How to Drive Feature Adoption in SaaS: Getting Users to Use What They Already Pay For

This article is for informational purposes only and does not constitute financial, tax, or legal advice. Consult a qualified professional for guidance specific to your situation.

Reviewed for financial accuracy by the Startup Finance Guide editorial team. Our editors cross-reference all claims against platform documentation, pricing pages, and primary regulatory sources. Last reviewed: October 12, 2026.

To get more users to adopt features they already pay for, join your entitlement data (what each account's plan includes) to your usage data (what each person in that account did), find the paid-for features with no meaningful use, work out which of four causes is behind each gap (discovery, permission or setup, relevance, habit), and apply the fix that matches the cause. The gap is usually large. Pendo's analysis of its own customers' usage data found that 80 percent of features in the average software product are rarely or never used, and Userpilot's product metrics benchmark report (547 SaaS companies, by its own count) puts the average core feature adoption rate at 24.5%.

The tools come second. The seven covered here do different jobs: Pendo, Appcues, Chameleon and Userpilot put guidance inside the product, Intercom adds tours and checklists to a helpdesk, Gainsight CS routes account-level adoption risk to customer success managers, and Userlens (through its Lumi agent) sends personal email and Slack nudges written from each user's behavior. Published entry prices run from $0 (Pendo Free up to 500 monthly active users, Userlens Free up to 100 profiles) through $100 a month (Userlens Starter), $279 (Chameleon Startup) and $299 (Userpilot Starter), while Gainsight and Pendo's paid plans publish no price at all. None of them fixes a gap you have not diagnosed.

Key takeaways

• The number to manage is the entitlement versus usage gap: features an account's plan includes, minus features its users have worked into a routine. Logins and overall "engagement" hide it.
• Measure it twice, per account and per user. An account can look adopted because one power user does everything while nine paid seats never touch the feature.
• Unused paid features have four common causes. Users never found it, could not get it set up or lacked permission, did not need it, or tried it and never built the habit. Each needs a different fix.
• Judge the work by behavior 30 and 90 days later against a holdout group, then by renewal and expansion. Guide views and email opens prove nothing.
• Published pricing differs sharply: Userlens Starter is $100 a month for 10,000 profiles, Chameleon Pro starts at $750 a month for 5,000 monthly tracked users, Intercom charges $29 to $132 per seat plus an add-on from $99 a month, and Gainsight CS is quote only.

What the entitlement versus usage gap is

An entitlement is a customer's right to use a feature. Stripe's billing documentation defines it plainly: an entitlement represents a customer's access to a feature, created when the customer subscribes to a product with that feature attached. Usage is what your product analytics records. The gap between the two is every feature an account pays for and does not use.

This is narrower than "feature adoption" in general. General adoption asks whether your user base uses feature X. The paid-for gap asks whether this account, which bought the plan that includes X, has anyone using X. The second is a revenue question, because the buyer will eventually compare the invoice with what the team does.

Two cautions on the headline statistics. Pendo's 80 percent figure dates from 2019, counts a "feature" as anything a Pendo customer tagged, and describes all features in a product, not only the ones a given account pays for. Userpilot's 24.5% is an average of core features across its own customers. Both tell you the gap is normal. Neither tells you what yours is.

Causes and matching fixes at a glance

Cause	What the data looks like	Intervention that fits	Who owns it
Discovery	Entitled users have never opened the feature; they do the same job a longer way	In-app guidance at the moment of the workaround	Product or growth
Permission or setup	Feature opened, then abandoned at a configuration step; or only admins can turn it on	CSM or support outreach to the admin, setup help, role changes	Customer success
Relevance	No user in the account performs the job the feature does	Packaging change, or stop pushing it	Product and pricing
Habit	One to three uses, then nothing	Lifecycle email, repeated nudges tied to the user's own activity	Lifecycle marketing or CS

Step 1: Measure the gap per account and per user

Start with three tables. You need one row per account per entitled feature (from billing or your plan configuration), one row per user with the account they belong to and their role, and feature-level usage events with a user ID and timestamp. If your plans are simple, the entitlement table can be a hand-built mapping of plan name to features.

