12. GATHERINGDATA TO BUILD A SEARCHEXPERIENCE
Customer
Profile
Searches Purchases
Usage
Patterns
Customer
Service
• Products in Use
• How are they
used?
• Segments
• Paths
• Search
Sessions
• Clicks
• Behavior
• Domain
purchased
• Other products
bought
• Cart size
• Top complaints
• How they sell?
Competitors
Industry
Trends
13. O FOR OUTCOME
PICK THE RIGHT METRICTO MEASURE
Q: Optimize the entire path or just a page?
• Alignment on metric to build to:
– Customer experience
– Revenue ($, margin)
• Set explicit and realistic goals
• Clearly state assumptions and dependencies
16. T FOR TEST
“Test fast, fail fast, adjust fast”
• Do you have an A/B testing platform?
• Build v/s Buy
• What should you test and when?
• Do you trust your data - gut check test results
19. M FOR MVP
• Build a multi year roadmap of “big
rocks”
• Pick the 1st “rock”
• Define the simplest happy path
• Simplify it even further
• Build that!
20. DOMAINSEARCH MVP
Rule
based
model
Algorithmic
suggestions
Global
Ranking
Model
Market
Ranking
Models
Customer
segment
models
Path
specific
models
t = 0 t = 1y t = 2y t = 3y
Reorder Tokenize
Complex
rules
MVP 1 MVP 2 MVP 3
22. L FOR LAUNCH
• Launch your MVP
• Log everything (even if you don’t need it)
• Measure test results
• Understand dynamic between metrics
– Conversion, Order side, New customer acquisition, LTV
• Share bad news quickly, wait to share good news
23. WHAT IS STATITICALSIGNIFICANCE?
• Not leaving outcome to chance
• Probability (p) is your friend – helps manage
risk
• Easy to compute tools available online
– 95% confidence => You are wrong 1 in 20 times
– 90% confidence => You are wrong 1 in 10 times
25. N FOR NEXT STEPS
• Launch your MVP
• Add new “big rocks”
• Leverage Lean canvases
Source: Steve Blank
Prioritize: Customer Experience, not Revenue / Effort
29. TAKEAWAYS
• Think of data as your road to
insights
• Optimize the right metric
• Test fast, fail fast, adjust fast
• Minimum Viable Product
• Optimize for customer experience,
not revenue or effort
29