Scoring Your Options: Cost-Benefit, Weighted Criteria, and RICE

Urgent vs Important helped you decide what deserves your attention at all. But once several real options survive that filter — three vendors, four features, two candidates, five projects and only budget for two — you still have to choose between them. And here gut feel starts to wobble. The options look similar. Each has a champion who argues loudly. You keep circling the same conversation without landing.

This page gives you three tools for exactly that moment: cost-benefit thinking, weighted criteria, and RICE-style scoring. They get progressively more structured, and they suit different decisions. The point of all three is not to compute an answer — it’s to force a clear conversation about what you actually care about, so the trade-offs stop hiding.

The baseline: cost-benefit thinking

Before any fancy method, there is one habit that carries most of the weight: for each option, write down what it costs you and what it returns — then compare like with like.

That sounds obvious, and yet most stuck decisions are stuck precisely because nobody has done it. People argue about options while holding different, unspoken pictures of what each one costs. One person is thinking about money; another about time; a third about the risk of annoying a customer. They’re not disagreeing — they’re comparing different things and don’t know it.

Cost-benefit thinking just makes those two columns explicit:

Option: Switch the ward to the new scheduling system

COSTS                              BENEFITS
- ~3 weeks of setup + training     - Nurses stop double-booking rooms
- $18k first-year licence          - ~4 hours/week of admin saved
- Risk: staff resist the change    - Fewer last-minute shift gaps
- One month of lower throughput    - Clearer picture of who's on call

Two things make this useful rather than a formality. First, count all the costs, not just the money — time, attention, risk, morale, the opportunity you give up by not doing something else. The most expensive cost is usually the invisible one. Second, be honest about which benefits are real and which are hoped-for. “Saves 4 hours a week” is checkable; “improves culture” is a wish until you say how you’d know.

Cost-benefit is enough on its own when a decision has one dominant dimension — mostly about money, or mostly about time. When the columns clearly favor one option, stop. You don’t need a heavier tool.

The common trap — counting only the costs you can measure

Because money and hours are easy to put numbers on, they crowd out costs that are harder to quantify but often larger: the risk of a change failing, the team’s goodwill, the customer relationship, the six other things you can’t do while you do this one. A decision that looks great on the measurable costs and quietly ignores the rest is not rigorous — it’s just precise about the wrong things. Name the fuzzy costs even if you can only rate them “high / medium / low.”

When one dimension isn’t enough: weighted criteria

Some decisions genuinely pull in several directions at once. Choosing a supplier, a job candidate, a location for a new cafe, a tool the whole team will live in — these are cheap on one axis and expensive on another, strong here and weak there. Gut feel struggles because it can’t hold five dimensions steady at the same time; it grabs whichever one felt loudest this morning.

Weighted criteria fixes that with three steps:

1. Pick the factors that actually matter. Not everything you could measure — the handful that would genuinely change your mind. For a cafe location: foot traffic, rent, kitchen size, lease flexibility, parking.
2. Weight them. Not all factors matter equally. Give each a weight (say, out of 100, summing to 100) that reflects how much you care. This is the real work — arguing about the weights is arguing about your priorities, which is the conversation you needed to have anyway.
3. Score each option on each factor, then multiply score × weight and total it.

Factor            Weight   Site A        Site B        Site C
                           score  ×w      score  ×w      score  ×w
Foot traffic        35       8   280       5   175       9   315
Rent (lower=better) 25       4   100       9   225       3    75
Kitchen size        20       7   140       6   120       8   160
Lease flexibility   12       6    72       8    96       5    60
Parking              8       9    72       4    32       6    48
                    ---                   ---           ---
TOTAL              100             664           648           658

Now look at what just happened. On gut feel, Site B (cheap rent) or Site C (great everything) probably felt like the winner. The table says all three are within a whisker of each other — and Site A edges ahead because foot traffic, which you said matters most, quietly outweighs its higher rent. The method didn’t decide for you; it showed you that your stated priorities and your gut were pointing in slightly different directions. That gap is the whole payoff.

