TURF (total unduplicated reach and frequency) finds the small set of flavors, products or features that appeals to the largest number of different people. It helps you build a range that covers different tastes instead of repeating one.
You bringThe items you could carry, and the ones that must stay
You getThe line-up that reaches the most different people
ModelTotal unduplicated reach
reach=people who would buy at least one
TURF counts people, not votes, and shows which item brings in the most new people.
ReachFrequencyAdded reachReach on its own
The problem
The five most popular items are not always the best five
Pick the top five by popularity and you can end up with five similar items, loved by the same people. TURF counts people, not votes. It asks how many different people find at least one item they would buy, and which item brings in the most new people when it joins the line-up.
Top five by popularity
The five flavors most people would definitely buy
Chocolate
Salted caramel
Vanilla
Cookie dough
Mint chip
53%of buyers reached
vs
Best five by reach
The five that reach the most different people
Vanilla
Chocolate
Salted caramel
Pistachio
Mango sorbet
60%of buyers reached
Play with the model
Build a line-up and watch the reach
Vanilla and Chocolate, the two best-sellers, are locked in. Tick flavors to add them: each one shows how many new people it would bring in.
Tick flavors: the number on each is the reach it would add
Best reach at each size, and where your line-up sits
Reach counts people who would buy at least one flavor in the line-up. Change the rule and every number moves.
Reach counts people, not volume or revenue. Read it next to what you know about sales.
02
Line-ups a point or two apart are not really different, so you see the leading few, not one winner.
03
The rule for what counts as reached changes the result, so it is fixed up front.
In CatalyseAI today
The questions TURF runs on are live in CatalyseAI today: multiple choice with an exclusive “None of these”, rating scales and matrix questions, with up to five audience groups compared.
20question types, the TURF inputs among them
309ready-made question templates
5audience groups compared side by side
How the model works
How TURF works
Step 01 / 04
Collect the right answers
TURF runs on answers to a question you would ask anyway: which items people would buy (select all that apply, with a 'None of these' option), a purchase-likelihood rating for each item, or MaxDiff scores.
Question 4 of 9
Which of these flavors would you buy?
Vanilla
Strawberry
Salted caramel
Mango sorbet
None of these
Step 02 / 04
Agree what counts as reached
For example, 'definitely would buy', or the top two answers on the scale. The rule changes the result, so it is fixed up front and printed on every chart.
What counts as reached
Definitely would buy
Probably would buy
Might or might not
Probably would not
Definitely would not
Top box only, or the top two answers: fixed before the analysis and printed on every chart
Step 03 / 04
Test the combinations
For each line-up size, the combinations are compared and ranked by reach. Items you must keep can be locked in, and items you cannot make can be left out.
Salted caramel, Mint chip, Pistachio, Mango sorbet65.2%
Mint chip, Cookie dough, Pistachio, Mango sorbet65.0%
Vanilla and Chocolate locked in; 3,060 line-ups compared
Step 04 / 04
Try your own line-up
Tick items in a simulator to see the reach, and how much each item you have not picked would add.
Your line-up reaches 50%
Vanilla
Chocolate
Strawberry
Salted caramel
Mint chip+5
Cookie dough+4
Coffee+5
The deliverables
What you get
Every CatalyseAI report downloads in these formats.
PowerPoint deck
Excel workbook
CSV
Chart images (PNG)
D1
Reach curve
The best reach at each line-up size, so you can see where one more item stops paying.
D2
Leading line-ups
Several top combinations for your chosen size, not one winner, because many land within a point or two of each other.
D3
Added reach per item
How many new people each item brings in, in the order it joins.
D4
Line-up simulator
Build and compare your own line-ups.
The terms, in plain words
What we measure
01Reach
The share of people who would buy at least one item in the line-up.
02Frequency
The average number of items in the line-up each person likes.
03Added reach
The extra share of people an item brings in that the rest of the line-up missed.
04Reach on its own
What each item reaches by itself.
Worked example
Example: six ice cream flavors from twenty
A frozen dessert brand can stock six pint flavors in a new store chain and has twenty candidates. US adults who bought ice cream in the past month say how likely they would be to buy each flavor. TURF finds the six-flavor line-up that reaches the most buyers with the two best-sellers locked in, and shows which new flavor brings in the most people the best-sellers do not already reach.
FAQ
Questions, answered
Do I need MaxDiff to run TURF?
No. A 'select all you would buy' question or a purchase-likelihood rating for each item works. MaxDiff helps when the list is long and you also want a ranking.
Does the line-up with the highest reach sell the most?
Not always. TURF counts people, not volume or revenue. A line-up with slightly lower reach can sell more if its buyers buy more often. Read reach next to what you know about sales.
Why show several line-ups instead of one answer?
At normal sample sizes, line-ups a point or two apart are not really different. Seeing the leading options lets you choose between them on cost, supply or fit with the brand.