PLG Metrics That Actually Matter (And the Ones You Can Drop)

The dashboard problem

Most PLG dashboards have the same flaw: twenty tiles, three of them worth checking twice a week. Signups, DAU, WAU, MAU, trial starts, feature adoption, NPS, session length — all tracked, most of it ignored the moment it stops being novel. The team that built the dashboard moves on. Six months later nobody remembers why "weekly active teams" has its own chart or what decision it was supposed to drive.

The fix isn't more metrics. It's cutting the list down to the ones that actually change what someone does on Monday morning, and being honest about which popular PLG metric is mostly vanity.

Start from the decision, not the number

Before tracking anything, write down the decision a metric is supposed to inform. If a number can go up or down and nobody's next action changes either way, it's not a metric — it's trivia.

• Activation rate should change whether you fix onboarding this sprint or not.
• PQL-to-paid conversion should change how sales triages the queue.
• Expansion rate by cohort should change which accounts get a check-in call.

If a metric doesn't map to a specific "then we do X," drop it from the weekly review. Keep it in a backup dashboard if someone wants it, but don't let it compete for attention with the ones that matter.

The core set: six metrics, in order

1. Activation rate

The percentage of signups who reach the moment your product actually becomes useful to them — not "logged in," the real activation event specific to your product (first report generated, first integration connected, first team invite sent).

Pick one activation event and defend it. Teams that track three "activation-ish" events end up with three slightly different activation rates and no clear read on whether onboarding is improving. Segment this by ICP fit: a 40% blended activation rate can hide a 70% rate among your best-fit users and a 15% rate among tire-kickers, and those two populations need completely different fixes.

2. Time to first value (TTFV)

How long from signup to that same activation event. Shorter is usually better, but "usually" is doing work there — a product that lets someone activate in 90 seconds by skipping setup might just be deferring the real friction to week two. Track TTFV alongside week-four retention, not alone.

3. PQL-to-paid conversion rate

Of the users your scoring model flags as product-qualified leads, what share convert to a paid plan within a defined window (30, 60, 90 days — pick one and stick with it so the number is comparable over time). This is the metric that tells you whether your PQL scoring model is actually predictive or just generating a list that looks busy.

If PQL-to-paid conversion is flat or declining while PQL volume climbs, the scoring criteria have drifted from what actually predicts revenue. Revisit the model before adding more sales headcount to work the queue.

4. Net revenue retention (NRR) by ICP segment

Blended NRR hides more than it reveals. A company at 105% blended NRR could be losing money on half its ICP segments and masking it with expansion in one segment that happens to be growing fast right now. Break NRR out by ICP tier and watch for segments trending down even while the topline number looks fine — that's usually the earliest warning a segment is drifting away from product fit before churn shows up in the aggregate number.

5. Expansion signal response rate

Of the accounts flagged with an expansion signal — new teammate joined, usage crossed a seat or tier threshold, a buying-committee member showed up — what share got a human follow-up within your target window (same day, same week, whatever your motion supports). A high volume of expansion signals with a low response rate means the detection logic works and the operational follow-through doesn't. That's a staffing and process problem, not a data problem, and no amount of better signal detection fixes it.

6. Time-in-stage for the sales-assist queue

For teams running a sales-assist motion on top of self-serve, track how long a qualified lead sits in queue before a rep touches it. This number quietly determines whether your enrichment and scoring investment pays off. A perfectly scored PQL that sits untouched for six days converts at roughly the same rate as an unscored one.

Metrics worth tracking but not obsessing over

DAU/MAU ratio. Useful as a rough stickiness signal, close to meaningless as a standalone KPI. A project-management tool used twice a week by a happy customer will always look worse on this ratio than a habit-forming consumer app, and comparing the two tells you nothing.

NPS. Fine as a long-horizon trend line, bad as a weekly decision input. Sample sizes are usually too small and response bias too large for week-over-week moves to mean anything. Check it quarterly, not every Monday.

Feature adoption breadth. Tracking how many features a user has touched sounds like a maturity signal, but a user who uses three features deeply is often healthier than one who's clicked through ten shallowly. Pair this with retention data before drawing conclusions, never read it alone.

Raw signup volume. Growth teams love this number because it always goes in a direction, but unsegmented signup volume mixes serious buyers with students, competitors doing research, and bots. Pair it with a view of who's actually signing up before treating a spike as good news.

Why most PLG metrics stacks fall apart at scale

The usual failure mode isn't picking bad metrics. It's picking reasonable metrics, then losing the ability to segment them as the user base diversifies. A team of 20 can eyeball which signups matter. A team processing a thousand signups a week can't, and blended metrics across a mixed population start lying by averaging good and bad together.

The fix is enrichment at the point of signup, not after the fact. If you know company size, role, and ICP fit the moment someone signs up, every metric above can be segmented from day one instead of retrofitted months later when someone finally asks "wait, is this 40% activation rate good or not?" and the honest answer is "we don't know, we never broke it down."

A short framework for building your own stack

1. List the decisions your team actually makes weekly. Onboarding changes, sales queue prioritization, expansion outreach, pricing experiments. Each decision needs exactly one or two metrics behind it.
2. Pick one activation event and one conversion window. Write both down somewhere permanent. Changing either without flagging it breaks every historical comparison.
3. Segment everything by ICP fit before trusting a blended number. A metric that looks stable in aggregate can be masking a segment in decline.
4. Assign an owner to each metric who has to explain a bad trend out loud. A metric with no owner degrades into decoration within a quarter.
5. Kill anything that hasn't changed a decision in two months. Archive it, don't delete the data, just stop reviewing it weekly.

Mistakes to avoid

Tracking activation before defining it precisely. "Activated" means different things to different people on the same team until someone writes the exact event down. Get this in writing before building a single chart.

Averaging across ICP segments. A blended conversion rate across your best-fit and worst-fit signups tells you almost nothing about either one. Segment first, then look at the blend if you still want it.

Changing the activation event or conversion window without a note. Six months of trend data becomes useless the moment someone quietly redefines what counts, and nobody remembers to mention it in the next QBR.

Confusing more signups with more pipeline. Volume and quality move independently. A 20% jump in signups from a content campaign that skews toward the wrong company size can shrink your effective pipeline while the topline chart looks great.

Where enrichment fits

None of the six core metrics above work without knowing who's actually signing up. Groful enriches every signup in real time — company, role, size, teammates already in the account — and scores it against your ICP so every metric on this list can be segmented from day one instead of patched together later from a CRM export. See the signup enrichment solution for how the data gets captured, check what this looks like for a growth manager's weekly review, review pricing, or get in touch to work through which of these metrics would actually move for your product.