## Conjoint / Discrete Choice Analysis: An Example

A simplified conjoint analysis example follows for the market for cellular telephone plans. The example walks through an illustrative conjoint question that respondents would see as well as illustrative modeling scenarios.

### Respondent Survey:  Example Conjoint Question

Assume that study participants are asked to trade-off the features and pricing for various cell phone plans. A typical conjoint question asked of respondents might look like the following:

Which of the following cell phone plans do you prefer?
1400 minutes1000 minutes800 minutes
Unused minutes rollover for one monthNo rollover of unused minutesUnused minutes rollover for one year
New phone every yearNew phone every two yearsNo new phone
Nights and weekends count toward monthly minutesFree nights and weekends (do not count toward monthly minutes)Free calling to top 5 contacts
Costs \$100 per monthCosts \$75 per monthCosts \$60 per month

Respondents are asked a series of these questions where they must trade-off the features and pricing of the various plans to select the offering they most prefer. The pattern of responses is analyzed for each respondent in order to determine the underlying value system (or “utilities”). The conjoint simulation model is then built based on these utilities and allows for “what if” testing of any combination of features and pricing, whether or not that combination was shown in the survey. Conjoint analysis example modeling scenarios follow.

### Simulation Model: Example “What If” Scenarios

This simplified example shows the model output for the client plan and two competitive plans. (Note that an actual model of this sort would almost certainly include a portfolio of products for each competitor.) The model predicts the following preference shares for the client’s current cell phone plan and major competitive plans:

Plan:Client PlanCompetitor Plan ACompetitor Plan B
DescriptionClient brandCompetitor ACompetitor B
1200 minutes800 minutes800 minutes
No rollover unused minutesUnused minutes rollover for 1 monthNo rollover unused minutes
New phone every 2 yearsNew phone every 2 yearsNo new phone
Free nights and weekendsFree nights and weekendsFree calls to top 5 contacts
\$80 per month\$65 per month\$60 per month
Model Results (predicted preference share)25%35%40%

Once these base preference shares are determined, the simulation model is used to test changes to the client’s plan in order to predict market response. Three example “what if” scenarios follow.

### Scenario 1 – Pricing Strategy

What if the client decreased its service plan price by \$5 per month? Would the plan become more competitive?  In the conjoint analysis example below, when the client’s service plan price is decreased by \$5, the predicted preference share increases from 25% to 30%. Share is drawn from both competitive plans.

Plan:Client PlanCompetitor Plan ACompetitor Plan B
DescriptionClient brandCompetitor ACompetitor B
1200 minutes800 minutes800 minutes
No rollover unused minutesUnused minutes rollover for 1 monthNo rollover unused minutes
New phone every 2 yearsNew phone every 2 yearsNo new phone
Free nights and weekendsFree nights and weekendsFree calls to top 5 contacts
\$75 per month\$65 per month\$60 per month
Model Results (predicted preference share)30%33%37%

### Scenario 2 – Feature Enhancement Strategy

What if the client enhances the features of its service plan? Does its plan capture more share? In the example below, when the client’s plan features are enhanced, the predicted preference share increases from 25% to 35%. Share is drawn primarily from Competitor Plan A, the other feature-rich plan.

Plan:Client PlanCompetitor Plan ACompetitor Plan B
DescriptionClient brandCompetitor ACompetitor B
1200 minutes800 minutes800 minutes
Unlimited minutes rollover for 2 monthsUnused minutes rollover for 1 monthNo rollover unused minutes
New phone every yearNew phone every 2 yearsNo new phone
Free nights and weekendsFree nights and weekendsFree calls to top 5 contacts
\$75 per month\$65 per month\$60 per month
Model Results (predicted preference share)30%33%37%

### Scenario 3 – Portfolio Strategy

What if the client introduces a second product to compete with the lower priced competitive products? Does total share increase across the two client plans? In the example below, when the client adds a lower priced plan, the predicted preference share increases from 25% (for the original plan) to 46% (for both plans), drawing share primarily from Competitor Plan B. While the new service plan draws some share from the client’s higher priced plan, the total predicted share significantly exceeds that of the previous strategies tested.

Plan:Client PlanClient Plan 2 (new)Competitor Plan ACompetitor Plan B
DescriptionClient brandClient brandCompetitor ACompetitor B
1200 minutes800 minutes800 minutes800 minutes
No rollover of unused minutesNo rollover of unused minutesUnused minutes rollover for 1 monthNo rollover unused minutes
New phone every 2 yearsNo new phoneNew phone every 2 yearsNo new phone
Free nights and weekendsFree nights and weekendsFree nights and weekendsFree calls to top 5 contacts
\$80 per month\$60 per month\$65 per month\$60 per month
Model results (predicted preference share)28%18%30%24%

While countless scenarios may be tested, Sawtooth Technologies has the expertise to focus in on the strategies that result in the greatest leverage for your business.  To determine the optimal strategy, estimated costs and profits can also be included in the analysis.

For the conjoint basics please read “Understanding Conjoint in 15 Minutes,” By Joseph Curry, President and Founder of Sawtooth Technologies.