TURF optimization

TURF, which stands for "Total Unduplicated Reach and Frequency," is an optimization method used to identify a subset (Portfolio) of items that "reaches" the maximum number of respondents possible.

Originally developed for broader survey research, TURF can also be applied to MaxDiff studies. The goal of TURF is to find combinations of MaxDiff items that best motivate, satisfy, or reach the largest number of respondents without simply selecting the most preferred items on average.

In a scenario like stocking ice cream flavors at a store, TURF helps decide which flavors to include to maximize customer appeal given limited shelf space.

The problem is not as simple as including the portfolio featuring the most preferred flavors on average across the sample. By doing this, niche flavors that appeal to segments of the population (and that can increase total reach) would be overlooked.

For example, most respondents may have favored vanilla, chocolate, and cookie dough flavors on average across the sample. So, the natural inclination would be to offer just these flavors. However, if most respondents would be happy with any one of these three flavors, you could replace one (or even two) of them with other flavors (like strawberry, or mint chocolate chip) which might reach new segments of the market and increase sales.

This approach enables businesses to uncover market preferences that enhance reach and profitability effectively.

Below, we describe TURF settings that are available to you.

Portfolios are groupings (sets) of items which TURF is asked to optimize for maximizing reach.

Multiple sets of items (with somewhat different items involved) can lead to very similar near-optimal reach. Specify how many of the top performing portfolios you want to see in the report by modifying this field. By default, the number of portfolios is set to 20.

Items per portfolio is exactly what it sounds like: how many items are included in each portfolio.

The Items configuration menu allows you to customize how items are included or excluded in portfolios with two settings: Include in analysis and Force in all portfolios.

The MaxDiff Analyzer provides three different options for assessing "reach" in TURF: Weighted by probability, First choice, and Threshold.

Different reach methods (Weighted by probability, First choice, Threshold) often result in different optimal portfolio outcomes. For MaxDiff data, we typically prefer the Weighted by probability approach because it utilizes more information from potentially sparse data. Minor changes to the data, such as adding a few respondents, have less impact on the rank-order of optimal portfolios compared to the First choice and Threshold approaches.

The First choice approach makes less efficient use of the detailed preference data provided by MaxDiff and is the most sensitive to small changes in data, affecting the rank-order of outcomes more significantly.

The Threshold approach falls between First choice and Weighted by probability, but determining the optimal probability threshold is challenging for MaxDiff data.

One challenge with TURF is that many solutions often yield similar levels of Reach. However, this can be viewed as an opportunity. Additional criteria, such as expert opinion, can help determine which Portfolio is best suited to address the business problem. For example, if a grocer knows that a flavor appearing frequently in top sets spoils quickly, such solutions might be avoided in favor of others with similar reach.