Three problems, told as you scroll.

Leadership sees five versions of revenue. Step one is a metric dictionary: one definition each for revenue, orders, AOV, CAC and repeat rate.

With definitions fixed, twelve months of revenue sit on one axis, computed once from the same table that feeds the MIS pack.

Overlay repeat rate and a second story appears: retention is doing real work in the growth.

And the uncomfortable line: acquisition cost rose over the year. One page makes the trade-off visible to everyone at once.

Two models predict churn on 3,000 held-out fictional subscribers: a logistic baseline and gradient boosting.

The fancy model wins by less than one point of AUC. Honest baselines keep us from overselling.

Where the value really is: ranking. The top three deciles concentrate most churners, so a retention team knows where to start.

And why: SHAP shows the drivers, from viewing trend and days since last login to contract type and promo expiry.

A scorer reads a personal-data inventory and checks each activity against DPDP obligations. Overall readiness on this synthetic inventory: 50%.

Scores vary sharply by activity: payroll is in good shape; marketing campaigns and the learning app are not.

Every gap becomes a tile in a register, weighted by priority.

Then a roadmap: priority-1 gaps first, then the rest, then a rehearsed breach drill. Readiness work, not legal advice.