--- title: "Feature Discovery — Expansion Growth Tactic" description: "Points subscribers locked in a single feature category toward a second and third, the breadth move that raises retention without nagging." canonical: "https://churnkey.co/growth/library/feature-discovery" category: "growth" related: - /growth/library/index.html.md - /growth/library/methodology.html.md - /tools/retention-rate-calculator.html.md last_updated: "2026-06-09" --- # Feature Discovery Points subscribers locked in a single feature category toward a second and third, the breadth move that raises retention without nagging. - **Stage:** Expansion - **Evidence grade:** strong (large, consistent correlation) - **Evidence:** +6.6 percentage points of retention from one to three feature categories - **Trigger type:** Threshold - **Channels:** in-app ## What is it? Subscribers who live in a single feature category are one workflow change away from leaving. To them, the product is exactly one tool, replaceable by any other tool that does that one thing. Moving from one feature category to three carries 6.6 percentage points of additional retention, because breadth turns a tool into a system. This tactic finds subscribers locked in one category and points the way to a second, then a third, chosen from what similar subscribers adopted next and offered inside the product at moments adjacent to the work they are already doing. ## When it fires The threshold is sustained single-category usage: a subscriber active in the product but confined to one feature category past the learning period. The suggestion renders in-app, framed from the subscriber’s existing work, with the second category presented as a natural extension rather than a separate product to learn. Pacing is the discipline. One category at a time, one suggestion per session, and silence after a dismissal. The corpus phrase is the operating instruction: point the way without nagging. ## What the evidence shows The breadth effect, 6.6 percentage points of retention from one category to three, holds across feature-category usage data from hundreds of organizations. It is one of the most actionable retention correlations in the dataset because the lever sits in the product’s hands: discovery costs nothing but placement. The churn-reason data explains why the lever exists at all: 19 percent of churn cites a missing feature, and a meaningful share of those features already exist. The subscriber never found them. Discovery is partly a retention play and partly a correction for value the product already shipped but never put in front of anyone. ## How it runs In production, the tactic maps each org’s features into categories from usage data, tracks per-subscriber category breadth, and resolves the next-best category per subscriber from the adoption paths of retained subscribers with similar usage. Suggestions render in-app at contextually adjacent moments. Guardrails keep discovery polite: hard frequency caps, an immediate stop at three categories (the measured effect plateaus there, and the suggestion budget is better spent elsewhere), and cooldowns after dismissal so the same suggestion never becomes wallpaper. ## Run it with the Churnkey MCP This tactic runs via the Churnkey MCP. Connect it to your favorite AI agent and it reads your own usage and billing data, then recommends and runs the plays most likely to move your LTV — starting with whether this one fits. **Trigger:** An active subscriber’s feature usage remains confined to a single category past the learning period, while overall engagement holds steady. **Surface:** In-app suggestion at a moment contextually adjacent to the subscriber’s current work, framing the next category as an extension of it. **Guardrails:** - Rate limit: at most one discovery suggestion per session and two per month per subscriber. - Sequencing: one new category at a time. Never present a menu of everything the subscriber has not tried. - Cooldown: after a dismissal, suppress suggestions for that category for 30 days and rotate to a different candidate on the next eligible surface. - Scope: stop suggesting once the subscriber is active in three categories. The measured retention effect is the one-to-three move. **Evidence:** Moving from one feature category to three carries 6.6 percentage points of additional retention, consistent across feature-category usage data from hundreds of organizations. Churn-reason data shows 19 percent of churn cites a missing feature, often one that exists unfound, making discovery both a retention lever and a correction for value the product shipped but never showed. A strong correlation, not a causally isolated effect. **How to run it:** 1. Map the product’s features into categories from usage data, grouping by distinct jobs rather than navigation structure. 2. Track per-subscriber category breadth and flag active subscribers confined to one category past the learning period. 3. Resolve the next-best category per subscriber from the adoption sequences of retained subscribers with similar usage profiles. 4. Render the suggestion in-app at a moment adjacent to the subscriber’s current work, framed as an extension of it. 5. On dismissal, start the per-category cooldown and rotate candidates. On adoption, advance toward the three-category target and then stop. 6. Track second- and third-category adoption rates and the retention of broadened subscribers against single-category peers. Tune suggestion placement against those outcomes. Install the MCP server: ```bash npm install -g @churnkey/mcp ``` The full reference is in the [Churnkey MCP docs](https://docs.churnkey.co/data-integrations/mcp). ## Related - [Growth Tactics Library](/growth/library/index.html.md) — The full register of graded growth plays. - [How Tactics Are Graded](/growth/library/methodology.html.md) — The four-grade evidence ladder behind this grade. - [Retention rate calculator](/tools/retention-rate-calculator.html.md) — Related calculator.