How fast do users adopt new features?
№ 194 · Feature adoption curve
Definition
MetricCumulative share of eligible active users who have used a feature, plotted by weeks since release
Unit% of monthly active users, against weeks since general availability
Line of % of users who have used feature X over weeks-since-launch. Usually S-curve. Build by counting users with at least one feature interaction. Plateau height tells you the addressable user share. Example: 12% at W1, 32% at W4, 52% at W12 (plateau).
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| Under 10% and flat by week four. That flat point is the natural ceiling for the feature as shipped, and further promotion will not move it without a change to the feature or its placement. | 16.5% median core feature adoption with a 24.5% average across Userpilot's panel. A newly released feature commonly reaches 20% to 30% within its first 30 days before the curve begins to flatten. | Above 45% of active users for a core feature. For a newly released feature, 40% to 60% within 90 days where the rollout is actively driven with in-app guidance rather than announced once. |
B2B SaaS products instrumented in Userpilot, n=181 companies for the core feature adoption figure, drawn from a 547-company panel, 2024 to 2025 data. Denominator is monthly active users rather than total signups. Pendo's concentration data covers n=615 instrumented cloud software subscriptions. · The denominator moves this number more than the feature does. Measuring against active users rather than total signups is the convention in the sources above, and measuring against signups will halve or worse any figure quoted here. Pendo's widely cited 6.4% is a different metric, the share of features that generate 80% of clicks, and must not be read as an adoption rate. A power-user feature sitting at 10% to 20% can be entirely healthy if the intended audience is that size, so the benchmark to apply is the share of the intended audience, not the share of all users. Note also that Userpilot's published industry table transposes the MarTech and CRM labels between two sections of the same report, so the two values for those verticals should be treated as approximately 26% to 28% rather than precisely assigned.
| Category | Bottom 30% | Median | Top 30% |
|---|---|---|---|
| HR software, core feature adoption | Not separately published | 31%, the highest vertical in the panel | Above 45% |
| MarTech and CRM or sales software, core feature adoption | Not separately published | Roughly 26% to 28% across the two verticals, which the source table labels inconsistently | Above 45% |
| AI and ML software, core feature adoption | Not separately published | 24.8% | Above 45% |
| Healthcare, FinTech and insurance software, core feature adoption | Under 15% | 22.8% Healthcare, 22.6% FinTech and Insurance, the weakest verticals in the panel | Above 45% |
| Product-led vs sales-led motion | Not separately published | 26.7% sales-led against 24.3% product-led, a smaller gap than most teams assume | Above 45% for both |
When it looks bad
The curve jumps in launch week on the back of the announcement, then goes flat by week three well under the core-feature median, so the release note drove the trial and the feature drove nothing after it.
Week 1 at 11% of active users, week 3 at 14%, week 12 at 14.5%. Of those adopters, only 22% used the feature more than once, so the real sustained number is closer to 3%.
What to do about it
- Separate discovery from value in the diagnosis before changing anything. Count users who reached the feature surface against users who completed its core action: high traffic with low completion is a usability fix, low traffic is a placement or targeting fix, and the two need opposite work.
- Deploy in-app guidance rather than relying on release notes and email. Pendo found that customers deploying at least 25 in-app guides increased the number of features used daily by about 25%, which is the most reliable published magnitude for this lever.
- Put the feature inside the workflow where the user already is rather than behind a new navigation item. The Userpilot panel shows sales-led products at 26.7% against product-led at 24.3%, so the guided-hand advantage is only about two points, which means placement is doing more work than assistance.
- Set an explicit kill or invest gate at day 90 and judge against the intended audience, not the whole base. Adoption velocity that is flat by week four has found its ceiling, and continuing to promote past that point spends attention on a feature that has already answered the question.
Sources
- Userpilot Average core feature adoption 24.5% and median 16.5%. By industry: HR 31%, MarTech 27.9%, CRM and Sales 25.6%, AI and ML 24.8%, Healthcare 22.8%, FinTech and Insurance 22.6%. Sales-led companies at 26.7% against product-led at 24.3%. userpilot.com ↗
- Pendo Median feature adoption of 6.4% measured as the share of features driving 80% of clicks, with roughly 94% of features untouched. Reports that industry materially affects adoption, with Manufacturing and Consumer Goods highest and Media lower. pendo.io ↗
- Pendo 80% of features are rarely or never used, and deploying at least 25 in-app guides increased the number of features used daily by about 25%. The clearest published magnitude for an adoption intervention. pendo.io ↗
- Pendo Reports customer feature usage increases of up to 154% from replicating the behaviour of highly engaged users. A vendor-published outcome figure and treated as an upper bound rather than a benchmark. pendo.io ↗
- Amplitude Names feature adoption rate as one of the three core engagement measures alongside DAU/MAU and session frequency, and reports that 10% of products account for 79% of all user engagement. amplitude.com ↗
- Appcues A core workflow feature might reach 80% to 90% adoption while a power-user reporting feature sits at 15% to 20% and remains healthy, so the benchmark has to match the intended audience. appcues.com ↗
- Appcues Separates breadth, depth, time-to-adopt and duration of adoption, and notes that high first-month usage followed by low six-month usage is the pattern that predicts renewal risk. appcues.com ↗
- Product Growth Reports the commonly cited 90-day ramp in which a newly launched feature reaches 20% to 30% in its first 30 days and 40% to 60% within 90 days once the rollout is optimised, and puts the Userpilot top quartile above 45%. productgrowth.blog ↗
- feeqd Flat adoption velocity by week four indicates the natural ceiling has been reached, while velocity still growing at week eight indicates untapped demand. Warns against treating launch week as the verdict. feeqd.com ↗
- Vemetric Recommends weekly tracking for the first four to six weeks after launch and monthly thereafter, and puts low adoption below 20% with high adoption above 60% for core features. vemetric.com ↗