The account view

For each account, list the features its plan includes and mark each one as adopted or not. "Adopted" needs a definition you choose and write down. A common starting point is some number of uses by some number of distinct users inside a rolling window (for example, three uses by two users in 30 days), but that threshold is a judgment call, and a monthly reporting feature should not be held to the same standard as a daily workflow.

Then compute two numbers per account:

• Entitled feature adoption: adopted entitled features divided by total entitled features.
• Paid-for unused value: the unadopted features, weighted by how much of the plan's price they justify. If the account upgraded specifically for one feature and that feature is unused, the weighting matters more than the count.

Most analytics tools can roll events up to the account if you send an account identifier. PostHog's group analytics aggregates events by entities like organizations or companies instead of individual users. Mixpanel does the same with Group Analytics, which is an add-on for Growth and Enterprise plans. Amplitude documents account-level reporting as well. We covered the tooling for this in more depth in our piece on platforms that show individual user activity within customer accounts.

The user view

Account-level adoption misleads in one specific way: a single enthusiast can make a 40-seat account look healthy. So for each entitled feature, also compute seat penetration, which is the number of users who met your adoption threshold divided by the number of users who could reasonably use the feature. That denominator is the hard part.

Pendo's glossary gives a simple product-wide version: monthly feature adoption rate equals feature monthly active users divided by monthly logins, times 100. Use the same shape, but restrict both sides to entitled users in the account.

A worked example

The numbers here are invented to show the arithmetic. An account on your top plan is entitled to 12 features and has 40 paid seats. Five features meet your adoption threshold, so entitled feature adoption is 5 of 12, or 42%. One of the seven unused features is bulk export, the stated reason for the upgrade. Thirty of the 40 seats are in roles that export data, and two of them have used bulk export. Seat penetration for the feature the customer upgraded to get is 2 of 30, about 7%.

Neither number shows up in a login report. The account logs in daily.

What to do with the list

Rank the gaps by renewal relevance, not by size: first the features tied to the reason the account bought or upgraded, then the features that separate one plan from the next. The rest may not deserve any effort. If you are still stitching usage and account data together, our guide to tools that unify product usage data and customer success workflows covers that plumbing.

Step 2: Diagnose why each feature goes unused

Usage data tells you where the gap is and rarely why. Guessing is expensive: a tooltip campaign aimed at a permissions problem produces impressive view counts and no adoption.

Discovery

The user does not know the feature exists, or does not connect it to the job in front of them. The signature in the data is zero opens of the feature combined with repeated use of a slower path to the same result. Someone exporting records one at a time has the need and has not found bulk export.

Permission or setup

The user found the feature and could not use it. Maybe it needs an integration connected, a data import, or an admin toggle. Maybe their role cannot see it. Look for users who opened the feature once or twice and stopped at a configuration screen, and for accounts where the only people with the required role have not logged in for weeks. In-app guidance does little here, because the person seeing the guide is often not the person who can act on it.

Relevance

Nobody in the account does the job the feature was built for. They are entitled to it because it came bundled with something they did want. The signature is the absence of both the feature and any workaround. The honest fix here is usually to stop nudging. Sometimes it is a packaging change.

Habit

The user tried the feature, perhaps several times, and went back to the old way. Usage shows a short burst after a launch or a guide, then a flat line. A single message will not fix it.

How to tell them apart

Run the checks from cheap to expensive. First, split entitled non-users into never opened, opened and abandoned, and used then stopped. That alone separates discovery, setup and habit for most users. Second, check for the workaround behavior to separate discovery from relevance among the never-opened group. Third, ask. A one-question in-product survey shown to people who opened and abandoned ("What stopped you?") or five calls with admins of unadopted accounts will tell you more than another dashboard. Chameleon, Userpilot, Pendo and Intercom all ship in-product surveys, so you may already own the tool.

Expect a mix: one feature can have a discovery problem in small accounts and a setup problem in enterprise ones.

