TURF Analysis: Complete Guide for Researchers

What Is TURF Analysis?

TURF analysis (Total Unduplicated Reach and Frequency) is a statistical method that finds the combination of products, features, or options that appeals to the largest number of unique customers. Instead of asking "which product is most popular?" TURF asks "which set of 5 products covers the most people?"

The distinction matters. Your two most popular flavors might appeal to the same group of customers. Adding a less popular flavor that reaches a completely different segment could increase total coverage more than doubling down on what's already selling. TURF finds those complementary combinations.

The method originated in media planning in the 1950s, where advertisers needed to maximize unduplicated audience reach across TV and radio buys. George Miaoulis and colleagues formalized TURF for product portfolio optimization in 1990, and it's since become standard in CPG, food service, SaaS feature planning, and retail assortment.

Why TURF Analysis Matters

Most product portfolio decisions are made by looking at individual item popularity. The bestselling items get shelf space, the least popular get cut. This approach misses overlap: if your top 3 flavors all appeal to the same demographic, you're serving one segment three ways and ignoring everyone else.

TURF quantifies this overlap. It evaluates every possible combination of N items and calculates the unduplicated reach of each combination, meaning the percentage of your audience that finds at least one item appealing. The winning combination isn't always the set of individually most popular items. It's the set that collectively covers the most ground.

For a restaurant chain with 37 menu items, TURF might show that you can cut to 25 items and still give 91% of guests their first or second choice. That's a 33% menu reduction with minimal customer impact, translating to lower food costs, faster operations, and simpler training.

When to Use TURF Analysis

Common applications:

• CPG product lines: Which 6 flavors out of 15 should we offer to reach the most consumers?
• Restaurant menus: How many items can we cut without losing guests?
• SaaS feature sets: Which 5 features should the free tier include to attract the widest user base?
• Retail assortment: Which brands should fill 4 shelf facings in a category?
• Media planning: Which 3 channels maximize unduplicated audience reach?
• Advertising claims: Which 3 messages from 12 candidates cover the broadest audience?

How TURF Analysis Works

The Input Data

TURF starts with preference data: for each respondent, which items from the list would they choose, accept, or consider? This data typically comes from:

• Binary preference surveys: "Would you buy this flavor? Yes/No" for each flavor
• MaxDiff studies: Use the utility scores and set a threshold (top 5 items for each respondent)
• Purchase data: Historical transaction records showing which items each customer has bought
• Concept testing scores: Use purchase intent scores above a threshold as "acceptance"

The key requirement: the data must show, for each person, which items they'd accept. TURF doesn't work with aggregated popularity data alone. It needs individual-level acceptance patterns to calculate overlap.

The Algorithm

TURF evaluates every possible combination of N items drawn from the full list and calculates two metrics:

• Reach: The percentage of respondents for whom at least one item in the combination is acceptable
• Frequency: The average number of acceptable items per respondent within the combination

For a list of 15 flavors where you want to pick 5, that's 3,003 possible combinations. TURF calculates reach and frequency for each one and ranks them.

The Output

TURF produces a table showing the optimal combination at each portfolio size:

Notice: Vanilla is individually one of the most popular flavors, but it enters the optimal portfolio last because most Vanilla fans also like Chocolate. Adding Mango at position 4 reaches more new people because Mango fans have less overlap with the existing set.

The incremental reach column shows diminishing returns. Going from 1 to 3 items gains 25 percentage points of reach. Going from 3 to 5 gains only 7 points. This helps you decide where the cost of adding another item no longer justifies the marginal audience gain.

How to Run a TURF Study

Step 1: Define Your Item List

Identify the full set of items you're evaluating. This could be 10-40 products, flavors, features, or messages. Fewer than 8 and the analysis is trivial. More than 40 and the computation becomes intensive (though modern software handles it).

Step 2: Collect Preference Data

The most common approaches:

Direct acceptance survey: Show each item and ask "Would you buy/choose/use this?" (Yes/Maybe/No). Use "Yes" as the acceptance threshold, or combine "Yes" + "Maybe" for a broader reach estimate.

MaxDiff + threshold: Run a MaxDiff study to get preference scores, then define acceptance as "items in each respondent's top 30% by utility." This produces richer data than binary acceptance but requires a larger sample.

Purchase history: If you have transaction data showing which items each customer has bought, use that directly. This is the most valid approach but only works for existing products.

Step 3: Set Your Portfolio Size Constraints

How many items can you offer? This is driven by business constraints: shelf space, manufacturing capacity, menu board space, engineering bandwidth. Set a target range (e.g., "we want to find the best 4-6 flavors") and TURF will show you the optimal combination at each size.

Step 4: Run the Analysis

Most research platforms with TURF support (including Quali-Fi) handle the computation automatically. Upload your preference data, set the portfolio size range, and the system evaluates all combinations.

For custom analysis, R packages (tuRf, support.CEs) and Python libraries handle TURF computation. Excel works for small lists (under 15 items) but becomes impractical for larger sets.

Step 5: Evaluate the Results

Look at three things:

1. Which items appear in the optimal set? Are there surprises (niche items that round out the portfolio)?
2. Where do diminishing returns kick in? The incremental reach curve flattens at some point. That's your natural portfolio size.
3. How much overlap exists in the current portfolio? If your existing 10-item portfolio has the same reach as an optimal 6-item portfolio, you have 4 items worth reconsidering.

