For Claude Code agents and AI coding assistants. This library documents 14+ real Ditto research studies run in production. Each study includes the exact questions asked, key findings, share links for viewing full results, and the techniques that made them successful.
How to use this library: Browse the study index to find studies similar to what you need to run. Click into any study for the full details. Use the questions, group configurations, and insight patterns as templates for your own research.
What it demonstrates: Multi-phase iterative research. Two AI systems (Claude Code + Ditto) founded a startup from scratch. Three research phases, 32 total personas, complete business validation in 4 hours.
Phase 1: Pain Discovery
Group: 12 US adults aged 45-65 managing aging parents
Questions: 7 open-ended about caregiving challenges
Key finding: "I'm responsible without real authority in a system that's chopped into pieces."
Phase 1 findings directly informed Phase 2 questions. The discovery that "authority without power" was the core pain led to Phase 2 questions about what "authority" means and what tools would restore it. Phase 2's finding that HIPAA-only (no POA) was the right starting point informed Phase 3's pricing validation. Each phase builds on the last.
3. ESPN DTC: Pricing Elasticity for Hedge Fund (30 Minutes)
What it demonstrates: Large panel pricing research for financial decisions. Real-money implications.
Group: 64 US personas from Ditto's "100 Americans" representative panel
64 personas (vs the standard 10) provided higher confidence on price sensitivity. The "100 Americans" pre-built panel ensured national representativeness. Multiple price points were tested in sequence to map the elasticity curve.
4. Loblaw / No Frills: M&A Due Diligence (16 Minutes)
What it demonstrates: Rapid voice-of-customer research for M&A evaluation.
Group: 20 Canada-based personas (ages 17-63), 80 total responses
Key finding: Sentiment decline reflected trust and consistency concerns, not price. "This was a credibility problem, not a cost problem."
Diligence impact: Changed recommendation from price-focused to trust-focused recovery strategy.
The M&A team had 30 days for diligence. Traditional consumer research would take weeks. Ditto provided directional consumer sentiment in 16 minutes, allowing the team to redirect their analysis from price to trust.
5. Michigan Secretary of State: Voter Sentiment (24 Minutes)
What it demonstrates: State-filtered political research with real constituent insights.
Group: 10 Michigan personas (ages 32-73), state-filtered with "state": "MI"
Topics: Digital ID acceptance, election integrity, data privacy, dark money
Key finding: Digital ID accepted only as "optional backup with offline verification." Privacy requires enforceable mechanisms, not promises.
The "state": "MI" filter (2-letter code) ensured all 10 personas were Michigan residents. This geographic specificity is the core value proposition for political research - generic "American voter" data has limited value for state-level campaigns.
6. MotorMinds: Auto Parts Sourcing Pain
What it demonstrates: Industry proxy filtering, 7-question non-leading framework, "ghost inventory" discovery.
Configuration
Group:
10 personas, industry: ["Automotive Manufacturing"], age 28-58
Study ID:
439
Questions:
7 (full non-leading framework)
Total responses:
70
Questions Asked
"In your current or past work, how often do you need to find, source, or order parts for vehicles or equipment? Walk me through what that process typically looks like."
"What's the most frustrating part of getting the parts you need? Tell me about a time when sourcing a part was particularly painful."
"Roughly how much time per week do you or your team spend hunting for parts, calling suppliers, comparing prices, or waiting on quotes? What's the cost of that time to your business?"
"What tools, websites, or methods do you currently use to find and order parts? What works well about them? What doesn't?"
"Have you ever tried a new tool or system specifically to make parts sourcing easier? What happened? Why did you stick with it or abandon it?"
"If you could wave a magic wand and fix ONE thing about how you source parts, what would it be?"
"What would make you hesitant to switch to a new parts ordering system, even if it promised to save you time and money?"
Key Findings
"Ghost inventory" is the #1 dealbreaker. Every participant described systems showing parts as in-stock when they are not. This was universal.
10-20+ hours/week spent on sourcing. One participant estimated "$400-$800/week in lost profit."
Users have tried and ABANDONED other tools. Reasons: fake inventory, VIN split errors, clunky UX, apps that crash.
Magic wand consensus: "Truthful inventory with guaranteed ETAs" - not faster search, not better prices, just HONEST information.
Question 4 ("When does screen time make you feel guilty, and when does it feel like good parenting?") was the breakthrough question. By explicitly naming the emotion (guilt), it gave participants permission to be honest about something they normally hide. This reframed the entire value proposition from "better games" to "guilt-free screen time."
9. PatientCompanion: Elder Care Communication
What it demonstrates: Healthcare industry filter for specialised roles.
Configuration
Group:
20 personas, Healthcare industry filter
Study ID:
442
Key Findings
Current call buttons are broken. Binary "help" button with no context. Staff arrive blind.
Cognitive decline is the core challenge. Many patients cannot articulate needs.
Family communication is a time sink. Repetitive updates drain staff.
Staffing shortages make everything worse. Any solution must save time, not add tasks.
Magic wand: "Context before arrival" - know what patients need BEFORE walking in.
This study revealed a critical pattern for any AI product in a regulated industry: the boundary between "AI assistance" (accepted) and "AI autonomy" (rejected) is the professional liability line. Engineers will use AI for suggestions but will not trust it for any decision that carries their professional stamp.
13. Mandel Diagnostics: Medical Device Adoption
Configuration
Group:
15 personas, Healthcare industry, eye care focus
Study ID:
446
Key Findings
Referral to specialists is the default. Most optometrists detect AMD but refer out.
Equipment ROI must be crystal clear. "How many patients per month to break even?" is the first question.
CPT codes are make-or-break. If insurance does not pay for the test, patients will not get it.
Space constraints are real. Small practices have no room for large equipment.
Clinical evidence is table stakes. Peer-reviewed studies required for credibility.
Question 6 ("Let's talk about the privacy tradeoff...") was direct about the tension. By naming the tradeoff explicitly, it produced honest responses about where the line is. The magic wand follow-up then revealed acceptable alternatives. This two-step approach (identify the boundary, then find what is on the other side) is replicable for any product with privacy implications.
16. Das Heilige Brot: German Bread Culture
What it demonstrates: Country-specific cultural research with small, focused group.
Configuration
Group:
6 German adults aged 30-65, country: "Germany"
Key Findings
Hotel toast was called "the carb mattress"
Brötchen described in near-spiritual terms
Bread is not a commodity in Germany - it is a cultural identity marker
Industrial bread is viewed as an insult to tradition
6 personas (smaller than the standard 10) produced deep cultural insights. For cultural research, smaller groups with tight demographic focus produce richer, more detailed responses than large diverse panels. The depth of individual responses matters more than statistical coverage.