What is an MVP, and what is it not?
An MVP—minimum viable product—is the smallest version sufficient to test a product idea's most important assumption with real users. The definition is simple; in practice it is one of the most misunderstood concepts.
In B2B, an MVP is not a half-finished product. A version with a little of every screen but no complete end-to-end task is useful to nobody: users return to old methods and feedback never arrives. A good first release is narrow but deep. It solves one workflow for one user group, end to end, using real data and at a quality suitable for daily work.
An MVP is not a prototype either. A prototype is a set of clickable screens prepared to explore and test an idea quickly; it can be set aside once it has served its purpose. The first release is live, works with real data, and meets basic requirements such as authorization and security. Its scope may be small; its quality cannot be low.
Narrow but deep: one user group, one flow, end to end.
Write the success criterion first
Scope discussions usually start with a feature list. A more effective approach answers one question first: what should have changed a few months after the first release goes live?
The answer should be one measurable sentence. “Improve dealer experience” is not a criterion. “The pilot dealer group places repeat orders through the portal rather than by phone” is: it specifies who performs which task through which channel. Then ask the same question of every feature: is it necessary to achieve this criterion?
The criterion also determines what data to collect. Decide during first-release design; retrospective measurement after launch is often impossible.
One user group, one main workflow
Most B2B products have multiple user groups: dealers, field representatives, operations, finance, management. Serving all from day one multiplies scope because each has its own screens, permissions, and exceptions.
Choose one primary user group for the first release. Two questions help: who experiences the problem most often, and who personally decides whether to use the product? Then identify the group's most frequent, time-consuming task and describe it end to end: where it starts, which steps it passes through, and where it ends.
Suppose a distributor with 300 dealers is planning a dealer portal. The primary user is the dealer and the main flow is repeat ordering. The dealer signs in, builds a basket from a previous order, sees their prices and stock, confirms the order, and tracks its status. The field representative's dashboard, management reports, and returns process are valuable but secondary.
Divide features into three groups: first release, later, never
Once the criterion and main flow are clear, divide the feature list into three groups. This simplifies prioritization methods such as MoSCoW:
- First release: essential for the main flow to work end to end.
- Later: valuable, but do not affect the first-release criterion. Put them on the roadmap in order.
- Never: frequently mentioned in meetings but not tied to any user need. Writing this down prevents reopening the same discussion at every meeting.
Example: a hypothetical dealer portal
| Feature | Decision | Reason |
|---|---|---|
| Sign-in and dealer-specific pricing | First release | If dealers cannot see their own prices, they will not order through the portal. |
| Basket from a previous order | First release | The main workflow itself; speeds up repeat ordering. |
| Order status tracking | First release | If “where is my order?” still requires a call, the channel has not changed. |
| Account summary (read-only) | First release | Dealers struggle to decide on orders without seeing their balance. |
| Payment through the portal | Later | Payment infrastructure and reconciliation are separate work; the main flow works without them. |
| Promotion management | Later | Initially, promotions can be reflected in the price list. |
| Field representative dashboard | Later | A separate user group; a second-stage topic. |
| Messaging between dealers | Never | Contributes neither to the criterion nor to the main flow. |
We discuss the essential features of a dealer portal in our article What is a B2B portal?, and the scope and delivery conditions of such a portal on our B2B portal development service page.
Bridge integrations manually first
Integration is one of the most common causes of first-release delays. A bidirectional real-time ERP connection may take weeks; it is not always necessary to test the main flow. Initially, prices and stock can be imported by file several times a day, and incoming orders transferred to ERP through the operations team's screen. This temporary solution is chosen deliberately and recorded on the roadmap: when order volume exceeds manual handling capacity, full integration is the next task.
Defer reports and exceptional cases
Management dashboards look attractive in a first release, but there is no data to measure yet. A simple export or weekly spreadsheet suffices in the first weeks; actual usage reveals which reports are read. The same logic applies to exceptional cases. Rare situations such as partial deliveries, special payment terms, or foreign-currency invoices can initially use existing methods. What matters is listing these exceptions and communicating them clearly to users.
Address the riskiest assumption first
Every first release rests on assumptions: users will adopt the new channel, data quality will be as expected, and integration will be ready on time. Write these down while defining scope and ask two questions about each: what happens if it is wrong, and how early can we find out?
Put the riskiest assumption at the start of the timeline. In the dealer portal example, this is whether dealers are willing to switch from phone orders to the portal. You need not wait for the first release: test a clickable prototype with a few dealers in week 3 and see where they struggle. If the uncertainty is technical—for example, the format in which ERP prices can be retrieved—run a small technical experiment in the first sprint.
Leaving risky work until the end moves surprises into the project's most expensive weeks.
A 6–12-week first-release timeline
With scope narrowed this way, the jointly defined first release can go live in 6–12 weeks. A typical timeline is:
- Discovery (1–2 weeks): Users are observed at work; the main flow, data sources, and risks are identified. The success criterion and “first release / later / never” list are clarified here.
- Definition and prototype (2–3 weeks): Screens for the main flow are designed; the clickable prototype is tested with real users in week 3.
- Build (2-week sprints): Working software is shown weekly. The first release goes live in weeks 6–12, usually with a pilot group first.
- Grow and operate (ongoing): Track the criterion; features in the “later” group enter the roadmap in sequence.
The difference between 6 and 12 weeks comes from the size of the main flow and integrations. For large programs such as core system modernization or a platform affecting several user groups at once, this is the first-stage duration, not the whole program; subsequent stages launch in roadmap order.
Success criteria, MVP definition, and a prioritized roadmap are core outputs of product strategy work. Discovery content and deliverables are covered in What is a discovery sprint?, while this timeline, team and delivery conditions are explained on our custom software development page.
When scope grows: warning signs
Scope usually grows through small additions rather than one decision. These signs indicate an overloaded first release:
- The first release has more than one primary user group.
- The success criterion cannot fit in one sentence or has not been written.
- The “later” group is empty; everything is in the first release.
- The main flow cannot be tested until full integration is complete.
- Management reports are discussed before the main flow.
- The estimate exceeds 12 weeks, with nothing reaching real users during that time.
If several signs appear, the most effective way to reduce scope is narrowing the user group: dealers in one region instead of all dealers, or the most frequently ordered products instead of every product group.
Checklist: first-release scope
- The success criterion is one sentence, and its measurement is defined.
- One primary user group is selected.
- The main flow is written step by step, start to finish.
- Every feature belongs to “first release,” “later,” or “never.”
- Manually bridged integrations and their operators are identified.
- Deferred reports and exceptions are listed and communicated to users.
- The riskiest assumption and the first step to test it are identified.
- The users testing the prototype in week 3 are identified.
- The pilot group and target go-live week are defined.
Explore what this approach might look like in a dealer portal through our concept for the fictitious X Dağıtım, Dealer Ordering Platform product concept: the first stage opens a smart basket and account view to a pilot dealer group; wider dealer rollout and the field sales dashboard follow later. If you would like to define your product's first release together, you can arrange a conversation about a discovery sprint or pilot.