--- name: product_onboarding_activation description: Design post-signup onboarding that activates users: define the activation event, shorten time-to-value, script the first session, schedule progressive disclosure, instrument the funnel, and run activation experiments. Triggers on "onboarding", "activation rate", "time to value", "first-run experience", "aha moment", "onboarding funnel", "new user experience". --- # Product Onboarding Activation ## Purpose Move a new user from signup to the moment they feel the product working for them, in as little time as possible, and prove the funnel is improving with real numbers. This is the craft of activation: one observable moment of value, a first session built around it, and metrics that keep the team honest. ## Workflow ### 1. Define the activation event Pick one binary user action that proves the user experienced core value. It must be: - A user action, not a page view and not something the system did for them. - Something that produces a user-visible outcome. - Reachable in the first session or first few days. - Predictive of retention or conversion (validate by comparing what retained users did in week one against what churned users never did). If there is no usage data yet, interview 5 to 10 retained users and ask: "What was the first moment you felt the product was useful?" The most common answer is the activation event. Worked definitions by product type live in `references/activation-event-cookbook.md`. ### 2. Audit the signup-to-value path List every step between "Sign Up" and the activation event. Classify each step into two buckets: required for the user to get value, or required for the business (profiling, billing details, legal). Move every business-only step to after activation. Any step with more than 15% step-to-step drop-off is a friction checkpoint: investigate and redesign it, do not leave it. ### 3. Script the first session Decide what the first 30 seconds look like: - **Product-first**: drop the user straight into the product. Fits simple products and consumer mobile. Risk: an empty screen with no idea what to do. - **Guided setup**: a short wizard that configures the product. Fits products where personalization genuinely changes the experience. Risk: asking for effort before delivering value. - **Value-first**: show the outcome immediately with sample or demo data. Fits products whose empty state has nothing to show. Risk: it can feel fake. Whichever lane you pick, the rules are the same: one clear next action at all times, no dead ends, visible progress if the flow is multi-step. The first session has exactly one goal: the activation event. One successful outcome beats a full feature tour. Supporting patterns: - **Checklist**: 3 to 7 items, ordered by value with quick wins first, progress shown, celebration on completion, always dismissible. - **Guided tour**: at most 3 to 5 steps, dismissible at any point, never repeated for returning users. - **Empty states**: explain what the area is for, show what it looks like with data, carry one primary action, and pre-populate with example data where it helps. Worked first-session scripts by product type live in `references/activation-event-cookbook.md`. ### 4. Schedule progressive disclosure Show only what the moment needs. Move billing, advanced settings, team permissions, and power features out of the first run. - Phase 1 (first session, 0-30 min): reach the activation event. Setup wizard, progress bar, inline tips only. - Phase 2 (days 1-3): secondary features and integrations. Contextual tips and email nudges. - Phase 3 (days 4-7): habit and power usage. Templates, shortcuts, advanced tips. - Phase 4 (day 8+): referral, premium upgrades, advocacy. Never require profile completion before demonstrating product value. People cannot make good profile decisions before they know what the product does. ### 5. Instrument the funnel Minimum instrumentation: signup timestamp, every onboarding step as a completion event, the activation event with timestamp, and day-1 / day-7 / day-30 retention. Segment everything by onboarding path taken, acquisition channel, persona or role, and plan type. Without instrumentation, onboarding improvements are guesses. ### 6. Run experiments Work the funnel in priority order: fix the step with the largest drop-off first. Test one variable at a time on that step: copy, CTA placement, number of required fields, social proof, tutorial format. Prove each change by comparing pre/post cohorts; seasonality and channel mix shift, so an experiment without a cohort comparison is a story, not a result. Out-of-product follow-up is part of the funnel: a welcome email immediately after signup, nudges for incomplete onboarding at 24 and 72 hours, a celebration email on activation, and feature discovery at days 3, 7, and 14. Every email links to the specific incomplete step, never to the dashboard home, so the user lands where they left off. ## Output Contract Deliver, in plain language: 1. **The activation event**: one sentence, binary (it happened or it did not), with the exact instrumented event name. 2. **The step audit**: every step from signup to activation, business-only steps moved or cut, and the single biggest drop-off identified. 3. **A first-session script**: the opening lane, the sequence of actions, a checklist or tour spec, and empty-state copy for key screens. 4. **A disclosure schedule**: what shows when, by phase. 5. **The measurement plan**: the event list, the formulas, the segments, and the experiment backlog in priority order. ## Operating Rules - Time-to-value beats tour completeness. If a step does not shorten the path to activation, it does not ship in phase 1. - The activation event must be something the user does. Anything else is a vanity metric. - A step with over 15% drop-off is a friction checkpoint, not a shrug. Investigate and redesign it. - Never invent benchmark numbers. Published claims (for example, "40% free-trial activation is strong") are directional at best; the team's own improving trend is the real target. - Never ask a user to finish a profile before they have felt the product work. - Keep the funnel instrumented before changing anything. A change without a before/after cohort comparison proves nothing. ## Supporting files - `references/activation-event-cookbook.md`: worked activation definitions and first-session scripts for common product types. - `references/funnel-math.md`: the formulas, a worked funnel analysis, and the experiment playbook. - `bin/activation_funnel.py`: computes activation rate, per-step conversion and drop-off, and time-to-value percentiles from an event CSV.