Prioritise the backlog of a tuition-class booking app with a scoring grid

A small team cannot build twelve features at once. The scoring grid replaces 'who argued loudest' with a fair, repeatable judgement, so the real work is fixing the scale before you score and then testing it against someone else's view.

• Who uses the app: students, parents, tutors, the class owner? Whose problem matters most right now?
• What does a high score mean on reach, on impact and on effort, in one sentence each?
• How will you combine the three scores into one ranking, and why that way?
• Which feature are you secretly fond of, and how will you stop that bending the scores?

A sheet with twelve features, a written scale for reach, impact and effort, a formula that ranks them, a comparison with a friend's scores on three features, and one paragraph defending the top three and the one you would cut.

Scoring first and writing the scale afterwards. The scale then quietly fits whatever you already wanted to build.

• Twelve feature ideas you write yourself, such as reminders, reschedule, parent view and fee receipts
• A short talk with a tutor or a parent about what annoys them about booking classes today, with their consent
• A friend willing to score three features on your scale without seeing your scores

The features and scores are yours and should be labelled as a practice exercise. AI must not claim how many users would use a feature or quote usage data for any real app.

1. 3.1 Write the scoring scaleCLAUDE CHAT

   Why this tool: Claude can spot vague scale wording before you start, while changing it is still cheap.

   Write a one-line meaning for each score on reach, impact and effort. Ask Claude where two people could read a line differently, and tighten it.

   Prompt to copy · fill in the [brackets]Copy

   Here is my scoring scale for features in a tuition booking app:
   [paste your scale]
   For each line, tell me how two different people might score the same feature differently because the wording is unclear. Suggest one question I could add to make it sharper. Do not score any features.

2. 3.2 Score every feature the same way with a SkillCLAUDE SKILLS

   Why this tool: A saved Skill applies your exact scale to each feature in turn, which exposes features you scored generously by hand.

   Save your scale as a Skill that asks for a score and a one-line reason per criterion. Run it on all twelve features, then compare with your own scores and decide each difference yourself.

   Prompt to copy · fill in the [brackets]Copy

   Score this feature for a tuition booking app using only the scale below. For reach, impact and effort, give a score and one line of reason. If the scale does not decide it, say 'unclear' instead of guessing.
   Scale: [paste your scale]
   Feature: [one feature from your list]

   Before you let it act: Read the saved instructions before running: they must contain your scale word for word and must ask for reasons. The Skill suggests scores; you decide the final ones in your sheet.

3. 3.3 Rank in the sheetGOOGLE SHEETS

   Why this tool: A formula keeps the ranking honest: change a score and the order updates.

   Put each score in its own column, add a combined score with a simple formula, and sort. Add your friend's three scores in a separate column.

4. 3.4 Ask what you missedCLAUDE CHAT

   Why this tool: Claude can suggest features and risks you did not list, which you then judge against what a tutor or parent would want.

   Paste your twelve features and ranking. Keep only suggestions your tutor or parent conversation supports.

• Your scale was written before any feature was scored
• Recompute two combined scores by hand and confirm the sheet agrees
• Where your friend scored differently, you wrote down why
• The feature you cut is cut for a reason on your scale, not a feeling
• Recompute two numbers by hand
• Check the row count before and after cleaning

AI tends to rate every feature as high impact because each one sounds useful. If most features score near the top, your scale is too soft, not your backlog too good.

• Twelve features with three scores each
• Written scale and ranking formula
• Friend comparison on three features
• One paragraph on the top three and the cut

Show the ranked table first, then the scale in a small box, then your paragraph. Highlight where your friend disagreed.

Explain one feature that moved a lot once you fixed the scale, and what that taught you about prioritising.