There is a particular comfort in opening a prioritisation spreadsheet. The chaos receives columns. Opinions receive decimals. By cell G27, an argument about the future has become a sortable table.

This is useful. It is also not a decision.

A framework can structure judgment, expose assumptions and make inconsistent reasoning easier to spot. It cannot tell you what your strategy should value, how much trust a customer promise deserves, or whether the team should accept short-term pain to create a new advantage.

Those choices entered the spreadsheet before the formula began.

What frameworks genuinely help with

Take RICE: reach, impact, confidence and effort. Sean McBride’s original explanation is careful about what the framework does. It creates a consistent way to compare ideas and makes teams state estimates they might otherwise leave vague.

That is valuable. “I love this idea” has to become “I believe it reaches this group, changes this behaviour and costs roughly this much.” Now colleagues can question the belief rather than the enthusiasm.

Frameworks are especially useful when:

  • many options are similar enough to compare;
  • the team uses shared definitions;
  • confidence is expressed honestly;
  • the score begins a discussion rather than ending one.

They improve the hygiene of a decision. Hygiene matters. It is not strategy.

The decimal-point costume

False precision arrives when uncertain inputs produce a very certain-looking output.

An impact score of 2.0 and an effort estimate of three person-months are not facts simply because they fit inside cells. They are judgments, often made with different evidence by people facing different incentives.

If one team interprets “high impact” as revenue this quarter and another interprets it as customer value over two years, the combined ranking is numerically elegant and conceptually broken.

The right response is not to ban scoring. It is to annotate it. What evidence supports the estimate? What would change it? Which number is carrying the most uncertainty? A score without its assumptions is a rumour wearing business casual.

Strategy must enter before scoring

Suppose one initiative improves the current product for a large audience. Another tests whether the company can serve a smaller, strategically important segment. The first will often win a reach-based model.

That may be correct. It may also quietly prevent the organisation from ever making an exploratory bet.

Before scoring, name the decision policy. Are we maximising near-term customer value? Protecting reliability? Learning about a new market? Fulfilling a contractual obligation? The policy determines which dimensions matter and which options should not share one ranking at all.

You can also reserve capacity by type of work rather than forcing maintenance, experiments and growth bets into a single beauty contest.

The politics do not disappear

Frameworks are sometimes introduced to “remove emotion” or “take politics out of prioritisation.” Neither will happen. Product decisions distribute money, attention, risk and status. That is political in the ordinary organisational sense.

The healthier goal is visible politics. Let people see why a strategic exception outranked a higher score. Record the trade-off. Revisit whether the promised consequence occurred.

If leadership overrides the model, the model did not fail. It revealed where authority entered the decision.

The framework can reveal strategic snacking

Des Traynor’s warning about “snacking” describes the attraction of low-effort, visible work that feels productive while larger strategic problems remain untouched. A scoring model can either expose or accelerate this behaviour.

If effort is estimated confidently while impact remains vague, small initiatives repeatedly rise to the top. The organisation becomes efficient at finishing things that do not materially change its position. Every item is rational in isolation; the portfolio is incoherent.

Review the ranked list as a portfolio, not ten independent winners. Does it contain enough work connected to the strategic bet? Is reliability receiving the investment needed to protect trust? Are there deliberate learning bets whose uncertainty would otherwise punish their score? A portfolio needs balance that no single-item formula can produce.

Separate eligibility from ranking

Before applying a framework, decide which ideas are eligible. Remove work that conflicts with strategy, lacks a clear problem or violates a non-negotiable constraint. Route legal obligations, critical incidents and small operational maintenance through policies designed for those classes of work.

Then compare the remaining options. This prevents a formula from pretending that a security remediation and a speculative growth feature are interchangeable because both have numbers.

Finally, record exceptions. “Leadership chose B despite its lower score because it tests our entry into segment Y” is more useful than quietly changing the inputs until B wins.

Recalibrate after reality arrives

Return to old estimates after delivery. Which teams systematically underestimate effort? Which impact assumptions survive contact with customers? Does “high confidence” mean the same thing across the organisation?

The framework becomes more valuable when it learns. Otherwise it is a permanent formula built from temporary guesses.

Questions to take back to your team

  • Which strategic choice determines what “impact” means in your scoring model?
  • Where are uncertain estimates wearing suspiciously precise decimal points?
  • Which important work loses because your framework rewards small, easily estimated initiatives?
  • What exception did leadership make recently, and was the real reason recorded honestly?
  • Have you recalibrated old reach, impact, confidence and effort estimates against what actually happened?

Look at the top five ranked items as one portfolio. Do they tell a coherent strategic story, or have you built a tasting menu of organisational snacks?

My take

Use a framework as a set of questions, not an answer machine.

After the ranking appears, ask:

  • Which result surprises us?
  • Which input would we least want to defend publicly?
  • What strategic obligation is missing from the model?
  • What lower-ranked item might still deserve an explicit exception?

The framework earns its keep when it makes judgment discussable. The difficult decision remains yours. That is not a flaw in the tool. It is the job the tool was never qualified to take.