PLG Benchmarks 2026: Activation, Retention & Experimentation KPIs

Most PLG benchmark posts are recycled from three 2019 SaaS surveys and a Twitter thread. The numbers below come from operating growth programs — including experimentation across $2B+ in annual mortgage origination volume — and from what I see when I audit SaaS and FinTech funnels. Treat them as calibration ranges, not targets. A benchmark's only job is to tell you whether the problem you're staring at is normal or actually broken.

The benchmark table that matters

Activation is the benchmark everything else inherits

Signup-to-activation is the only metric on that table that mechanically constrains all the others. Retention, PQL conversion, and expansion are all multiples of the activated population, so a 10-point activation gain compounds through the entire funnel while a 10-point pricing-page gain does not.

Two rules for reading your own number honestly:

• Segment by acquisition channel. Paid traffic routinely activates at a third of the rate of referral traffic. A blended 38% can hide a 12% paid cohort that you're still buying.
• Define activation empirically. Compare retained and churned cohorts, find the earliest action that separates them, and use that. Teams that pick activation by committee inflate the number and then wonder why retention doesn't follow.

Retention curves beat retention points

A single D30 number is nearly useless without the curve shape. What you want to see is a curve that flattens — declining, then stabilizing at a durable plateau. A 30% D30 that keeps sliding to 8% by D90 is worse than a 24% D30 that flattens at 21%.

For FinTech, the equivalent of D30 is not daily usage — nobody logs into a mortgage product daily. It's return intent: repeat visits to the rate tool, saved scenarios, re-opened applications. Instrument those explicitly or you will report a retention story that has nothing to do with the actual purchase cycle.

Experimentation benchmarks: velocity, win rate, and effect size

Velocity is the metric growth teams underreport and overpromise. Ten tests a month is a mature, well-instrumented org with a dedicated analyst and pre-built surfaces. Four to eight is a healthy growth-stage team. In regulated funnels — where legal review, disclosure requirements, and lower page traffic all bite — two to three a month is a strong program, not a weak one.

Win rate is where teams misread themselves in the other direction. A 60% win rate means the hypotheses are too safe. Healthy programs sit in the 25–35% band, with a handful of large wins funding a long tail of flat results. Track cumulative validated lift per quarter, not the win/loss ledger.

On low-traffic pages — quote flows, application steps, pricing — sequential testing methods reach decisions materially faster than fixed-horizon tests, which is often the difference between running two experiments a quarter and running six.

FinTech-specific benchmarks most frameworks skip

Standard SaaS frameworks assume a feature-adoption model that doesn't exist in lending, insurance, or investing products. The benchmarks that actually diagnose a FinTech funnel:

• Step-level completion. Pages 3–5 of a typical application are where 60%+ of total drop-off concentrates. Anything below 70% completion per step in that band warrants a dedicated experiment queue.
• Verification pass rate. Document/identity verification success below 80% on first attempt is a product problem, not a compliance one.
• Quote-to-application rate. The single strongest leading indicator of funded volume, typically 8–20% depending on channel intent.
• Time-to-decision. Every additional hour between application submission and a decision measurably lowers close rate on competitive products.

What Series A/B investors actually check

Investors don't ask for your activation rate by name. They ask questions whose answers are your activation rate:

• Net revenue retention. Above 100% for SMB-focused PLG, above 110–120% for products with usage-based expansion.
• Organic share of signups. Above 50% suggests the product distributes itself; below 25% reads as a paid-acquisition business regardless of the growth rate.
• CAC payback. Under 12 months for PLG at Series A/B. Longer requires an unusually strong retention curve to defend.
• Cohort curve flattening. Shown as a chart, not a number. This is the single artifact that most often decides the meeting.
• Activation trend, not level. A 25% activation rate improving 3 points a quarter tells a better story than a static 40%.

If you're preparing a raise, build the cohort chart first and back into the rest. Every other metric on this page is a supporting argument for what that chart shows.

The next step

Pick one metric from the table where you can't confidently state your current number segmented by acquisition channel. That gap — not the benchmark itself — is your first project. Instrument it, get four weeks of clean data, then decide whether you have a benchmark problem or a measurement problem. Most teams discover it's the second.

Working on a PLG or FinTech growth challenge and want a second read on your numbers? Let's talk.