Understanding what customers do.
WHAT TO INSTRUMENT
Signup and activation Reaching the core value Use of each significant feature Frequency of return Errors encountered
WHY ERRORS SPECIFICALLY
Customers rarely report them, and they predict churn.
WHAT TO DEFINE
An active customer, precisely.
WHY
Every subsequent figure depends on it, and vague definitions produce meaningless trends.
WHAT TO AVOID MEASURING
Everything, indiscriminately.
WHY
Unused data costs storage and attention, and obscures what matters.
WHAT TO LOOK AT PER ACCOUNT
Whether usage is growing or declining How many people are active Which features are adopted
WHAT DECLINING USAGE INDICATES
Churn, ahead of the cancellation.
WHAT FEATURE DATA REVEALS
What to invest in, and what to remove.
WHY REMOVING MATTERS
Unused features still require testing, support and documentation.
WHAT TO BE CAREFUL WITH
Removing something a few customers depend on entirely.
WHAT TO DO
Contact them before removing anything.
WHAT PRIVACY REQUIRES
Collecting only what is needed Disclosing what is collected Not capturing customer content in analytics
WHY THAT LAST POINT
Sending customer data to third-party analytics is a breach of your obligations.
