Retention strategies for single-use products – Vivek Kumar (Investor and Advisor, Atlys)

Product decisions built on daily-active metrics fall apart when your customers show up once a year, or once a decade. In this episode, Randy Silver talks to Vivek Kumar about building and growing low-frequency products, from property and tax to jobs and dating.

He explains why PMF should be measured by market penetration, how to prioritise enduring problems with the BELT framework, and how the ICE model blends engagement, lifetime retention, and distinctiveness to reduce CAC and lift LTV. We also discuss research that actually works for episodic journeys and debate “cures vs. treatments” in product ethics.

Chapters
04:25 — What makes a product “infrequent”? Episodic use and recall decay
07:05 — Rethinking PMF: penetration and market share over retention curves
10:36 — When iteration is slow: prioritising problems under seasonal cycles
14:28 — BELT framework: behaviours, enduring vs transient problems, lock-ins
21:56 — Spotting enduring problems: “what will still matter in 10 years?”
24:11 — ICE framework overview for infrequent products
26:03 — Engagement: active retention, complexity, single- vs constant-touch
29:55 — Predictable vs unpredictable retention; referrals as a strategy
31:06 — Lifetime retention: seeding frequency hooks (e.g., estimates, salary data)
33:01 — Distinctiveness and brand: why CAC collapses when you own the memory
33:48 — Control over experience: monetisation through end-to-end journeys
36:13 — Research that works: ethnography, diary studies, “follow-me-home”
40:22 — Example: discovering the real tax filing pain (document collection)
43:04 — Ethics and value: “cures vs treatments”, utility vs entertainment products

Key takeaways
— Infrequent products are defined by episodic transactions, not daily use; recall decays as time between events grows.
— Classic PMF heuristics (retention curves) mislead here — penetration and share of the in-market cohort are better signals.
— Prioritise enduring, value-driving problems over transient irritations; satisfaction beats momentary delight.
— Use BELT to frame opportunities: existing behaviours, enduring problems, lock-ins, and transient problems.
— Design engagement for the transaction window: assess product complexity and whether it’s single- or constant-touch.
— Treat retention in two modes: predictable (tax, insurance) vs unpredictable (jobs, dating); build referrals for the latter.
— Seed lifetime hooks (e.g., valuations, salary checks) to create reasons to return between buying cycles.
— Brand distinctiveness is a moat in low-frequency categories; it drives direct/organic traffic and lowers CAC.
—Monetise by expanding control over the experience (own more of the offline/opaque steps).
— For research, rely on ethnography and observation over memory-based interviews; people recall peaks and troughs, not the middle.