Pricing study with a choice-based conjoint for a subscription product
Design a choice-based conjoint survey, estimate what customers value, simulate prices against competitors, and recommend a price and plan structure with a test plan.
Industry project · for Pantry & Pan, a UK meal-kit subscription company with 41,000 active subscribers
Pantry & Pan delivers recipe boxes across Great Britain. It has 41,000 active subscribers on a plan of 3 recipes for 2 people a week at £4.20 per serving, about £25.20 a box, with delivery included. Ingredient and delivery costs have risen by about 14% in two years, and gross margin has fallen from 38% to 29%. The board wants margin back above 34%.
The commercial director does not want to simply raise prices, because monthly churn is already 9% and two competitors are discounting heavily. She wants to know what customers actually value and what they will pay for: more recipe choice, faster delivery slots, recyclable packaging, or a cheaper plan with fewer options.
You are the pricing analyst. In eight weeks you will: - frame the pricing problem and the hypotheses to test; - design a choice-based conjoint survey with attributes such as price per serving, recipe choice, delivery window, packaging and a pause option, plus a Van Westendorp price sensitivity block; - simulate 600 responses from a stated utility model, so you can check that your analysis recovers the values you put in, and treat the result as your survey data; - estimate part-worth utilities with a multinomial logit model; - build a market simulator with two competitor offers and find the revenue- and margin-maximising price and plan; - recommend a price and plan structure, with a live A/B test plan to confirm it.
Done means a recommendation that names the price per serving and plan features, the expected effect on take-up and margin, the risk to churn, and a test that would confirm or reject it within six weeks.
Pricing frame + survey design + synthetic response data with generator + conjoint estimation notebook + market simulator + recommendation deck with test plan
Stack: Python with pandas and statsmodels or pylogit, or R, Excel or Google Sheets, A survey tool such as Google Forms or Qualtrics (for the design mock-up), PowerPoint or Google Slides
- Sharp framing15
The pricing decision, margin target, constraints and hypotheses are written before the survey design.
- Sound survey design20
Attributes and levels are realistic, the design is balanced, and the price block is well formed.
- Correct estimation25
The model recovers the known utilities from the synthetic data and part-worths and willingness to pay are reported with uncertainty.
- Useful simulation25
The simulator compares plans against two competitors and finds a price that meets the margin target with stated churn risk.
- Testable recommendation15
The recommendation names price and features and includes an A/B test with sample size and decision rule.
Pass 60% · Merit 75% · Distinction 90%
Engineering Students
Diploma, B.Tech, B.E., BCA, MCA, or M.Tech students seeking practical experience.
Aspiring Developers
Developers who want to build proof-of-work capstone projects for tech resumes.
Job Seekers
Graduates looking for AI-evaluated code review & verified industry credentials.