Sample Size Requirements

The key constraint isn't statistical power (TURF's math works with any sample size) but representation. Your sample needs to reflect the full diversity of your target audience's preferences. An all-millennial sample will produce a different optimal portfolio than a representative age distribution.

Real-World Examples

CPG: Sparkling Water Flavor Launch

A beverage company was launching a sparkling water line with shelf space for 5 flavors. They tested 12 candidate flavors with 400 category buyers using binary acceptance ("Would you buy this flavor?").

Individual popularity ranking: Lime (68%), Lemon (61%), Peach (54%), Black Cherry (51%), Mango (48%).

TURF optimal 5: Lime, Peach, Black Cherry, Mango, Cucumber. Total reach: 89%.

The popularity-based pick (Lime, Lemon, Peach, Black Cherry, Mango) had only 83% reach because Lime and Lemon fans overlap heavily. Replacing Lemon with Cucumber (individually only 31% acceptance) gained 6 percentage points of reach by capturing a segment that no other flavor attracted.

Restaurant: Menu Rationalization

A fast-casual chain had 37 core menu items and wanted to simplify operations. TURF analysis on 600 regular customers' stated preferences showed that 25 items achieved 91% reach for customers' first or second choice. Cutting 12 items would affect fewer than 9% of guests, and most of those guests had an acceptable alternative in the remaining set.

The chain phased out 10 items (keeping 2 of the 12 for regional variation), reducing food waste by 15% and cutting average order time by 22 seconds.

SaaS: Free Tier Feature Selection

A project management tool tested 18 features with 350 free-tier users to determine which 5 features should remain in the free plan (others would move to paid). TURF on MaxDiff-based acceptance data showed that task lists, calendar view, file sharing, basic reporting, and team chat reached 94% of users. The popularity-based top 5 (which included Gantt charts instead of team chat) reached only 88%, because Gantt chart users overwhelmingly also valued task lists and calendar view.

Common Pitfalls

1. Ignoring overlap between popular items. The whole point of TURF is that popular items often serve the same audience. If you just pick the most popular items, you're leaving reach on the table.

2. Using aggregated data instead of individual-level data. TURF needs respondent-level acceptance patterns to calculate overlap. Average popularity scores across the sample don't work.

3. Setting the acceptance threshold wrong. Too generous ("Yes" + "Maybe" + "Neutral") inflates reach numbers and makes every combination look equally good. Too strict (only "Definitely would buy") excludes legitimate customers. Test both thresholds and compare.

4. Treating TURF as the only input. TURF optimizes for reach but doesn't account for margin, brand strategy, manufacturing constraints, or customer lifetime value. The TURF-optimal portfolio is a starting point for discussion, not the final answer.

5. Ignoring frequency. Reach tells you how many people find at least one item acceptable. Frequency tells you how many items each person likes. A high-reach, low-frequency portfolio means most customers have exactly one option. A moderate-reach, high-frequency portfolio means customers have several options. The right balance depends on your business (one-and-done purchases vs. repeat/variety seeking).

TURF Analysis vs Alternatives

When to pick each: TURF when you're selecting a subset from a larger set and need to maximize audience coverage. MaxDiff when you need a rank order of items by preference. Conjoint when features interact and you need trade-off modeling with pricing.

How Quali-Fi Supports TURF

Quali-Fi includes TURF analysis as a built-in analytical tool. You can run TURF on binary acceptance data collected through standard survey questions, or on MaxDiff utility scores from a MaxDiff exercise within the same survey.

The platform evaluates all combinations up to your specified portfolio size and displays reach curves, optimal combinations, and incremental reach tables. Segment-level TURF is available for comparing optimal portfolios across audience groups.

Frequently Asked Questions

How many items can TURF handle?

Most software handles 30-40 items comfortably. Beyond that, the number of combinations grows rapidly (40 items choosing 8 = 76 million combinations), and computation time increases. For very large item sets, heuristic algorithms find near-optimal solutions without evaluating every combination.

Can TURF use conjoint data?

Yes, but indirectly. Run the conjoint analysis first to produce utility scores, then derive acceptance (e.g., products with total utility above a threshold are "accepted" by that respondent). Run TURF on the resulting binary acceptance matrix. This combines conjoint's trade-off precision with TURF's portfolio optimization.

What's the difference between reach and frequency?

Reach is the percentage of people for whom at least one item in the set is acceptable. Frequency is the average number of acceptable items per person. You want high reach (everyone has an option) and moderate-to-high frequency (customers have variety to choose from).

Does TURF work for services, not just products?

Absolutely. TURF works for any scenario where you're selecting a subset of options to cover the broadest audience: service offerings, advertising messages, media channels, event programming, course offerings, and membership tiers.

TURF answers a question that most teams skip entirely: not 'what's most popular?' but 'what combination reaches the most people?' Those are different questions with different answers, and confusing them is how you end up with a product portfolio where every item appeals to the same customer.

Related Guides

• TURF Analysis for Product Development -- CPG and SaaS portfolio optimization
• TURF Analysis for Menu Optimization -- Restaurant and food service applications
• How to Run TURF Analysis -- Step-by-step execution guide
• How to Interpret TURF Results -- Reading reach curves and combination tables
• TURF vs Conjoint -- When to use each method
• TURF Analysis Examples -- CPG, F&B, and retail case studies
• MaxDiff Analysis -- Complementary method for item prioritization
• Conjoint Analysis -- For feature trade-off and pricing analysis
• TURF Analysis Survey Template -- Ready-to-use template

Optimize your product portfolio -- try Quali-Fi free for 14 days.