Kill‑or‑Build Decision Kit: 7 Rapid Tests to Prioritize App Features and Avoid Wasted Engineering

Treat every feature hypothesis as a single test. Define the riskiest assumption (demand, willingness to pay, or retention impact) and pick the experiment that most directly challenges that assumption. Each test below is designed to be run in 2–14 days, with a one‑metric decision rule (example thresholds below).

Run experiments sequentially or in parallel depending on risk. If the idea affects activation or retention, prioritize activation smoke tests first. If it’s monetization‑led, run paid preorders or a price anchor fake door. After each experiment, record the outcome as one of: Kill (stop investing), Iterate (change hypothesis and re‑test), or Build Brief (green light to write specs and estimate engineering).

Create a simple landing page or in‑app CTA that presents the feature as if it exists, and measure clicks and signups. The goal is behavioral evidence: people who click a CTA or join a waitlist reveal more than survey answers. Use copy that clarifies value and an explicit next step (join waitlist, pre-order, schedule demo).

Decision rule: if conversion to meaningful action (click → email or preorder) meets your threshold (example: 2–5% of reachable audience via a short ad or mailing), move toward paid‑preorder or concierge; if <0.5% across a credible sample, kill or pivot the idea.

Deliver the feature manually to a small set of users. For subscription or service features this exposes genuine product‑market fit signals: are users willing to trade time or money for the manual workflow? Concierge tests are especially valuable for personalization, onboarding or complicated flows where automated engineering is expensive.

Decision rule: if 4–8 pilot users repeatedly return within the manual flow and report concrete value (NPS-like qualitative signal + willingness to continue), convert the manual steps into acceptance criteria for a build brief. If manual delivery is unsustainably expensive with weak retention, kill or rework the value prop.

A paid preorder — charging before the feature exists — is the cleanest signal for monetizable demand. Offer an early‑access price or lifetime discount with an explicit ship window and refund policy. Even a modest paid‑preorder conversion rate from your engaged audience demonstrates real willingness to buy.

Decision rule: if your preorder conversions reach a prespecified revenue threshold (for example, cover expected initial engineering + 20%), proceed to build; if you get strong interest but low willingness to pay, iterate pricing or packaging; if interest is negligible, kill.

Combine a short in‑product or email micro‑survey with a follow‑up behavioral ask (e.g., 'Would you like to join a pilot? Click to schedule'). Micro‑surveys convert stated preference into a small action and filter out noise from vague opinions.

Decision rule: if >20% of respondents take the follow‑up action or self‑select into pilots, treat that as positive signal to run a concierge or preorder. If survey responses cluster around 'not for me' or show low follow‑up, kill or reframe the target persona.