Playbooks
Playbooks: how a growth decision actually gets run
Each playbook takes one revenue decision and walks the whole way through it: who it applies to, the minimum data it needs, what Markin decides, what your existing stack executes, how the result is read against a holdout, and what normally goes wrong.
- Stripe/Execute a decisionRun a retention offer before renewal without paying customers who were staying anywayPre-renewal save offers on Stripe, sized by risk instead of by list membership.Read the playbook
- Braze/Work with the current stackAdd next best action to journeys you already run in BrazePer-customer arbitration written into the canvases your team already operates.Read the playbook
- Salesforce/Execute a decisionCross-sell a second product without cannibalising the plan the customer already pays forSecond-product offers valued on incremental margin, not on attach rate.Read the playbook
- Klaviyo/Execute a decisionWin back lapsed customers with a silence window instead of a discount ladderReactivation timed on returning intent, not on days since last order.Read the playbook
What a playbook contains
- What is a growth playbook?
- A growth playbook is a documented decision: the customers it applies to, the data it needs, the action taken, the system that executes it, and the experiment that proves whether it worked. It is more specific than a strategy and more durable than a campaign.
- Do I need Markin to run these?
- No. Every playbook is written so a team can run it with the stack it already has. Markin automates the parts that are hard to do repeatedly by hand: valuing every candidate action per customer, arbitrating to one, and reading each result against a randomised control.
- Why does every playbook include a holdout?
- Because without a randomised control the result absorbs seasonality, halo effects and customers who would have acted anyway. Independent research on incrementality testing finds a large share of active next-best-action programmes deliver marginal or negative lift once measured properly.
- Are the numbers in the worked examples real results?
- No. Every worked example is labelled illustrative and built on made-up inputs, to show how the decision is evaluated. Markin does not publish an average uplift figure, because the only number that matters is the one from your own holdout.