Weighted scoring models: turning gut feel into a defensible decision

"We should do this one first." Ask why, and you'll often get a confident answer that turns out to mean "I feel strongly about it." That's not a knock on instinct — experienced people have good instincts. The problem is that instinct doesn't scale, doesn't transfer, and doesn't hold up when two experienced people disagree.

A weighted scoring model turns that instinct into something you can write down, apply consistently, and defend when someone asks why this and not that. Here's how to build one that actually helps.

What a weighted scoring model is

It's simple. You pick a handful of criteria that matter for the decision, assign each a weight reflecting how much it matters, rate each request on every criterion, and combine the ratings into a single comparable score:

Score = Σ (rating ÷ max rating × weight)

The output is a starting point for judgment: a way to see the whole field on the same axes before you bring human context back in.

Choose criteria that actually discriminate

The most common mistake is picking criteria everyone scores the same. If every request is "high impact," impact isn't telling you anything. Good criteria spread requests out.

For project intake, a durable set is usually some combination of:

• Impact — how much this moves an outcome you care about.
• Effort — the cost to deliver, often the strongest differentiator.
• Risk — what happens if it goes wrong, or if you don't do it.
• Strategic fit — alignment with where the organization is headed.

Four to six criteria is the sweet spot. Fewer, and you lose nuance; more, and every request regresses toward the middle as the weights get diluted.

Set weights deliberately — they encode your strategy

Weights are where your priorities become explicit, so treat them as a real decision rather than round numbers. If impact carries 40% and effort 20%, you've just told everyone that you'll accept expensive work when the payoff is big. If you flip them, you've said you optimize for quick wins. Neither is wrong — but the weights should reflect the choice you actually mean to make.

Two practical rules:

1. Weights should sum to 100%. It keeps scores comparable and forces trade-offs — you can't raise one without lowering another.
2. Revisit them quarterly, not weekly. Weights encode strategy; if they change constantly, they're not weights, they're moods.

Use a rating scale people can apply consistently

A five-point scale is enough for almost everyone. The key isn't the number of points — it's that each point means the same thing to every reviewer. "4 out of 5 on impact" should have a shared definition, even a rough one, or your scores just launder individual bias through a spreadsheet.

Write a one-line anchor for the top and bottom of each scale. It takes ten minutes and does more for consistency than any amount of calibration meetings.

Watch for the traps

Weighted scoring fails in predictable ways. Know them going in:

• False precision. A score of 7.4 vs 7.1 is not a meaningful gap. Treat scores as bands — top tier, middle, bottom — not a strict ranking.
• Gaming. When people know the model, they'll describe requests to score well. Transparent criteria plus a human review step keeps this honest.
• Set-and-forget. A model that never changes slowly drifts out of step with reality. Review the criteria themselves once or twice a year.
• Letting the number decide. The score orders the conversation; it doesn't end it. Keep a deliberate step where a human can override with a stated reason.

The score orders the conversation

Used well, a weighted scoring model moves the argument from which request wins to what we value and how much. That's a far more productive disagreement to have, and once it's settled, most of the individual decisions fall out of it.

That's the model Admisio builds in: you define the criteria and weights that fit your organization, reviewers score against a shared framework, and every request gets a transparent, comparable total — with the final call still firmly in human hands. Pair it with a clean intake process (see building intake without bottlenecks) and prioritization stops being the meeting everyone dreads.