Does equipment guidance support better buying decisions?
Task, season and available-size context may improve paid conversion without reducing margin or recommendation quality.
- Primary metric
- Paid purchase rate among test participants.
RA LAB · EXPERIMENT CATALOGUE
Hypotheses that need testing.
Three directions for HONOR and three for 1 MLN KM. Choose a specific question and propose how to produce useful evidence.
The catalogue presents potential research directions. A published hypothesis does not imply open recruitment, an approved budget or data access. We agree on each task’s terms before work begins.
HONOR
Project context ↗What is a relevant reason for a customer to return?
Purchase context and seasonal needs may produce more useful offers than a uniform campaign.
- Primary metric
- Repeat paid purchase rate within the agreed window.
Where does product data break the online-to-store journey?
Checking stock, sizes and variants may reduce recommendation errors and drop-off between basket and fitting.
- Primary metric
- Share of scenarios without a critical stock or variant error.
1 MLN KM
Project context ↗How can we tell whether an AI itinerary is trustworthy?
Verifiable sources, constraints and explanations may support itinerary acceptance without critical errors.
- Primary metric
- Share of itineraries accepted after critical-condition checks.
Which personalisation actually helps a traveller?
A small set of goal-focused questions may yield a more useful itinerary than a long preference questionnaire.
- Primary metric
- Share of users accepting a verified plan.
Which planning assistance is worth paying for?
A clear verified outcome, such as an agreed itinerary, may support willingness to pay better than the general promise of an AI assistant.
- Primary metric
- Confirmed order rate for an agreed outcome.
Have a new lab challenge?
Bring the industry, problem, hypothesis, baseline, owner and boundaries. After discussion, we can define a project brief, skills and test plan.