Conjoint Analysis for CPG Product Development
Why CPG Companies Use Conjoint
Conjoint analysis is a quantitative research method that measures how consumers trade off product features against each other, and it fits CPG product decisions like few other methods can. CPG companies face a specific challenge: their products are defined by a small number of attributes (flavor, size, packaging, price, brand claim) that interact in ways direct questioning can't capture. Asking consumers "Do you prefer glass or plastic bottles?" gives you an answer. Asking them to choose between a glass bottle at $3.49 and a plastic bottle at $2.29 gives you a decision.
The method is standard practice at companies like Procter & Gamble, Unilever, PepsiCo, and Mars Wrigley, where product line decisions affect millions of units and small shifts in preference translate to large revenue impacts.
Common CPG Applications
Package Design Optimization
The highest-frequency CPG conjoint use case. Attributes typically include container type, size, label design, closure mechanism, and price. The output tells you which packaging elements drive shelf selection and how much consumers will pay for packaging changes.
A snack company might test:
- Bag size: 3 oz, 7 oz, 12 oz, Family (16 oz)
- Bag material: Standard film, Resealable pouch, Stand-up pouch
- Window: No window, Small window, Full front window
- Price: $2.99, $3.49, $3.99, $4.49
- Flavor claim: Original, "New Recipe," "Limited Edition"
The market simulator then answers: "Does the resealable pouch justify a $0.50 premium?" and "Does the 'Limited Edition' claim increase preference enough to offset a smaller package size?"
Pricing and Size Architecture
CPG pricing isn't just "how much to charge." It's about the relationship between sizes, price per unit, and competitive positioning. Conjoint lets you test the full price-pack architecture simultaneously.
This matters because size and price interact. A 12 oz package at $3.99 and a 16 oz package at $4.49 look different to consumers than you'd expect from the price-per-ounce math. Conjoint captures these non-linear preferences where straight calculation doesn't.
Product Line Rationalization
Large CPG portfolios accumulate SKUs over time. Conjoint helps answer: "If we drop 3 of our 12 flavors, how much preference do we lose?" The TURF analysis extension takes this further by finding the smallest set of products that reaches the largest share of consumers.
A condiment brand with 15 sauce varieties might discover that 6 of those varieties each appeal to fewer than 3% of buyers and overlap almost entirely with more popular flavors. Cutting them simplifies the shelf, reduces manufacturing complexity, and loses minimal total reach.
Shelf Simulation
Advanced CPG conjoint studies incorporate shelf position as an attribute, turning the analysis into a virtual shelf test. Combined with visual concept display (showing actual package images rather than text descriptions), this gets as close to real shopping behavior as survey research can.
CPG-Specific Design Considerations
Use Visual Stimuli
Text descriptions of packaging don't activate the same consumer response as seeing the actual package. For CPG conjoint, build product profiles as visual cards showing the package render at each attribute level combination. This increases ecological validity and produces utilities that better predict in-store behavior.
Most conjoint platforms support image-based profiles. Budget extra time for creating the visual stimuli, as a study with 4 attributes and 4 levels each could require dozens of unique package renders.
Include Competitive Products
CPG purchasing happens on a shelf surrounded by competitors. Running a conjoint study with only your brand's configurations misses the competitive context. Include 1-2 key competitor profiles as fixed alternatives in your choice tasks. This gives your market simulator a competitive baseline and produces more realistic share predictions.
Price Ranges: 60-140% of Market Price
CPG price sensitivity is real but bounded. Testing a $1 price for a premium ice cream produces meaningless data. Set price levels to span roughly 60% to 140% of the current shelf price for the category. This range captures genuine price sensitivity without introducing unrealistic options.
Recruit Category Buyers
Sample from people who actually buy the category. A conjoint study about frozen pizza should recruit respondents who've purchased frozen pizza in the last 30-60 days. General population samples dilute your data with respondents who have no category expertise and produce importance scores that don't reflect buyer behavior.
