The Guided Value Path
Define four moments for your app and write the real sequence, because activation and payment do not always happen in this order.
| Moment | What it means |
|---|---|
| 1. Installed | The app is connected to the store |
| 2. Setup complete | Required configuration is finished |
| 3. Activated | The merchant experiences a defined first value |
| 4. Paid | A paying subscription or agreed arrangement |
Never define activation as clicking a button when no meaningful value follows. In one founder-reported dropshipping app example, replacing a clunky onboarding flow with a wizard that led toward paid signup shortened time to paid conversion by approximately 90%. That is evidence about speed. It does not by itself establish a higher conversion rate or better retention. A 90% reduction would mean a 10-day interval became one day, and that is a mathematical illustration, not the recorded timing.
Measure where merchants stop
Count what happens within a fixed window of install, such as 30 days.
| Stage | Merchants | Rate |
|---|---|---|
| Installs | 100 | |
| Complete setup | 70 | 70% of installs |
| Activate | 40 | 40% of installs |
| Pay | 20 | 20% of installs, 50% of activated |
The figures locate a loss. They do not explain its cause, so inspect the failing step through observed behavior and merchant feedback. Separate required setup from optional preferences, explain the action, purpose and next step at each stage, preserve progress where supported, and show what remains. A wizard helps when it improves sequence and clarity. Dividing the same confusion into several screens does not.
Compare cohorts, not hopes
You improve the setup instructions and compare the next comparable cohort.
| Earlier cohort | Later cohort | |
|---|---|---|
| Installs | 100 | 100 |
| Setup complete | 60 | 75 |
| Activated | 30 | 45 |
| Paid by day 30 | 15 | 18 |
The later cohort looks better at every stage, and the largest relative gain is activation. Check whether traffic quality, product changes or support also changed. Without controlling those, describe an observed improvement and do not credit one screen with every gain. It is still useful operating evidence: it tells you where to look and what to preserve.
Follow up in a way that resumes the task
A generic "finish setting up" email asks the merchant to remember the work. A useful message names the incomplete task and the route back to it: "Your store is connected. The remaining step is [required configuration]. Complete it here: [link], or reply if you need help." Use accurate account state, and maintain suppression rules for paid customers, completed steps, cancellations and requests to stop. A stale reminder makes a well-built product feel poorly run.
Time to payment is only one view
Report three views together: paid conversion within a defined window, time to payment among converters, and later retention. A fast median among five payers can hide 95 abandoned installs, and a faster charge followed by refunds does not improve the business. For trials, track whether merchants reach first value before the evaluation ends, since registration alone is not enough.
Decide who gets human help
Assistance needs a trigger and a boundary. You might prioritize merchants who fit your profile, have completed part of setup and are blocked on a known issue, or offer help to every early customer while you learn the flow. Label founder-assisted onboarding during the learning stage: if the flow needs a founder for every install, you do not yet know how the unassisted product performs.
Ten assisted merchants take 90 minutes each, or 15 hours. At an internal planning value of $50 an hour that is $750 of effort. If three become paying customers, assistance adds $250 per payer before other acquisition costs. The rate is a costing assumption, not a statement about market wages.
For assisted accounts, hand the onboarding owner the merchant's desired outcome, the agreed scope, known dependencies, commitments and what was explicitly excluded. If setup begins with the merchant repeating the demo conversation, the process has spent some of the trust the sale earned.
Do this week
- Define activation for your app and map the essential setup steps.
- Pull five recent installs and map install, setup, activation and payment with dates.
- Contact a small sample of incomplete accounts, pick one change and evaluate a comparable cohort.
- Track setup completion, activation rate, paid conversion, median time to activation and payment, and support effort per activated account.
Questions
What counts as activation for a Shopify app?
A product-specific first-value event: the point where the merchant experiences the outcome they installed the app for. Logging in or clicking a button is access, not value. Define it for your app and measure it separately from setup and payment.
Should a Shopify app use an onboarding wizard?
A wizard helps when it improves the sequence and clarity of required steps. It does not help if it only spreads the same confusion across several screens. Test it against a comparable cohort and watch activation, not just completion of the screens.
How do I compare onboarding cohorts fairly?
Use equally mature cohorts over the same window, and note changes in traffic quality, product and support. Treat the result as an observed improvement unless you have controlled the other factors.
Related guides
Chapter 8: How much should you pay for a Shopify App Store ad click?Chapter 17: A weekly operating system for a Shopify app: from MRR goal to weekly activityNumbers marked as planning examples are arithmetic illustrations, not Shopify benchmarks or promised results. Founder-reported figures are not independently audited. Platform rules change: confirm current requirements with Shopify before you act.