Step 3: Match the intervention to the cause

In-app guidance, for discovery

A tooltip, banner or short walkthrough works when it appears at the moment of the workaround and names the user's own task. It works badly as a day-one tour of everything in the plan. Target on behavior (did the slow path three times this week) and on entitlement (plan includes the feature), and cap how often any user sees a prompt. Pendo, Appcues, Chameleon and Userpilot are built for this, and Intercom's Outbound product covers tours, checklists and tooltips as well.

CSM outreach, for setup and permissions

When the blocker is an admin action, a human message to the right person beats any automated prompt to the wrong one. Give the CSM the specific evidence: which feature, which users opened it and stopped, what step they stopped at. This is where customer success platforms earn their place, because the trigger (an entitled feature unused 30 days after purchase, say) can open a task automatically. Our roundup of AI customer intelligence platforms for SaaS compares that category.

Lifecycle email and repeated nudges, for habit

Habit needs repetition tied to real occasions. Email (or Slack, if that is where your users work) reaches people who are not in the product at the moment the occasion arises, which in-app guidance cannot do. The message should refer to something the user did, such as "you built this report by hand again on Monday", and offer the feature as the shorter route.

One compliance point for US senders. The CAN-SPAM Act treats a "transactional or relationship message" differently from a commercial one, and the definition covers messages whose primary purpose is to notify a customer of a change in the terms or features of a subscription or account, or to deliver product updates the recipient is entitled to receive. A nudge about a feature the customer already pays for may fit that definition. One that mainly promotes an upgrade probably does not, and the primary purpose test in 16 CFR 316.3 decides mixed cases. We are not lawyers, and other jurisdictions have different rules. Include an opt-out regardless.

Packaging changes, for relevance

If a third of the accounts on a plan have no use for a feature bundled into it, more messaging will not help. The options are to move the feature to an add-on, to offer a cheaper plan without it, or to accept the gap. That can protect revenue, because an account that renews at a lower price on a plan it fully uses is worth more than one that churns after discovering it paid for shelfware.

Whatever the channel, cap message frequency per user, exclude accounts in an open escalation, and keep a holdout group that gets nothing. The holdout is what makes Step 4 possible.

Step 4: Check whether adoption moved retention or expansion

There are three levels of evidence, and most teams stop at the first. Level one is delivery: the guide was shown, the email was opened. That tells you the plumbing works.

Level two is behavior. Did entitled non-users start using the feature, and are they still using it 30 and 90 days later? Compare against the holdout. Without one you cannot separate your campaign from a product release, a seasonal spike, or users who would have found the feature anyway. Ron Kohavi and colleagues at Microsoft made the general case years ago: controlled experiments are the best scientific design for establishing a causal relationship between a change and user behavior. Small products will have to live with noisy answers.

Level three is money: did accounts whose gap closed renew at a higher rate, or expand more, than accounts whose gap stayed open? Accounts that adopt more features were often healthier to begin with, so a raw comparison of adopters and non-adopters overstates the effect. The holdout helps again. Compare renewal and expansion for treated accounts against held-out accounts, not adopters against non-adopters.

Be honest about sample size. With annual contracts you may wait a year for a renewal result, and with 60 accounts in a cohort the difference may never be statistically clear. In that case report the behavior result as proven and the revenue result as directional. We would rather show a board a real 90-day usage lift than a confident retention claim built on 14 renewals.

Tools that help, and where each one fits

This is not a ranking. The seven products below are grouped by the job they do in the process above, and most teams need one from the first group plus either the second or the third. Prices are from each vendor's own pricing page on 12 October 2026.

Tool	Job in this process	Channel	Published pricing
Pendo	Analytics plus in-app guides	In-app	Free up to 500 MAUs; Base, Core, Ultimate by quote
Appcues	No-code in-app experiences	In-app, email	Start, Grow, Enterprise not priced publicly; Spark $3,600 a year up to 1,000 MAUs
Chameleon	In-app tours, tooltips, microsurveys	In-app	Startup from $279 for 2,000 MTUs; Pro from $750 a month for 5,000 MTUs
Userpilot	In-app engagement, analytics, surveys	In-app, email	Starter $299 a month up to 2,000 MAUs; Growth from $849 a month
Intercom	Tours and checklists inside a helpdesk	In-app, email, chat	$29, $85 or $132 per seat a month paid annually; Proactive Support Plus from $99 a month
Gainsight CS	Account health, playbooks, CSM tasks	Human outreach	Not published
Userlens (Lumi)	Agent-written nudges from usage data	Email, Slack	Free (100 profiles); Starter $100 a month; Growth from $1,000 a month billed annually

In-app guidance: Pendo, Appcues, Chameleon, Userpilot

These four are the closest substitutes for one another, and all address the discovery cause.