Panel providers can screen on category purchase recency. Typical CPG conjoint studies require 300-500 category buyers, which means screening 3-5x that number from a general panel.
Real-World CPG Example: Ice Cream Portfolio
A mid-size ice cream brand was preparing to launch a premium line extension. They needed to determine: pint vs. tub format, flavor count, pricing position, and whether a "locally sourced" claim would drive enough incremental preference to justify the sourcing cost.
Attributes tested:
| Attribute | Levels |
|---|---|
| Format | Pint (16 oz), Quart (32 oz), Half gallon (64 oz) |
| Price | $4.99, $6.49, $7.99, $9.99 |
| Flavor range | 4 flavors, 8 flavors, 12 flavors |
| Sourcing claim | No claim, "Farm fresh," "Locally sourced" |
| Package design | Current design, Premium redesign A, Premium redesign B |
Sample: 450 ice cream category buyers (purchased at least 2x in past 60 days), recruited from a consumer panel.
Findings: Pint format had the highest utility in the format attribute, driven primarily by the 25-44 age segment who associated pints with premium positioning. The "locally sourced" claim added meaningful utility, worth approximately $1.20 in willingness to pay. But "farm fresh" performed statistically identically to no claim at all, meaning its premium perception was minimal.
Premium redesign A outperformed both the current design and redesign B, but only when paired with the pint format. In the quart and half gallon formats, package design differences washed out. This interaction effect would have been invisible without conjoint.
Decision: The brand launched a pint-format premium line with 8 flavors, redesign A packaging, and the "locally sourced" claim at $6.49 (confirmed as the optimal price point by the market simulator). They skipped "farm fresh" and the half gallon format entirely.
Tips for Getting It Right
Invest in visual stimuli. Package renders cost $2,000-$5,000 to produce but dramatically improve data quality for CPG studies. Text-only conjoint in CPG is a false economy.
Test at the shelf level, not the product level. Include competitors. Include shelf position if possible. The closer your conjoint mimics the real purchase environment, the more predictive the results.
Watch for interaction effects. CPG attributes interact heavily. Price sensitivity changes by pack size. Claim effectiveness changes by package design. Your analysis should look for these interactions, not just main effects.
Use TURF for portfolio decisions. Conjoint tells you what individual products people prefer. TURF analysis tells you which combination of products covers the most consumers with the fewest SKUs. For line planning, you need both.
Validate with in-market testing. Conjoint predicts stated preference, not actual purchase. Follow up your highest-confidence scenarios with a small in-market test (controlled store test or regional launch) before committing to national distribution.
Frequently Asked Questions
How many respondents do I need for a CPG conjoint study?
300-500 category buyers for a standard study. If you're testing segment-level differences (e.g., heavy buyers vs. light buyers), plan for 200+ per segment. Recruit from category-specific panels to avoid dilution from non-buyers. See the sample size guide for formulas.
Should I use images or text descriptions in my CPG conjoint?
Images whenever possible. CPG purchase decisions are highly visual, and text descriptions of packaging don't activate the same evaluative processes. Image-based profiles produce more externally valid results, especially for package design and shelf optimization studies.
How does conjoint differ from concept testing for CPG?
Concept testing evaluates complete product concepts (often including messaging, imagery, and positioning) as whole units. Conjoint isolates the contribution of individual attributes. Use concept testing when you want an overall go/no-go on a product idea. Use conjoint when you need to optimize the configuration of attributes within a product.
Related Guides
- Conjoint Analysis: Complete Guide -- Full methodology overview
- Conjoint Analysis Examples by Industry -- Examples across healthcare, SaaS, telecom, and automotive
- TURF Analysis -- For portfolio optimization and SKU rationalization
- Concept Testing -- For evaluating complete product concepts
- Conjoint Analysis Survey Template -- Adaptable CBC survey template
- Van Westendorp Pricing Model -- Simpler pricing research for single-product studies
Optimize your CPG product lineup -- try Quali-Fi free for 14 days.