TURF Analysis Simulator

How to access the TURF Analysis Simulator

The TURF analysis simulator is accessible in the following Conjointly experiment outputs:

• Generic Conjoint: levels per attribute.
• Brand-Specific Conjoint: levels by attribute by brand.
• Claims Test: Preference scores *.
• Product Variant Selector: Preference scores *.
• Survey questions: multiple choice, star ratings, ranking, and constant sum.

It is not available for Brand-Price Trade-Off preference scores.

Conjointly also provides a dedicated TURF analysis tool, enabling you to conduct TURF analysis on your custom dataset.

* Conjointly generally limits the TURF analysis to 50 items because it is not realistic to determine relative preferences for more than 50 items for each individual. For Claims Tests and Product Variant Selector, it is only available if HB analysis is enabled, which can be done for tests with under 130 claims.

TURF analysis simulator settings

The following settings of the TURF analysis simulator can be easily modified within the Configuration tab:

• Number of items in combination
  Let you specify the number of items to be included in the combinations tested.
• Reach method
  As different data sources will have different representations of how respondents show preference, the TURF analysis tool allows you to choose how respondents are determined to be activated by combinations of items through one of the following reach methods:
  • Among top options for a person
    Respondents are considered activated by items if their scores rank among the top N items, where N can be adjusted by changing the threshold value. This reach method can be used where higher scores indicate a stronger preference for a particular item. For most applications, we recommend using this method, setting N = 1.
  • Among bottom options for a person
    Respondents are considered activated by items if their scores rank among the bottom N items, where N can be adjusted by changing the threshold value. This reach method should be used with data where lower scores represent higher preferences.
  • Greater than
    Respondents are considered activated by an item if its score exceeds the threshold value. This reach method can be used with data where scores above a certain value indicate preference.
  • Greater than or equal to
    Respondents are considered activated by an item if its score is greater than or equal to the threshold value. This reach method can be used with data where scores above a certain value indicate preference.
  • Equal to
    Respondents are considered activated by an item if its score equals the threshold value. This reach method is best used when scores of a specific value represent activation, such as data from a multiple choice question.
  • Less than
    Respondents are considered activated by an item if its score is less than the threshold value. This reach method can be used with data where scores below a certain value indicate preference.
  • Less than or equal
    Respondents are considered activated by an item if its score is less than or equal to the threshold value. This reach method can be used with data where scores below a certain value indicate preference.
  • Utils
    Utils calculates the average probability an item in a combination will reach the respondent if placed among a set of other combinations. This reach method is better at considering strong secondary preferences and is best used for utility scores from conjoints and MaxDiffs. It can also be used for any data with mean-centred scores (the mean is 0). There is no frequency value when using the “Utils” method.
• Threshold value
  Specify the threshold value to be used in the reach method.
• Ensure each combination has at least one high-ranking item
  This option forces combinations to contain at least one of the top N items, in which N can be adjusted through the option labelled Specifically, at least one item that is among top.`
• Combinations must include all of these
  This option lets you specify items that must be present in every combination in the output. For example, you may want to include all items that have already been released.
• Combinations must include any of these
  This option lets you select multiple items, ensuring that the combinations in the output contain at least one of the selected items. For example, when you are choosing any item from a specific category.
• Combinations must exclude all of these
  This option allows you to specify items to ensure they never appear in any combinations in the output, such as those identified as underperforming.
• Prohibited pairs
  This option allows you to specify the combinations of items to prevent them from appearing together in the output. Use this option to avoid combinations that may seem unrealistic to consumers.
• Exclude the following from analysis
  This option allows you to specify the items to be excluded from analysis, such as those belonging to competitors.

You can conduct multiple TURF analyses with different settings by duplicating an existing configuration and modifying it as needed.

Conduct TURF analysis on your custom dataset

Upload your own dataset

When you open the TURF analysis tool , you will see three preloaded example datasets. To upload a custom dataset into the app, click on the Upload a new dataset button:

The dataset should be in a .csv file where each row represents a respondent and each column represents an item ( Example dataset).

If you want to use weighting for your dataset, ensure that one of the columns is called weights and all its values are between 0 and 1.

You can then configure the settings as described above.

Export into PowerPoint and Excel

You can export the output of all tabs into either Excel or PowerPoint by clicking on the Export into Excel or Export into PowerPoint button. Alternatively, you can export tabs individually by clicking the Quick export button on the top right corner of each tab.

Please note that when there are multiple TURF analyses for the same question or item, only the first analysis will be exported.