How Pare works

From study set up to a board-ready decision in five working days. Each stage is instrumented; nothing happens off-platform.

Pare's concept screener uses the methodology BASES, Nielsen, and Quantilope use for serious concept evaluation. Respondents see one concept at a time, in randomised order, and answer a battery of questions on each.

The five pillars

• Purchase Intent. The headline commercial signal. Top-2-Box converted into a single comparable score across concepts and benchmarked against a 50% commercial floor.
• Appeal. Emotional response, measured independently from intent. Concepts with high appeal but lower intent often need a pricing or distribution rethink, not a recipe change.
• Uniqueness. Differentiation vs the brand's existing menu and the category. Low uniqueness is acceptable for safe extensions, fatal for innovation pipelines.
• Brand Fit. Does the concept feel like the brand. Catches positioning errors before they reach the menu.
• Believability. Does the brand have permission to deliver this. Premium items from value brands, authentic regional dishes from chain operators — both flagged here.
• Relevance. Sixth pillar. How well the concept fits the customer's use case — daypart, occasion, party.

Sequential monadic design

Respondents see one concept at a time, never side by side. This eliminates direct comparison bias — every concept is rated on its own merits, the way customers encounter menus in the real world. Order is randomised across respondents to remove position effects. Comprehension questions are asked before evaluative ratings so the rating reflects understanding, not priming.

The Concept Strength Index

A weighted composite across the six attributes, calibrated to commercial weight in published industry benchmarks:

Weights configurable for studies where brand priorities differ — defaults reflect commercial weight in published industry benchmarks.

TURF (Total Unduplicated Reach and Frequency) identifies the smallest combination of menu items that reaches the largest share of customers. Pare runs both greedy (incremental) and exhaustive search to verify the optimum.

Reach measurement

Binary "would order" responses build a reach matrix. The greedy algorithm adds items one at a time, each maximising new reach. Exhaustive search verifies for menu sizes where it is computationally tractable (up to ~25 items).

Intent strength layering

On top of binary reach, intent ratings (1–5) reveal which items pull hardest. A high-reach, low-intent item is a "default" — ordered when nothing else appeals. A low-reach, high-intent item is a niche with loyal fans. Both are different strategic decisions.

The reach curve and elbow point

The curve plots cumulative reach against menu size. The elbow point marks where adding more items stops adding meaningful new reach — the optimal menu size for the customer base. Beyond it: diminishing returns and operational complexity.

Segment reach analysis

TURF results filtered by diet, dining party, visit frequency. Confirms the optimal menu works across audiences and surfaces underserved segments.

Van Westendorp's PSM identifies the optimal and acceptable price range for any menu item by asking respondents four price questions: too cheap, bargain, expensive, too expensive.

The four questions

Each respondent gives four prices per item. The cumulative distributions of these prices, plotted together, reveal four critical price points:

• PMC — Point of Marginal Cheapness. Below this, quality is questioned.
• PME — Point of Marginal Expensiveness. Above this, customers walk.
• OPP — Optimal Price Point. Minimises rejection.
• IPP — Indifference Price Point. Perceived as standard / fair.

The acceptable range and verdict

[PMC, PME] is the acceptable range. Current prices outside this range are flagged: above PME = overpriced, below PMC = underpriced.

Revenue-optimal pricing

Beyond OPP (which minimises rejection), Pare also computes revenue-optimal price — the price that maximises expected revenue given customer willingness to pay. Often different from OPP, and often more useful for commercial decisions.

Highlighter testing lets respondents mark which specific words in a concept description they react positively or negatively to. The output is a sentiment-coloured heatmap of the description, plus diagnostic identification of which ingredients drive appeal or divide opinion.

Why it matters

Most concept tests give you a score. Pare gives you a score AND the reason for the score. "Truffle and hand-rolled drive 23% of the positive response — parmesan is dragging it by 6%." Actionable copy revision in one diagnostic view.

How it works

Concepts are pre-tokenised on creation. Respondents tap words to mark positive or negative reactions during the survey. Aggregated across respondents, each token gets a sentiment score and engagement rate.

The polarising tokens insight

Tokens with high positive AND high negative rates are polarising — not failures, but segmentation signals. Identifies which elements divide the audience and informs targeting strategy.

Survey data is only as good as the quality controls behind it. Pare applies automatic quality screens that match enterprise research standards.

Speeder detection

Respondents completing faster than a methodology-appropriate minimum (45 seconds for menu screener, 90 seconds for concept screener) are flagged. Default behaviour excludes them from analysis with a toggle to include.

Straightliner detection

Respondents giving the same rating to every item or every concept attribute are flagged as low-effort responses.

Duplicate detection

IP-based detection flags responses from the same IP within 24 hours without blocking — accepts but flags for review.

Logical-ordering validation

Van Westendorp specifically: prices must order as too_cheap < bargain < expensive < too_expensive. Illogical sets are excluded from price analysis.

Sample-size guidance

Pare surfaces warnings when results are based on under-target sample sizes. Recommended minimums: 200 for headline metrics, 400+ for segment cuts.

BASES five-pillar concept testing

Nielsen BASES methodology documentation; Quantilope concept testing white papers.

TURF analysis

Sawtooth Software TURF documentation; Krieger & Green academic literature.

Van Westendorp Price Sensitivity Meter

Original Van Westendorp (1976) paper; Quirks Marketing Research methodology guides.

Highlighter testing

Quantilope highlighter methodology; Ipsos qualitative-quantitative hybrid documentation.

Sequential monadic design

ESOMAR methodology guidelines; standard concept testing literature.