What Metrics to Put in a Product Manager Portfolio (and Which to Cut)
Metrics in a PM portfolio aren't there to prove a win — they're there to show judgment. How to pick a primary metric, handle projects without a clean number, and state baseline and attribution so a hiring manager actually trusts it.
What Metrics to Put in a Product Manager Portfolio (and Which to Cut)
Most PM portfolios either drown a case study in numbers or skip metrics entirely because the project didn't have a clean one. Both mistakes come from the same misunderstanding: metrics in a portfolio aren't there to prove the project succeeded. They're there to show you knew what to measure and why. Here's how to pick the numbers that actually do that.
1. A metric's job is to show judgment, not a win
Hiring managers reading a case study aren't scoring whether your project "won." They're checking whether you picked a metric that was actually connected to the problem, and whether you can explain the connection in one sentence. "Activation went from 24% to 31%" says nothing on its own — "we picked activation because the drop-off was happening before users saw any value, not after" is the part that shows product thinking. Lead with the reasoning, then the number, not the other way around.
2. One primary metric per case study, not five
A common instinct is to list every number that moved — activation, retention, NPS, revenue, support tickets — to look more rigorous. It reads the opposite way. A reader can't tell which number you actually cared about, which usually means you can't either. Pick the one metric that was your actual target for that project, mention one or two supporting signals if they add context, and leave the rest out. This is the same discipline our case study guide covers for structure generally — cutting is the hard part, not adding.
3. What to do when you don't have a clean number
Plenty of real PM work — a platform migration, a policy change, an internal tool — doesn't produce a tidy percentage. Don't force one. A qualitative outcome stated precisely ("cut the average support-ticket resolution time from three touches to one" or "the team shipped the next two features on this system without needing a rewrite") reads as more credible than a vague quantitative one ("significantly improved efficiency"). Vagueness is what gets a metric-free case study flagged as weak, not the absence of a percentage.
4. Baseline and timeframe, or the number means nothing
"Increased conversion by 20%" is a different claim if it's 20% relative (2.0% to 2.4%) versus 20% absolute (20% to 40%), and a different claim again over two weeks versus two quarters. Give the baseline, the result, and the window in the same sentence. It's one clause, and it's the difference between a number a hiring manager trusts and one they quietly discount.
5. Attribution: say what else was happening
If three other changes shipped in the same window, a single-cause claim won't survive a follow-up question — and in 2026, with AI-polished case studies now the norm, follow-up questions are exactly what a skeptical interviewer asks. One line of honest attribution ("this shipped alongside a pricing change, so we isolated impact using a holdout cohort" or "we can't fully isolate this from the seasonal lift, but the trend held after we controlled for it") does more for your credibility than a clean number ever will. See showing shipped work without the vibes for the same principle applied to AI-assisted projects specifically.
6. Metrics belong in the case study, not the summary line
Keep the resume-style compression — "improved activation 12%" — on your resume. In the portfolio case study itself, the metric should show up inside the narrative, next to the decision it was measuring, not as a headline stat sitting above the fold. A reader who has to hunt for what a number is attached to won't trust it. Our resume vs. portfolio guide covers this split in more depth: compress on the resume, expand in the case study.
Final thoughts
The strongest metric in a portfolio isn't the biggest one — it's the one paired with a clear reason you chose it, an honest baseline, and an acknowledgment of what else might have caused it. That combination is what separates a case study that reads as self-aware product judgment from one that reads as a highlight reel.
Start your portfolio here — free, no credit card.
FAQ
How many metrics should I include per case study in a PM portfolio?
One primary metric, plus at most one or two supporting signals if they add real context. Listing every number that moved makes it unclear which one you actually cared about, which reads as less rigorous, not more.
What if my project doesn't have a clean percentage or metric?
Don't force one. State the qualitative outcome precisely — for example, the number of steps a process went from and to — rather than reaching for a vague quantitative claim. A precise qualitative result is more credible than a soft quantitative one.
Should I mention other things that could have caused the result?
Yes, in one line. Acknowledging a confound — a pricing change in the same window, seasonal lift, a parallel launch — and how you accounted for it builds more credibility than presenting a single clean number as the whole story.
Where should metrics appear in a portfolio: the summary or the case study?
Inside the case study narrative, next to the decision the metric was measuring. Save the compressed, resume-style version ("improved activation 12%") for your resume; the portfolio's job is to show the reasoning behind the number, not just restate it.
Is it bad to include a baseline and timeframe with a metric?
It's the opposite — leaving them out is what makes a metric untrustworthy. "20% increase" means different things as a relative versus absolute change, and over different timeframes. Stating both in the same sentence as the result costs one clause and removes the ambiguity.
Related reading