Channel is not motion
The two mix freely. An app can live on the App Store and still need a salesperson for its larger merchants. A merchant can arrive through a cold email and activate without ever seeing a demo. Treating either category as an identity you adopt is how founders end up forcing a motion their customers never asked for.
Pick the shortest credible buying path
| Motion | Fits when | Founder responsibility |
|---|---|---|
| Product-led | Merchants can understand, set up and get value independently | Make evaluation and activation clear |
| Sales-led | Value, setup or operational change needs discussion | Qualify, demonstrate and coordinate delivery |
| Hybrid | Some merchants self-serve while others need help | Provide a clear route into assistance |
For most founders, the App Store is the foundation, with a self-serve website route alongside it wherever the product supports one. The objective is not to eliminate human help. It is to remove friction that buys you nothing and keep the help where it changes the outcome.
Write down the routes you actually support, not the ones you intend to. Self-serve runs from listing or website to install, setup, first value and paid plan. Assisted runs from a relevant contact to a qualified conversation, demo, decision, setup, first value and paid customer. A trial can sit on either path. Define activation and payment as separate events, and make sure every self-serve merchant can find a human without having to work for it.
Match effort to customer value
What you can afford to do to win an account depends on what the account is worth.
| Lower price | Higher price | |
|---|---|---|
| MRR per merchant | $50 | $500 |
| Monthly contribution | $40 | $400 |
| Founder time and onboarding to win | $300 | $300 |
| Time to recover the effort | 7.5 months | about three weeks |
Both figures are before churn enters the picture.
Nothing about the work changed. What changed is what you can afford to do to win and serve the account. Founder help is a reasonable temporary investment in learning. It stops being that when every low-priced install needs a demo. At that point the work is a permanent cost, and it should be labelled and priced before anyone calls the model scalable.
Model contribution before choosing the service level
Build the contribution estimate from the costs your product creates. Hosting is often trivial while support, third-party usage, fulfillment, incentives or claims-related costs are not. A blanket software-margin assumption hides the difference you need to see.
At $100 MRR, variable costs total $30 a month, leaving $70 of contribution. A $210 acquisition and setup investment gives a three-month simple payback. If support cost rises by $20 a month, contribution falls to $50 and payback moves out to 4.2 months.
The price never moved, yet the service model shifted the economics further than a respectable improvement in ad performance would have. Track setup effort and ongoing support as separate lines: setup is concentrated at the start, while support compounds across every month the customer stays.
If you offer assistance, make it an intentional offer. Define who performs setup, what the merchant supplies, how testing works and what counts as complete. "White-glove onboarding" should name a bounded service. Used loosely, it promises unlimited work that someone will have to do.
Use five questions to select a route
For each target segment, ask:
- Can the merchant understand the value without talking to anyone?
- Can they complete setup alone?
- Can they tell whether it worked during an evaluation?
- Does the price support human help?
- Does the change carry operational risk serious enough to need a conversation?
Do not score the answers by counting. Weight them by what they are, because two hard setup answers can outweigh five encouraging ones about the website.
| Situation | Practical starting route |
|---|---|
| Low setup effort and clear value | Self-serve with accessible support |
| Clear value but occasional setup blockers | Self-serve with optional assistance |
| Material migration or negotiated economics | Assisted evaluation and setup |
| Small and large merchants buy differently | Segment-specific hybrid routes |
Track handoffs so one customer counts once
Hybrid motions produce a reporting error almost automatically. The merchant installs independently, books help, then pays. Self-serve claims the win and so does sales. Both contributed, but the company acquired one customer.
Keep three separate fields. Source is how the merchant arrived. Assistance is what your team did. Payment is the commercial outcome. Separated, they let you ask whether sales assistance improves a segment enough to pay for itself. When you compare assisted and unassisted cohorts, remember that merchants who seek help select themselves. A higher conversion rate in the assisted group does not prove the call caused it.
Know when the motion should change
Watch two triggers. Add sales when larger merchants repeatedly need help and can carry its cost. Add self-serve when the setup path becomes repeatable enough to automate. Before hiring anyone to do either, document the repeated work, the volume and the process.
Do this week
- Draw your primary buying path in words and name every step that requires a human.
- Map the last five real buying routes, the hours each consumed and the first-value event in each.
- Compare what you observed with the motion you believe you are running.
- Track time to first value, time to payment, human hours per acquired merchant and conversion at each step.
Questions
Should a Shopify app be product-led or sales-led?
Neither is an identity. Product-led fits when merchants can understand, set up and get value independently. Sales-led fits when value, setup or operational change needs discussion. Hybrid fits when some merchants self-serve and others need help. Choose by the shortest credible buying path for each segment.
When does a Shopify app need a demo before purchase?
When the change carries material migration, negotiated economics or operational risk serious enough to need a conversation. If every low-priced install needs a demo, the help has become a permanent cost and should be priced before the model is called scalable.
How do I attribute a customer who installed alone, then booked a demo?
Count one customer and track three fields: source (how they arrived), assistance (what your team did) and payment (the outcome). That avoids both self-serve and sales claiming the same win, and it lets you test whether assistance pays for itself in a segment.
Related guides
Chapter 7: How to turn Shopify app installs into paying customersChapter 12: How to run a Shopify app demo that earns a decisionNumbers 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.