Impact Estimation in the Unit of Your Goal
Prioritize by estimating each option's impact in the exact same unit as your team goal, not an abstract score.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 3
- Confidence
- 90%
The third impact-first step: when prioritizing, express the 'impact' input to any scoring model (ICE, RICE, etc.) in the same unit of measure as your team's goal rather than an abstract number. This keeps prioritization honest, forces you to ask whether each item can actually move the goal, and surfaces the high-effort bet that's the only real path to the number.
Origin
Matt LeMay's third step in the impact-first method from 'Impact First Product Teams,' framed as the tactical payoff of setting a one-step-away goal.
Core principles
- 01Prioritization frameworks are worthless if 'impact' has no clear, specific definition
- 02Impact should be estimated as a function of your specific team goal (e.g. users converted), not an abstract 1-10 score
- 03It's called impact estimation, not impact-getting-perfect-number — approximation is fine and expected
- 04A proprietary scoring system can make you feel smart while severing the connection to why the work exists
How to run it
- 1
Anchor impact to the goal's unit
Before scoring anything, restate impact in your goal's unit. If the goal is converting N users, ask of each option: what is the most users this could convert?
Pro tip Do this even when it feels imprecise — 'let's just work through it, we don't have to be precise.'
Watch out Accepting an abstract impact score ('it's a lot of different things') hides whether the item can move the goal at all.
- 2
Estimate each option honestly, including the small ones
Walk each candidate: a landing page that only a few hundred people hit yields a few hundred; be willing to write down small numbers.
- 3
Weigh the high-effort bet that could reach the whole goal
Identify the option that, if it worked, would get you 'all the way there' even if it's high-effort, lower-certainty, and dependency-heavy — and seriously consider going for it.
Pro tip If it's the only option with any real chance of hitting the goal, the dependencies and cross-team conversations are worth it.
Watch out Don't reflexively deprioritize the ambitious redesign just because it requires working with marketing and other teams — the safe small items may not reach the goal at all.
In the wild
A PM on the multi-product conversion team built an ICE matrix scoring impact, certainty, and effort with abstract numbers. LeMay had them re-express impact as users converted. A landing page scored low (only a few hundred visitors). One high-effort, medium-certainty option they were reluctant to pursue was redesigning the entire onboarding so users were sorted into the right category on joining — which they admitted 'would get us all the way there if we did it right,' but required working with marketing and many other teams.
→ By scoring impact in the goal's unit, the team saw the onboarding redesign was the only item with a real chance of hitting the year-end conversion goal, justifying the effort and cross-team coordination rather than defaulting to safe, low-impact items.
Common mistakes
Scoring impact as an abstract number
Assigning impact a generic score derived from 'a lot of different things' means you can't tell whether an item actually contributes to your goal, defeating the purpose of prioritization.
Rewarding cleverness over connection
Inventing a proprietary scoring system with many intermediate steps feels like 'very terribly smart' work but risks losing the connection to why the work was undertaken in the first place.
Is it for you?
Best for
Teams that already have a specific one-step-away goal and are choosing what to build next.
Not ideal for
Teams that haven't yet defined a concrete, measurable goal — they must do that first before this step is usable.
From the transcript
“if you don't have a clear and specific sense of what impact means, you can't really prioritize effectively”
“let's try to express impact as a function of that goal. So, rather than an abstract score, what is the most number of users we…”
“It's called impact estimation, not impact getting perfect number”
“are we estimating and measuring impact in the same unit of measure as our goals?”
“if this thing could get us all the way there, if it's really the only thing we're looking at, which has really any chance of…”
From the episode
The one question that saves product careers
Matt LeMay