# Feature adoption marketing for SaaS > Run adoption campaigns that move usage: targeting by entitlement gap, the in app and email mix, launch tiers for existing customers, and holdout testing. Source: https://saas-marketing.net/guides/saas-feature-adoption-marketing/ Topic: SaaS Customer Marketing Type: guide Published: 2026-09-11 Last updated: 2026-09-11 Publisher: SaaS Marketing (saas-marketing.net) License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/saas-feature-adoption-marketing/ ## Short answer Feature adoption marketing targets existing customers who are entitled to a capability but not using it. The funnel runs entitled, aware, tried, habitual, and the largest drop is usually between aware and tried. Target by entitlement gap and usage pattern rather than by segment, sequence in app messages ahead of email because in app engagement is far higher, size the campaign to a launch tier, and always hold out a control group so the reported lift is real. ## Key takeaways - Any adoption number reported without a holdout group is a usage trend standing next to a campaign. - The entitled-to-aware and aware-to-tried steps lose the most accounts, and email alone rarely fixes either. - In app messages reach users inside the product where intent already exists, so they consistently out-engage email. - Targeting by entitlement gap beats targeting by segment, because entitlement gap is the actual eligibility criterion. - Applying tier one launch effort to a tier three feature burns campaign attention you need for real launches. - Habitual use, not first use, is the number that predicts renewal and expansion, so measure at day 30 and 60. --- Feature launches get built for prospects. The positioning, the landing page, the launch post, all of it is written to explain why someone should buy. Then, a week before ship date, someone asks what we are telling existing customers, and the answer is a forwarded version of the same message. That is why adoption numbers disappoint. Customers already bought. They need a different sentence. ## The adoption funnel: entitled, aware, tried, habitual Four states, and each transition loses people for a different reason. Mapping your own numbers against this takes an afternoon in Amplitude or Pendo and it is the most clarifying thing you will do all quarter. | Stage | Definition | Why accounts drop out | |---|---|---| | Entitled | Has access on their current plan | Plan gating they do not know about | | Aware | Knows the feature exists | Never logged in, or ignored the announcement | | Tried | Used it at least once | Setup cost, unclear first step, no obvious trigger | | Habitual | Used it in 30 of the last 90 days | It did not replace the old way of working | In most B2B products the entitled-to-aware and aware-to-tried transitions are where the campaign lives. Awareness is a distribution problem you can solve with channels. Trial is a friction problem you usually cannot solve with a message at all, which is the uncomfortable finding most adoption campaigns produce. Ever-tried is the number that goes in the launch retro because it looks good. Habitual use at day 60 is the number that correlates with renewal and expansion. Report both, lead with the second, and expect the second to be a fraction of the first. ## Target by entitlement gap, not by segment The standard approach is to pick a segment, usually by plan or industry, and message everyone in it. That is wrong because plan membership is not the eligibility criterion. Entitlement plus non-use is. Build the list this way: accounts entitled to the feature, that have not used it in the last 90 days, and that show the usage pattern the feature is designed to improve. That third condition is the one that lifts response rates, because it is the difference between "here is a new thing" and "here is a faster way to do the thing you did 40 times last month". An example. If you ship a bulk edit capability, the target list is not "all Pro accounts". It is Pro accounts whose users performed more than 20 single edits in a session at least three times in the last month. That list will be a fraction of the segment and it will convert several times better. **Building an entitlement gap audience** ## The channel mix and the order it runs in In app first, always, for accounts that log in. The user is already inside the product with the relevant screen in front of them, which is context no email can manufacture. Pendo, Intercom and Appcues all do this competently and the choice between them matters far less than the targeting quality behind them. Email second, roughly two weeks later, and only to accounts that either did not see the in app message or saw it and did nothing. Email's job here is reach into dormant accounts, not persuasion. Release notes are a reference surface, not a campaign. Notion's changelog is a good model: clear, dated, linked from the product, and nobody pretends it drives adoption on its own. It exists so that a customer who hears about the feature elsewhere can confirm the details. Then two human channels that carry more weight than their volume suggests. Office hours or a live walkthrough, which works well for features with a real setup cost, and CSM outreach for named accounts where the feature maps to an expansion conversation. Reserve both for the top launch tier, because they do not scale. ## Launch tiers applied to the installed base Most companies have launch tiers for prospect-facing announcements and no equivalent for customers, so every feature either gets the full treatment or nothing. Both are wrong. A tier three feature that receives a tier one campaign consumes in app attention you needed for the launch two months later. Three tiers is enough. **Tier one.** Changes what the product is or unlocks a new use case. Full sequence: in app, email, office hours, CSM outreach on named accounts, release note, community post. Runs six to eight weeks with a holdout. Two or three of these a year. - **Tier two.** Meaningful improvement to an existing workflow. In app to the entitlement gap audience, one email to non-responders, release note. Four weeks, holdout if the audience is large enough to split. Maybe one a month. **Tier three.