Get value sooner

using data

by Adam Straughan

• Old site - complex SQL schema
• Customers renewing; are valuable, worth migrating

Traditional Approach

• 2-3 Month Development

• No Feedback

• Stored Value

• Waste
• Work in progress
• Store value
• Inventory

Incremental Approach

• Data driven choices

• Get value asap

• Agile team
• Incremental approach
• Use data to decide what is in each increment we release

Weighted Graph - weight is complexity to migrate, number of questions/nodes

• Buildings only
• Employed
• Joint policy holder

• Simplest thing
• least effort to migrate
• least number of questions answered

built 2 things

• Simple categorise system & log it
• Migrate one type of risk

• Choose an intercept point in the system
• Silently analyse the data (Comparison)

• Categorisation is really cheap compared to migration and log it
• Match against what we have implemented
• Process the data (Migrate in our case)
• Catch any errors log them and direct on to old system
  • missing fields for old data
  • renamed fields, etc

Value!

• Migrated in 2-3 weeks
• No 2-3 months
• Only one customer on day 1 - simplest (slightly unrealistic journey)

• Capturing data
• The is more value to the team

• Capture data
• First version not even categorising all the data event that was data driven

• populatity based on data
• team knowledge of comlexity and risk
• Do you want to deal with eye colour stored in Fax field now?

• Itterate
• We even did not implement everything, left it blank! based on data

Take Aways

• Aim for incremental

• Measure & Evaluate

• Get value early