What is TURF analysis and when to use it?


Posted on 18 November 2019


TURF stands for Total Unduplicated Reach and Frequency. It is a technique that came into prominence during the 1950s in the space of media planning. In this report, we are only using the “reach” component of this technique.

Importantly, claims are tested one-by-one (not in combinations). Therefore combinations analysis through TURF is not the optimal way to assess preferences for combinations of claims.

On Conjoint.ly, TURF analysis is available in Claims Test and Product Variant Selector studies with under 50 claims or product ideas. Reach is the percentage of respondents for whom at least one of the claims in a particular combination is their most preferred claim. That is, it is a measure of how many respondents can be “activated” by a combination of claims.

Conjoint.ly’s user-friendly open-source TURF Analysis tool is currently under development.

A simple example of TURF analysis

Imagine you are launching a new brand of 🍹 vegetable juices. As you are preparing for the launch, you want to have a range of flavours that will appeal to (“reach”) the largest number of potential customers. But what flavours should you offer if your budget allows only two flavours?

Let’s start by listing all possibilities:

  • 🥑 avocado
  • 🥔 potato
  • 🥕 carrot
  • 🌽 corn
  • 🌶 pepper
  • 🥒 cucumber
  • 🥦 broccoli
  • 🍄 mushroom
  • 🌰 chestnut

What do you do? One way to solve this problem is to run a Product Variant Selector study, which will help you rank flavours by consumer preference. The same study will also give you inidividual-level preferences, such as:

Respondent ID 🥑 🥔 🥕 🌽 🌶 🥒 🥦 🍄 🌰
1 4.9 1.5 0.6 -0.8 1.2 1.8 3.9 -8.5 4.0
2 4.5 0.8 -1.3 0.1 -0.2 1.7 3.8 -8.4 2.0
3 5.5 -1.0 0.3 -0.3 1.4 0.7 3.6 -7.3 2.9
4 5.7 -1.2 -1.0 6.0 0.0 0.1 5.2 -8.9 3.1
5 4.3 -1.5 -0.3 -1.5 -0.8 0.8 5.9 -7.8 3.7
6 3.2 0.8 -1.0 -0.2 0.2 0.2 3.7 -8.5 3.5
7 3.1 0.2 1.2 -0.2 -0.9 0.5 3.1 -7.5 1.5
8 5.1 0.5 -0.1 -1.4 -1.4 0.3 3.1 -6.5 3.8
9 4.6 0.9 -0.9 -1.4 1.0 0.1 3.4 -8.0 1.2
10 3.2 0.9 0.5 -1.0 0.4 0.3 3.6 -7.8 1.7
11 3.4 -0.5 -1.4 5.0 0.2 1.5 5.8 -7.9 2.2
12 4.8 -0.9 0.5 0.3 -1.0 2.9 3.0 -7.2 3.7

Each cell shows the partworth utility of a certain flavour for a particular respondent. We can then assume that if a particular flavour is among the top two most liked by a person, then we call it appealing to them ✔. Therefore, these scores can be used to identify which flavour will be the most or the second most liked by each respondent:

Respondent ID 🥑 🥔 🥕 🌽 🌶 🥒 🥦 🍄 🌰
1              
2              
3              
4              
5              
6              
7              
8              
9              
10              
11              
12              

Next, we assemble several possible combinations of flavours and calculate a couple of metrics:

  • (Unduplicated) Reach, the percentage of people for whom at least one of the flavours is appealing.
  • Frequency, the average number of appealing flavours per respondent.
Respondent ID 🥑 + 🌽 🥑 + 🥦 🥑 + 🌰 🌽 + 🥦 🌽 + 🌰 🥦 + 🌰
1 ✔✔  
2 ✔✔  
3 ✔✔  
4 ✔✔  
5 ✔✔  
6   ✔✔
7 ✔✔  
8 ✔✔  
9 ✔✔  
10 ✔✔  
11   ✔✔
12 ✔✔  
Reach 11 12 11 9 6 11
Reach % 92% 100% 92% 75% 50% 92%
Frequency 1.1 1.5 1.3 1.1 1.0 1.1

As you see, everyone likes at least one of 🥑 avocado + 🥦 broccoli (Reach = 100%). This combination is a winner!

If you have more budget to launch three combinations, you can do the same analysis with three-way combinations:

Respondent ID 🥑 +🌽 + 🥦 🥑 +🌽 + 🌰 🥑 + 🥦 + 🌰 🌽 + 🥦 + 🌰
1 ✔✔ ✔✔
2 ✔✔ ✔✔
3 ✔✔ ✔✔
4 ✔✔ ✔✔
5 ✔✔ ✔✔
6 ✔✔ ✔✔
7 ✔✔ ✔✔
8 ✔✔ ✔✔
9 ✔✔ ✔✔
10 ✔✔ ✔✔
11 ✔✔ ✔✔
12 ✔✔ ✔✔
 
Reach 12 12 12 12
Reach % 100% 100% 100% 100%
Frequency 1.7 1.3 1.8 1.2

This time, all combinations have equally good reach. Now you need to look at frequency. If you offer the combination of 🥑 avocado + 🥦 broccoli + 🌰 chestnut, for an average consumer, there will 1.8 liked flavours from your brand. This is the way to go.

NB: This is not an analysis of mixing flavours in one juice product. If you want to look into combinations of features in one product, you are better off with conjoint analysis.