Feature Prioritization SOP

A feature prioritization SOP turns a long wish list into a short first version that customers will actually miss if it is gone, and it takes one to two weeks. You ask target customers how they feel about each feature using the Kano model, score the gaps between importance and satisfaction, then rank the survivors by reach, impact and effort. Every extra feature in version one delays the launch and dilutes the few that matter. This SOP gives you a written cut line you can defend to a co-founder, a developer or an investor.

What is feature prioritization and when should you use it?

Feature prioritization is the process of deciding which features go into the next build and which wait. This SOP combines four named methods, each doing one job:

• The Kano model (Noriaki Kano, 1984) sorts features into must-be, performance, attractive, indifferent and reverse, based on how customers feel when a feature is present and when it is absent.
• Opportunity scoring from Anthony Ulwick's outcome-driven method ranks features by how important a need is and how poorly current tools meet it.
• RICE (Reach, Impact, Confidence, Effort), published by Sean McBride at Intercom, adds the cost of building.
• MoSCoW (Must, Should, Could, Won't), created by Dai Clegg at Oracle, draws the final line.

Use this SOP:

• Before your first build, when the feature list has 10+ items and the budget covers five.
• When co-founders disagree on what matters and nobody has customer data.
• After a concierge or Wizard of Oz test, to turn manual steps into a build order.

This SOP ranks your own ideas with customer input. To compare your features against rivals, use the feature comparison SOP. To arrange features along the customer's workflow, use user story mapping.

What do you need before you start?

• A list of candidate features, each written as one plain sentence about what the customer can do.
• The jobs or outcomes behind them, from interviews or a customer journey map.
• 30 or more reachable target customers for a 10-minute survey.
• A survey tool such as Typeform or Tally, and a spreadsheet for scoring.
• A rough effort estimate for each feature from whoever will build it.

How do you run it, step by step?

1. Clean the feature list

Merge duplicates, split vague items ("better reporting") into concrete ones and drop anything that serves you more than the customer. Keep 10 to 20 features. Each one should describe a customer action: "Get a text when a waitlist spot opens."

2. Link each feature to a customer outcome

Next to each feature, write the outcome it serves and the interview quote that supports it. Features with no outcome and no quote go to a "parking" list. Expect this step to shrink the list noticeably.

3. Write the Kano survey

For each feature, ask two questions: "How would you feel if the product had this?" and "How would you feel if it did not?" Answers: I like it, I expect it, I'm neutral, I can live with it, I dislike it. Add two more per feature: how important the outcome is (1 to 10) and how satisfied they are with their current tool (1 to 10).

4. Send it to 30+ target customers

Recruit from your interview list, communities and a screened panel. Keep it under 10 minutes by limiting it to 15 features. Use the online survey validation SOP for recruiting and screening rules.

5. Classify and score

Use the standard Kano evaluation table to turn each answer pair into a category, then take the most common category per feature. Compute Ulwick's opportunity score: importance + max(importance minus satisfaction, 0). Ulwick treats scores above 10 as worth pursuing.

6. Rank the survivors with RICE

Take the must-be and performance features and the top attractive feature. Score each: Reach (customers affected per quarter), Impact (3, 2, 1, 0.5 or 0.25), Confidence (100%, 80% or 50%) and Effort (person-months). RICE = Reach × Impact × Confidence ÷ Effort.

7. Draw the cut line with MoSCoW

Must: every must-be feature, because customers punish their absence. Should: the top one or two performance features by RICE. Could: one attractive feature if it is cheap. Won't (this time): everything else, written down so it stops coming back in meetings.

Template you can copy

KANO SURVEY BLOCK (repeat per feature)
Feature: [customer can ...]
Q1 If the product had this, how would you feel?
   I like it / I expect it / I'm neutral / I can live with it / I dislike it
Q2 If the product did NOT have this, how would you feel?
   I like it / I expect it / I'm neutral / I can live with it / I dislike it
Q3 How important is [outcome] to you? 1-10
Q4 How satisfied are you with how you handle it today? 1-10

SCORING SHEET
Feature | Outcome | Supporting quote | Kano category (mode) | % agreeing |
Importance avg | Satisfaction avg | Opportunity score |
Reach | Impact | Confidence | Effort | RICE | MoSCoW

Worked example: booking software for dance studios

This example is illustrative. A founder building booking software for small dance studios has 14 candidate features and budget for about six. She surveys 36 studio owners.

The waitlist scored importance 8.6 and satisfaction 3.0, giving an opportunity score of 14.2. Its RICE score came from reach 30 studios per quarter, impact 2, confidence 80% and effort 1 person-month. The first version shipped with six features instead of 14, and the social feed, the founder's own favourite, went to the Won't list.

What should you measure, and when do you stop?

Stop when the cut line is written and agreed. If categories are spread evenly, check whether several segments are mixed together in your sample: split them by profile and recount. Before you build, check the top feature with a fake door test if you have traffic, or with prototype testing if you don't.

Which mistakes should you avoid?

• Asking "Which features do you want?" Customers say yes to everything. The paired Kano questions force a trade-off.
• Skipping must-be features because they are boring. Payments and booking excite nobody, and their absence kills the product.
• Letting effort decide alone. Cheap features nobody needs still waste time.
• Mixing segments. Studio owners and their students value opposite things. Survey one group at a time.
• Never revisiting the list. Re-run the scoring after the first version is in customers' hands.

How can AI speed this up?

An AI assistant can turn interview notes into a clean feature list, write the Kano questions for each feature and classify answer pairs with the evaluation table once you paste in the responses. Ask it to flag features with no supporting quote. The human check: you set the RICE effort numbers with your developer, and you make the final MoSCoW call, because a spreadsheet can't weigh strategy. The F/MS AI co-founder can work through the scoring with you step by step, and the startup tool comparisons help you pick software for tracking the list.

Where does this SOP fit in your validation plan?

Feature prioritization sits in the middle of Solution Validation, after you know what problem you solve and before the first build.

1. Before: arrange the customer's workflow with user story mapping.
2. After: test the top feature with a fake door test or a prototype, and confirm it can be built with a technical feasibility assessment.
3. Later: compare your feature set with rivals using the feature comparison SOP.

All Solution Validation SOPs are listed in one place.

FAQ

Which feature prioritization framework is best for a startup?

Use Kano to learn what customers expect, RICE to add the cost of building and MoSCoW to cut. Each covers a gap the others leave open.

How many people do I need for a Kano survey?

30 or more from one segment gives you stable categories for the top features. Below 20, treat the results as a conversation starter.

What is a good RICE score?

RICE scores only mean something compared with each other inside one list. Look for a clear gap between the top few and the rest.

How many features should my first version have?

As few as deliver the main outcome: every must-be feature plus the one or two that customers rate highest by opportunity score. This SOP sets five to eight as the target range for a first build.