
Using it well
When to use it
Use it when
- Use: Choosing which flavors, variants, SKUs or features to launch when you can only offer a few.
- Use: Media and content planning: which channels or titles together reach the most of the audience.
- Use: Checking whether a proposed line-up has redundant items that reach the same people.
Not the right tool when
- Avoid: When items compete with each other for the same purchase and you need shares: use a choice model such as MaxDiff or conjoint with a simulator.
- Avoid: When interest is not a good proxy for behaviour (stated purchase interest overstates real trial).
- Avoid: With very small bases: reach differences of a point or two between line-ups are within sampling noise.
How to interpret it
Rule of thumb
Report reach first, then use frequency to break near-ties: a line-up with higher frequency gives more of its reached respondents a second item they like. Treat line-ups within about a point of each other as equivalent and choose between them on cost, margin or production constraints.
Choosing the reach rule
The rule decides what “reached” means, and the answer changes with it. Top-2 box on a five-point purchase-interest scale is the common default; top box is stricter and suits a launch where only definite interest counts. If the questionnaire already has a 0/1 “would buy” question per item, use it directly.
Fix items that must be in the line-up (a current best-seller) and exclude ones that cannot be made, then let TURF fill the open slots. Crosstabs supports both, plus respondent weights.
Exhaustive search versus greedy
With eight items and a line-up of three there are only 56 combinations, so every one is evaluated. The number grows fast — 30 items taken 8 at a time is almost 5.9 million — so beyond 50,000 combinations (or when a very large file would make the search too heavy) Crosstabs adds one item at a time, each time the one that adds the most reach. Greedy search is usually close to the optimum but is not guaranteed to find it; the incremental ladder above is built the same way.
Questions
Frequently asked questions
What does TURF stand for?
Total unduplicated reach and frequency. Reach counts each respondent once if any item in the line-up appeals to them; frequency is the average number of items in the line-up that appeal to each respondent.
What is the difference between reach and frequency in TURF?
Reach is the share of respondents who like at least one item in the line-up. Frequency is the average number of line-up items each respondent likes. Maximise reach first; use frequency to choose between line-ups with similar reach.
What data do I need for a TURF analysis?
One row per respondent and one column per candidate item, holding a rating (such as purchase interest on 1–5) or a 0/1 flag. A reach rule such as top-2 box turns each answer into reached or not reached.
How many items should the line-up have?
Look at the incremental reach ladder: each extra item adds less new reach. Stop where the added reach no longer justifies the cost of another item, or at the number of slots you actually have.
Is TURF the same as MaxDiff?
No. MaxDiff measures how much people prefer each item relative to the others; TURF finds the combination that reaches the most people. MaxDiff scores can feed a TURF by treating an item as reaching a respondent when it is among their top choices.
Sources and next steps
References & further reading
- [1]Miaoulis, G., Free, V., & Parsons, H. (1990). TURF: A new planning approach for product line extensions. Marketing Research, 2(1), 28–40.
- [2]Cohen, E. (1993). Data use: TURF analysis. Quirk's Marketing Research Review.quirks.com
Run a TURF analysis — free, no signup
Open a respondent-level file with one rating or 0/1 column per item, choose the reach rule and the line-up size, and get the best combinations and the incremental reach ladder. Free and browser-local.
Run a TURF analysis →