3. Sales Rep Compensation Example
• Base pay + variable commission
• For monthly <50k, commission = 3%
• For monthly 50-99k, commission = 5%
• For monthly 100k+, commission = 7%
3
4. Some Examples
Spam filtering arms race
Search engine ranking
Clearing cookies to get better airline prices
Keep account open to manipulate FICO score
Retail discounting/couponing strategies
Bidding in AdTech marketplaces
4
5. Measuring with Cross Validation
Cross Validation
• You should be doing this anyway!
• Set production performance expectation
• Measure post deployment
• Total deviation =
deviation due to overfit
+ deviation due to incomplete training
+ deviation due to Goodhart’s Law
5
6. Measuring via Homogeneity Assumption
Can you train a model to accurately
predict the date at which the observation was created?
6
10. Dealing with it
• Detection is key
• Experimentation is required
• Agile methods for model deployment
10
11. Causal Impact
• An approach to estimating
the causal effect of a
designed intervention on a
time series.
• Predicts counterfactual
(how response likely would
have evolved absent the
intervention)
11
12. Self Fulfilling Prophecies
• Beware!
• Case study: lead qualification
– Try to predict leads that will close
– Relearn the bias of your training
12
13. Fast Iterations
• Outside normal SWLC release cycle
– State updates
– Parameter tuning
• Run experiments
13
14. Explanatory power
• Goodhart’s law will often manifest on only a
subset of (possibly significant) instances.
• Model interpretability for effected instances is
key
14
17. Why Should I Trust You?
Explaining the Predictions of Any Classifier
Ribeiro, Singh, Guestrin
17
Model Interpretability
18. Summary
• Goodhart’s law: When a measure becomes a target, it ceases to be
a good measure
• As a data scientist, if your work is meaningful, you will encounter it
• Try to measure it in the data
• Work on explanatory models to mitigate
• Don’t let the average case blind you
18