Pendo pairs product analytics with guides. Pendo Free covers up to 500 monthly active users with analytics, in-app guides and Pendo-branded NPS surveys. The paid plans (Base, Core, Ultimate) show "Request pricing", and survey-based sentiment is included only in Ultimate and sold as an add-on below that. Quote-only pricing makes budgeting hard for a small team, and Pendo is a lot of platform if you only need tooltips.

Appcues is a no-code builder for in-app experiences. Its pricing page lists Start (up to 3,000 MAUs, 10 published experiences), Grow (25 experiences) and Enterprise (100) without dollar figures, and says price depends on MAUs, installations and published experiences. The one published number is the Spark program for teams of 25 or fewer: $3,600 a year for up to 1,000 MAUs plus a $250 implementation fee. Watch the caps on published experiences: one flow per feature per segment runs through 10 quickly.

Chameleon sells tours, tooltips, launchers, a HelpBar search and microsurveys. Its plans page lists a Startup plan from $279 for 2,000 monthly tracked users, Pro from $750 a month for 5,000 MTUs with up to 20 live experiences, Growth from $1,250 a month for 5,000 MTUs billed annually, and a custom Enterprise tier. Microsurveys make it useful for the diagnosis step. It is the most expensive of the four at the published entry points, and it has no account-health or CSM workflow.

Userpilot combines in-app engagement, product analytics, NPS and behavior-triggered email. Starter is listed at $299 a month for up to 2,000 MAUs and Growth from $849 a month. The resource center and advanced in-app surveys start at Growth, as does event autocapture, so the Starter plan is thinner than the feature list suggests.

Customer success routing: Gainsight CS

Gainsight CS is where setup and permission problems go when the fix is a person. Gainsight's features overview lists Adoption Explorer (which manages customer usage data inside Gainsight), Scorecards, Cockpit and Playbooks, Success Plans and Journey Orchestrator. In-app guidance is a separate product, Gainsight PX. Gainsight's pricing page shows no dollar figures for either. Consider the implementation effort: this is a platform for teams with CSMs and an admin to run it, and a 15-person startup will get little from it.

Support-led discovery: Intercom

Intercom is a helpdesk first. Its Outbound product adds product tours, checklists, tooltips, banners, surveys and a Series journey builder. On the pricing page, seats cost $29 (Essential), $85 (Advanced) or $132 (Expert) a month paid annually, the Fin AI agent is $0.99 per outcome, and the Proactive Support Plus add-on that covers tours, checklists and surveys starts at $99 a workspace a month with 500 messages included. If you already run support on Intercom, the add-on is the cheapest way to test in-app discovery. If you do not, buying a helpdesk to get tooltips makes little sense.

Agent-written nudges: Userlens

Userlens fits the habit cause and the long tail of accounts no CSM will ever call. Its site describes it as a customer adoption and retention engine for product-led growth and product teams that connects a backend database (such as Postgres) with product analytics (PostHog, Mixpanel or Amplitude), builds a profile of every account and person, and has its agent, Lumi, send personal email or Slack nudges inside a campaign a human approves. Approval covers the goal, audience, tone and limits, not each message. Frequency caps, quiet hours, never-contact flags and skip-if-unsure rules are listed as built in, and success is measured as product behavior, not opens.

The database-plus-analytics design maps neatly onto the entitlement versus usage join in Step 1, which is the reason it belongs in this article. Pricing is public: Free includes 100 profiles, 300 personalized emails a month and one active campaign; Starter is $100 a month for 10,000 profiles, 10,000 emails and five campaigns; Growth starts at $1,000 a month billed annually; Enterprise is by quote. A profile is one person or one company.