In practice — the argument moves to where it belongs

A team spent a week deadlocked over two software tools, each camp certain the other was blind. When they built a weighted table, the scores came out nearly tied — and they realized the real disagreement wasn’t about the tools at all. It was about the weights: one camp valued ease of setup, the other valued long-term flexibility. Once that was named, the decision took twenty minutes, because it was finally the right argument. The table’s job was to relocate the disagreement, not resolve it.

How fine-grained should the scores be?

Resist the urge for precision you don’t have. A 1–5 or 1–10 scale, filled in by the people who actually know each option, is plenty. If you find yourself debating whether something is a 7 or a 7.5, you’ve left the zone where the numbers mean anything. The scale exists to sort options into “clearly better,” “roughly equal,” and “clearly worse” — not to rank them to the decimal.

Ranking a whole backlog: RICE-style scoring

The first two tools compare a small set of options head to head. But managers often face a different shape of problem: a long list of things all worth doing — feature requests, improvement ideas, projects, fixes — and far too little time to do them all. You don’t need to pick the single best; you need to rank the pile so you work the top of it first.

RICE is a simple, popular way to do that. For each item you estimate four things and combine them:

• Reach — how many people (or cases, or customers) does this affect in a given period?
• Impact — how much does it help each one? (A rough scale: 3 = massive, 2 = high, 1 = medium, 0.5 = low, 0.25 = minimal.)
• Confidence — how sure are you of the reach and impact estimates? (100% = solid data, 80% = reasonable, 50% = a guess.)
• Effort — how much work will it take? (Usually person-weeks or person-days.)

Then:

        Reach × Impact × Confidence
RICE =  ---------------------------
                 Effort

The logic is intuitive once you see it: things that help more people, more deeply, more certainly, for less work float to the top. The division by effort is what makes it useful — it stops a shiny, expensive idea from beating three cheap, high-value ones.

Item                     Reach  Impact  Conf.  Effort   RICE
Fix the checkout error    2000    2.0    0.9      2      1800
Redesign the dashboard     800    1.0    0.5      8        50
Add dark mode             1500    0.5    0.8      3       200
Speed up search            900    2.0    0.8      1      1440

Ranked by RICE, the checkout fix and the search speed-up clearly lead — and the dashboard redesign, which probably felt like the exciting big project, lands dead last because it’s low-confidence, medium-impact, and expensive. That’s RICE doing its one honest job: dragging the seductive-but-costly item down the list where it belongs.

RICE fits when you have many roughly-comparable items and need a defensible order. It’s overkill for a two-option, high-stakes call — use cost-benefit or weighted criteria there.

Case study — the six-page memo and the number that had to be argued

Amazon, since roughly the early 2010s, famously bans slide decks in its big decision meetings in favor of a densely-written six-page memo, read in silence at the start of the meeting. The reason connects directly to scoring: a bullet on a slide can hide a shaky assumption, but a sentence that has to justify a number cannot. The discipline isn’t the format — it’s that every figure has to survive being written down in prose and read by skeptics. A RICE score or a weighted table works the same way only if the numbers behind it can be defended out loud. The artifact is worthless; the forced argument is everything.

The number is a thinking aid, not a verdict

Here is the mistake that ruins all three tools: treating the total as the decision. It isn’t. Every one of these methods is a machine for turning fuzzy priorities into an explicit, arguable structure — and the value is entirely in building it, not in obeying the output.

So when the numbers come out, do the opposite of surrendering to them. Look at the winner and ask, “Does this feel right?” If it does, good — you’ve confirmed your instinct and now understand why. If it doesn’t, that mismatch is a gift: either your gut is catching something you left out of the model (a missing criterion, a cost you underweighted), or the numbers are catching a bias in your gut. Both are worth an extra ten minutes. The score’s best moment is often the one where it surprises you.

