- Ron Kohavi et al
- Data Mining and Knowledge Discovery
- manual (eg OLAP) or machine learning
Controlled Experiments
- simplest controlled experiment à A/B tes
- Overall Evaluation Criterion
o scorecard approach – with single metric
o forces tardeoffs
- A/A test
- use OEC components that have inherently lower variability
o impact on sample size
- effect of robots on experimental results
Analysis
- mine the data
- speed matters
- run continuous A/A tests
- Beware day of week effects
- many organisations have strong managers who have strong opinions , but lack data, so we started using the term HiPPO (Highest Paid Persons Opinion) to remind every one that success really depends on users perceptions.
No comments:
Post a Comment