What to consider:

• Lumi sends email and Slack messages. It is not an in-app guidance tool, so it does not replace a tooltip at the moment of the workaround.
• It is new. Userlens announced Lumi on 1 September 2026, and the three customers quoted on its homepage are early-stage startups. We found no independent evidence of results.
• Free and Starter list only the PostHog and database integrations. "Additional data integrations" begin at Growth, a tenfold price step.
• You approve a brief, not each message. Teams in regulated industries may want per-message review, which is not how the product is described.
• We found no security attestation named on the homepage, Lumi page or pricing page. "Security and governance controls" appear only under Enterprise. Ask before connecting a production database.

How we evaluated

We read each vendor's own product and pricing pages on 12 October 2026 and recorded only what those pages state: product names, plan limits and prices. Where a vendor publishes no price, the table says so. We did not sign up for trials, run campaigns or measure results in any of the seven tools, and we did not rely on review sites or other vendors' comparison posts. The process in Steps 1 to 4 is our own synthesis. Userlens is described only from userlens.io.

Limitations

• No hands-on testing. Feature descriptions are vendor claims, and in-app builders in particular differ in ways a pricing page does not show.
• The two adoption statistics come from vendors (Pendo, 2019, and Userpilot) and describe their own customer bases. We could not find an independent, current benchmark for paid-for but unused features specifically, so this article does not offer one.
• Gainsight's marketing site blocked automated requests, so its feature list comes from its support documentation and its pricing status from a single page read.
• Userpilot's page lists monthly prices without stating the billing term. Chameleon's Startup price is quoted as written ("starts at $279 for 2,000 MTUs").
• The CAN-SPAM discussion covers US federal law only and is not legal advice. GDPR, ePrivacy and CASL were not assessed.
• The link between feature adoption and retention is plausible and widely assumed. We cite no study proving it, and small cohorts may never give a clear answer.

Frequently asked questions

Why do users not adopt features they already pay for?

There are four usual reasons: they never discovered the feature, they could not complete setup or lacked the permission, the feature is not relevant to their job, or they tried it and did not form a habit. Split entitled non-users into never opened, opened and abandoned, and used then stopped. That sorts most of them.

How do I calculate feature adoption for a single account?

Divide the number of entitled features that meet your adoption threshold by the number of features the account's plan includes. Then, for each important feature, divide the users who met the threshold by the users whose role could use it. Pendo's general formula is feature monthly active users divided by monthly logins, times 100. The account version restricts both numbers to that account's entitled users.

What is a good feature adoption rate?

No single number applies. Userpilot reports an average core feature adoption rate of 24.5% across the SaaS companies in its benchmark, and notes that a 15% rate can be healthy for an optional power feature and alarming for a core workflow. Set a target per feature based on how many entitled users have the job it serves.

How do I know if feature adoption improved retention?

Hold out a random group of eligible users or accounts, run the campaign for everyone else, and compare renewal and expansion between the two groups once contracts come up. Comparing adopters with non-adopters is misleading because adopters were often healthier accounts already.

When should I change packaging instead of pushing adoption?

When the data shows a relevance problem: a meaningful share of accounts on a plan have neither used a bundled feature nor attempted the job it does. Moving that feature to an add-on or creating a plan without it can retain accounts that would otherwise question the price at renewal. Test on a small cohort first.

Conclusion

Start with the join. Until you can list, for each account, the features it pays for and who uses them, every adoption campaign is a guess. Diagnose before you build anything, because only one of the four causes is fixed by a tooltip. Keep a holdout.

On tools, buy for the cause you have. Discovery points to Pendo, Appcues, Chameleon or Userpilot. Setup problems in managed accounts point to Gainsight CS. An existing Intercom helpdesk makes its add-on the cheap first test. Userlens is worth a trial at $0 or $100 a month if your gap is habit across thousands of unmanaged users and email or Slack is an acceptable channel, with the caveat that it launched its agent in September 2026 and has a short track record.

---

This article reflects publicly available information as of October 2026 and does not endorse any specific platform. Vendor pricing, plan limits and product names change often, and the effect of feature adoption on retention or revenue is not guaranteed for any product. Consult a licensed professional for guidance specific to your business.