Industry
Retail
Use Case
Data Actionability
The Challenge
For years, HEB's insights team ran conjoint studies once or twice annually, each one a stressful, elaborate undertaking. That changed after attending Sawtooth training in 2017, which opened the team's eyes to what more frequent MaxDiff studies could really do. Adoption accelerated fast. But with scale came a new problem: stakeholders kept asking questions the data alone couldn't answer.
"This is so cool, but what do I do with it?" The scores showed what customers wanted. They didn't show whether acting on those preferences would grow basket size, change trip behavior, or actually move the needle for the business. HEB needed a way to close that gap and make MaxDiff results not just insightful, but immediately actionable for product, digital, and brand teams across a complex, fast-moving retail operation.
The Solution
HEB's insights team didn't abandon MaxDiff—they built around it. Working with Sawtooth’s tools, they developed four iterative techniques that layered context, nuance, and business relevance on top of the raw scores.
First, they added unaided respondent voice by placing open-ended questions before and after the MaxDiff to capture what was top of mind before customers were cued by a list of options. Text analytics on these responses revealed insights the scored results alone couldn't surface. In a sustainability study testing 33 topics, MaxDiff scores ranked "product sourcing" below the priority threshold. But unaided responses showed that the more specific phrase, "local sourcing from Texas", was the third most-mentioned topic unprompted, a distinction that became the foundation of a major HEB marketing campaign celebrating local farmers and businesses.
Second, they built business decision matrices by pairing MaxDiff scores with secondary metrics including current HEB share by category, purchase frequency, customer satisfaction, and brand fit. For their "Texas Size Packs" initiative, this approach revealed not just what customers wanted in bulk, but which categories HEB was already losing to Costco and Sam's Club. Those became the priority categories for the new product line. For inflation sensitivity research, they assigned memorable quadrant names (High-Stakes Change, Guilty Sacrifices, Consistently Essential, and Come and Take It) that gave the entire business a shared language for navigating price strategy.
Third, they developed multi-level anchoring to move beyond the traditional must-have / neutral / not-important structure. By running Hierarchical Bayes (HB) at multiple anchor thresholds, they could sort 40+ digital features into four tiers: must-have, ideal, nice-to-have, and low priority. For HEB's recipe experience launch, this gave the digital team a clear product roadmap rather than an undifferentiated pile of "important" items.
Fourth, they introduced baseline comparisons for brand trait research. HEB maintains a consistent library of 93 brand traits tested via MaxDiff across product lines and services. When "Texan" kept scoring in the top five across wildly different categories, including a health clinic, the team suspected the trait might simply be universally appealing regardless of context. A blinded baseline study confirmed that "Texan" scored fourth out of 93 for any brand at all. With that benchmark in place, they could identify when a concept genuinely needed a strong Texas identity versus when it was just riding the baseline, a distinction that shaped the branding of everything from their Kindly Cultivated produce line to their HEB Wellness Clinics.
Sawtooth tools and methods used:
- MaxDiff (standard and anchored)
- Hierarchical Bayes (HB) analysis
- On-the-fly MaxDiff scores with contextual follow-up questions
- Multi-level anchoring
- Baseline normalization for brand trait research
The Outcome
- From scores to strategy: HEB evolved MaxDiff from a prioritization tool into a full business decision framework, enabling product, digital, and brand teams to act on research results without needing a follow-up analysis.
- Uncovering what customers actually mean: Unaided voice questions revealed that "local sourcing" was a top sustainability priority for customers, sparking a major campaign that likely would've been overlooked entirely.
- Sharper portfolio decisions: Business matrix quadrants helped HEB identify which bulk product categories to pursue in their Texas Size Packs line based on where they were actually losing share to competitors.
- Cleaner product roadmaps: Multi-level anchoring transformed long lists of "important" features into tiered, actionable roadmaps, giving digital teams clarity on what to build now versus what to plan for later.
- Smarter brand building: Baseline comparisons for brand traits ensured that standout scores actually meant something, preventing over-indexing on familiar signals and leading to more intentional brand identities for new product lines and services.