The common trap — false precision that launders a gut decision

The dangerous failure mode isn’t messy numbers — it’s tidy ones that were quietly reverse-engineered. Someone already wants Option B, so they nudge a weight here, inflate a confidence there, until the table dutifully outputs B. Now the gut call wears a lab coat: “the analysis says B.” This is worse than no analysis at all, because it shuts down the very argument the tool was meant to open. Guard against it two ways. First, assign the weights before you score the options, so you can’t tune them to fit a favorite. Second, run a quick sanity check: change one debatable weight or estimate and see if the ranking flips. If a small, arguable tweak reorders everything, your “answer” is really a coin flip wearing a spreadsheet, and you should say so out loud rather than hide behind the total.

The evidence — structure beats intuition on repeatable calls

Research on decision-making generally finds that simple, consistent scoring rules tend to outperform expert intuition on the kind of decisions you make over and over with comparable inputs — hiring, lending, forecasting — largely because a rule doesn’t get tired, moody, or swayed by the last vivid story. The catch, hedged appropriately: this holds for repeatable decisions with decent inputs, not rare one-offs, and only when the criteria themselves are sound. A scoring model is only ever as good as the priorities you built into it. Treat it as a reliable co-pilot for recurring choices, not an oracle for singular, high-uncertainty bets.

Which tool for which decision?

Situation                                   Reach for...
One dominant dimension (mostly cost/time)   Cost-benefit
A few options, several dimensions,          Weighted criteria
  genuine trade-offs
A long backlog to rank by value-for-effort  RICE-style scoring
High stakes, deep uncertainty, one-off      A lighter touch — see next page

Match the weight of the tool to the weight of the decision. Reaching for a five-criterion weighted table to pick where to eat lunch is its own kind of failure — analysis where instinct would do. The next page, Deciding Under Uncertainty, takes on the decisions these tools can’t clean up: the ones where the inputs themselves are guesses.

Try this

Take one real decision you’re currently sitting on — a purchase, a hire, a project choice, or a backlog you can’t order. Pick the tool that fits its shape (cost-benefit, weighted criteria, or RICE) and actually build the artifact this week, on paper, with the people who know the options. Then do the sanity check: change one debatable weight or estimate and see whether the ranking holds. Notice which part of the exercise was most useful — you’ll usually find it was the argument about priorities, not the final number.

Reflect

1. Think of a recent group decision that circled for too long. Were people secretly comparing different costs and benefits without realizing it?
2. For a choice you face now, what are the three or four factors that would genuinely change your mind — and how would you weight them?
3. When have you seen a number used to launder a decision someone had already made? What tipped you off?
4. Which of your recurring decisions (hiring, prioritizing, buying) could benefit from a consistent scoring rule rather than fresh gut feel each time?
5. Recall a time the “right” option on paper felt wrong. Was your gut catching a missing criterion, or was the paper catching your bias?

Show reflections

1. Most stuck decisions are stuck here. The tell is people talking past each other with equal confidence — usually a sign they hold different unspoken pictures of what each option costs. Naming the two columns out loud often unsticks it faster than any vote.
2. The useful discipline is picking few factors and being honest about weights. If you can’t defend why one factor outweighs another, that’s the priority conversation you’ve been avoiding — have it before you score anything.
3. Good answers spot reverse-engineered tidiness: weights or confidences that happen to favor whoever built the model. The defense is assigning weights before scoring, and welcoming the sanity check that tries to flip the result.
4. Repeatable decisions with comparable inputs are where a consistent rule earns its keep — it removes mood, fatigue, and recency bias. One-off, high-uncertainty bets are not; a rule there gives false comfort. Sorting your decisions into those two buckets is the real skill.
5. Both are possible and both are valuable. If your gut found a missing criterion, add it and re-score. If the paper caught a bias, that’s exactly what the tool is for. Either way the mismatch, not the agreement, is where the learning lives.