** Quality of life improvements and fixes. Release note only, plus a contextual tooltip where the change appears. No campaign. The majority of what you ship. Every in app message you show spends a little of your users' willingness to read the next one. Teams that message everything find their tier one launches landing at tier three engagement rates, and the decline is very hard to reverse. The [feature launch tiers playbook](/playbooks/saas-feature-launch-tiers/) has the full decision criteria, and the [launch week social distribution checklist](/checklists/saas-launch-social-distribution/) covers the external half of a tier one launch. ## The holdout, and why most reported lift is fiction Here is the position I will defend anywhere: any adoption number reported without a holdout is a usage trend with a campaign standing next to it. Features get discovered organically. Usage has seasonality. A product-side change shipping the same month can move the number more than your campaign did. Without a control group you have no way to separate any of that from your work, and the number you put in the board deck is guesswork dressed as measurement. The mechanics are not hard. Hash the account ID, hold back 10 to 20 percent, ship nothing to them, and compare adoption rates at day 30 and day 60. Lift is the difference. If treated accounts hit 34 percent adoption and the holdout hit 26 percent, your campaign produced 8 points, not 34. **10 to 20%** Holdout share that gives usable lift measurement on most B2B adoption audiences Two honest limitations. Small audiences cannot support a holdout: below roughly 300 eligible accounts the confidence interval swallows any realistic lift, so measure a pre and post window instead and label it as weaker evidence. And holdouts have a cost, because you are deliberately withholding a useful capability announcement from paying customers. For a feature that solves an active support burden, that cost is real and you should skip the holdout and say why. ## What to do when the campaign works and habitual use does not follow This is the most common outcome and it is genuinely useful information. Trial goes up, day-60 habitual use does not move, and the feature quietly reverts to background usage. It means the message worked and the product did not. The feature did not replace the existing way of doing the job, either because it was slower in practice, because it required setup the user never finished, or because the old method was good enough. No amount of additional messaging fixes that, and running a second campaign is the wrong response. The right response is a handful of calls with accounts that tried it once and stopped. Five conversations usually produce the answer, and the answer usually belongs to product. That is a better outcome than a lift number, even though it is harder to put in a deck. ## What to do next Pick the last feature you shipped to existing customers and calculate the four funnel numbers for it. Entitled, aware, tried, habitual at day 60. Most teams have never done this for a single feature and the first one is always instructive. Then build the entitlement gap audience for your next launch and hold out 15 percent before anything ships. The full campaign structure is in [feature adoption campaigns](/playbooks/feature-adoption-campaigns/), and there is a guided version in [lesson 2: run an adoption campaign with a holdout](/courses/saas-customer-marketing-sprint/02-run-an-adoption-campaign/). For where these numbers should land, see [SaaS product launch benchmarks](/research/saas-product-launch-benchmarks/) and the definition work in [what is product adoption rate](/glossary/product-adoption-rate/). Once adoption is moving, the revenue connection runs through [how marketing moves NRR](/playbooks/nrr-expansion-levers/) and [expansion marketing plays](/playbooks/expansion-marketing-plays/). The wider discipline sits at [SaaS customer marketing](/saas-customer-marketing/). ## Frequently asked questions ### What is feature adoption marketing? It is the practice of driving usage of an existing capability among customers who already have access to it. Unlike a product launch aimed at prospects, the audience is the installed base and the constraint is attention rather than awareness of the company. The work is targeting by entitlement gap, sequencing in app and email touches, and measuring lift against a holdout group. ### How do you measure feature adoption rate? Divide accounts or users who have used the feature by those entitled to use it, over a defined window. Report it at two depths: ever tried, and habitual use over the last 30 days. Ever-tried numbers flatter campaigns badly. Habitual use is the number that correlates with renewal, so report both and lead with the second. ### Should feature announcements go in app or by email? Both, in app first. In app messages reach users while they are already in the product with the relevant context in front of them, so engagement runs far ahead of email. Email reaches people who have not logged in recently, which is exactly the group in app cannot touch. Sequence in app for two weeks, then email the non-responders. ### What is a holdout group in an adoption campaign? A randomly selected slice of the eligible audience, usually 10 to 20 percent, that receives no campaign messages. Adoption in the holdout is your baseline. Lift is the difference between the treated group and the holdout, not the raw adoption number. Without one, you cannot separate campaign effect from organic discovery and seasonal usage. ### How long should a feature adoption campaign run? Six to eight weeks for the active phase, then measure habitual use at day 60 and day 90. Shorter windows overstate results because they catch first-use spikes that decay. Anything longer becomes background noise and the messages start damaging trust in your in app channel, which is a finite resource. ### Why do feature launches underperform with existing customers? Because the campaign was built for prospects and bolted onto customers afterwards. Prospect launches lead with the category problem, which customers already solved when they bought. Customers need to know what changes in their current workflow, whether they are entitled, and what it costs them to switch to the new way. Different message, different channel, different success metric.