Skip to main content
Chat with us

Search

What can we help you with?

Advanced MaxDiff HB TURF Analysis

The Advanced MaxDiff HB can also be used with the TURF Analysis export feature.

In This Article

Before You Start

TURF is only available on a MaxDiff question when all of the following are true:

  • The MaxDiff is built in Advanced (HB) mode. Aggregate and Express MaxDiff questions produce item-level scores for the sample as a whole, not a utility score per respondent—so there is no individual to mark as "reached," and these questions will not appear in the TURF question picker at all.
  • The question has at least 150 completes, the minimum for the HB model to converge. (Roughly 400 responses is a separate, higher bar recommended for confident interpretation of the results.)
  • The question returned usable data (it is not flagged as bad data).
  • The TURF export feature is enabled for your team.

⚠️ Note: If your MaxDiff question is missing from the TURF question picker, the most common reason is that it was built in Aggregate mode rather than Advanced HB. This cannot be changed after fielding—the HB model has to be selected when the question is authored.

What is TURF Analysis?

TURF stands for Total Unduplicated Reach and Frequency Analysis. The main objective is to show how having multiple items "enabled" at the same time affects the total share of the audience those items reach between them. Reach is often used as an indicator of overall appeal, supporting conclusions about how a certain set of items/features/products/descriptions might perform.

Example study questions:

  • Which combination of product variants appeals to the largest share of the audience?
  • Where should we promote this product to maximize our reach in the industry?
  • In how many distribution channels would a respondent be exposed to the product, on average?

⚠️ Note: TURF models reach across an audience. It does not model competitive context and does not simulate share of preference, and Frequency should not be interpreted as purchase frequency or volumetric purchase interest.

MaxDiff TURF Outputs

Once the MaxDiff HB analysis is complete, individual utility scores for the items become available. TURF uses these coefficients to produce its output. To convert utility scores into "reached" and "not reached" statuses for a given combination of items for a given respondent, aytm offers five industry-standard export methods, selected as radio buttons in the Export tab once a MaxDiff question is chosen:

Method How "reached" is defined Threshold?
Average-based PL (50% bl) The respondent shows a preference for the item over an average item, at a likelihood of at least the set threshold. "bl" is the baseline: 50% is the point at which an item performs exactly on par with the average. Yes
Preference Likelihood {x}/screen Where x is the number of items per screen programmed in the question; the respondent shows a preference for the item over x−1 others, at a likelihood of at least the set threshold. Yes
Weighted by probability The respondent shows a preference for the item over an average one, with results weighted by each item's preference probability rather than cut at a fixed point. No
Weighted by probability {x}/screen As above, but preference is assessed against x−1 other items, where x is the number of items per screen. No
First choice The respondent is counted as reached only by the item they prefer most among those enabled. No

The value of x (items per screen) is set when the question is authored—it defaults to 4 and cannot be changed at export time.

⚠️ Note: These export methods are not the same thing as the three chart view modes on the MaxDiff question block (Preference Likelihood, Average-based PL, Utility Scores), which control how the chart is drawn and appear on every MaxDiff. Two of the names are shared between the two lists but mean different things—a chart axis in one case, a reach rule in the other.

⚠️ Note: The available methods and their default are configured per team. Some teams see a shorter catalogue with different labels. If your Export tab does not match the list above, check with your account manager.

The Reach Threshold

For the two methods that use one—Average-based PL (50% bl) and Preference Likelihood {x}/screen—a threshold slider appears below the method selection. It offers seven stops: 20%, 25%, 33.3%, 50%, 60%, 70%, and 80%, defaulting to 70%. The other three methods have no threshold and the slider does not appear for them.

💡 Tip: A common approach is Average-based PL (50% bl) at a 70% threshold. Since the 50% baseline means an item performed on par with the average, a threshold below or only slightly above 50% allows too much leeway in defining what is "appealing," whereas 80% can feel too exclusive. A 70% threshold balances potential reach with exclusivity. If 70% yields unrealistically high reach, refine by selecting 80%; if it yields unrealistically low reach, try 60%. Using Preference Likelihood {x}/screen is another common approach, with a lower threshold reflecting relatively high preference for the given number of items per screen (the random-pick baseline for 3/screen is ~33%—so use 50% or higher; for 4/screen it is ~25% and for 5/screen ~20%—so use 33.3% or higher).

💡 Tip: If you just want the standard output, the MaxDiff question block's own export menu includes a one-click Download TURF (default method) row—no trip to the Export tab required.

HB TURF.png

Excel Simulator

The analysis takes the form of an Excel workbook containing three sheets: TURF SIMULATOR, ALL COMBINATIONS, and LISTS. On the TURF SIMULATOR sheet, you can activate or disable certain items to see how it affects the summary statistics. You can also enter a total budget and item-wise costs to see additional statistics on how well a certain selection fits the budget.

TURF SIMULATOR sheet with items enabled and summary statistics

Combinations

The ALL COMBINATIONS sheet lists every possible combination of activated items and allows in-depth filtering to suit particular needs. For instance, you may be interested in the best combination of 3 items excluding one particular item. With default Excel filters, this can be achieved quickly.

ALL COMBINATIONS sheet listing every item combination with Excel filters applied

Considerations

If more than 19 items are being tested and available for TURF simulation, the number of possible combinations that can be explored is limited. This is due to limitations both in computing and Excel file dimensions—not because combination sizes beyond the given limit are unreasonable or not recommended. For more information on TURF simulation limitations, see TURF Simulator Export.

Warning about the item limit for TURF simulation

⚠️ Note: If you have more than 19 items in your TURF analysis and want to look at more combinations than provided by the simulator export, reach out to your account manager or the Customer Engagement team (support@aytm.com). The research and professional services team can consult with you on a custom simulator.