My talk at Big Data & Analytics Innovation Summit, Singapore, March 2015. I explain how we create good user stories using user path analysis.
Look for video here - http://www.novozhilov.co/2015/04/following-user-discovering-value-of-big.html
Following the User – Discovering the Value of Big Data
1. Big Data & Analytics Innovation Summit
Nikolay Novozhilov, Wego.com
Following the user – discovering
the value of big data
2. About Wego
Wego compares hotel and airfares from
hundreds of leading travel sites from all
around the world on one simple screen
Wego was founded in 2005 in Singapore
7. User path seems like a good idea!
… but is it?
1. Landing page 2. Search results 3. Handoff
8. Look for the most common
paths… combinatorics kills you!
All revenue-generating
users have unique paths
Popular and bad solution - reduce details
Bounced user
9. Different approach: Look at
individual user story
User path
(way of analytics)
User story
(way of agile)
10. How does it work in practice?
Relational
tables with
events
Denormalized
sparse table
Natural language
generation
Human-readable
story
11. Final “user story”
• Visitor from SA with tablet came in
using organic search (google.com).
• Did a flight search from Jeddah to
Paris from 2015-02-06 to 2015-02-23
for 1 adult
• Did a flight search from Jeddah to
Dubai from 2015-02-06 to 2015-02-23
for 1 adult
• Did a flight search from Dubai to Paris
from 2015-02-06 to 2015-02-23 for 1
adult
• Clicked on airtickets.com for 1378 USD
12. How does it work? Lessons learned
You can include as many details as possible
User story allows to guess user’s motivation (can’t
be simulated!)
Most users will walk away silently but some - will
keep trying repeatedly and will show you the
problem
But do stats later! Don’t rely on user stories for
actions
Just random users will be a little boring. Look at
active users/revenue generating/returning users.
13. What I saw in my data?
Travel planning cycle and it’s length
Specific patterns like “leg” searches
Specific patterns for different
countries/languages
Bugs in Data and in Product
14. Technical details for geeks
Storage for events - Google BigQuery
Generating natural language – SQL query
Visualization – Google Spreadsheets
15. This is just one tool in the tool box but
often overlooked