A deep read of the Minimum Viable Segment (MVS) framework from Underscore VC (Lily Lyman and Gabrielle McCarthy), derived from Michael Skok's Startup Secrets curriculum at Harvard.
Why this analysis exists
This is one of the source layers behind the BuildOS cold-email-icp-signal-design child skill. It supplies the segment-testability gate — the three hard tests that decide whether a candidate segment is even valid before any timing signal, anchor, or offer work begins. A segment that fails the MVS check produces an uninterpretable campaign no matter how good the email is.
Core thesis
"Minimum Viable Segment is about focusing on a market segment of potential customers with the same needs. Without it, potential users who have divergent needs will quickly pull your MVP in many different directions."
The MVS framework's argument: when an MVP fails, the cause is usually a segment failure, not a product failure. The product is being pulled in three directions because the segment contains three sub-groups with three different pains. The fix is not to add features — the fix is to cut the segment until the remaining customers share a single dominant need.
For cold outreach, the same logic applies. A segment with divergent needs produces uninterpretable reply data. Some replies confirm the offer; others reject it; others ask for something the offer does not promise. The campaign cannot be diagnosed because the segment was never coherent.
The three MVS criteria
A segment is valid only if it passes all three:
1. Common Needs
"Customers must have the same or as similar a need, pain, or problem as possible."
Emphasis: needs, not features. Customers think in pain. Two customers who "use the same tech stack" but have different jobs-to-be-done are not in the same segment. Two customers in entirely different industries with the same job-to-be-done are.
Examples Underscore VC cites:
- ✗ "Financial services" — too broad, customers have divergent needs.
- ✗ "PCI compliance across eCommerce + merchants + software providers" — same regulatory pain, but mixed buyer types and channels.
- ✓ "PCI compliance for eCommerce software providers" — single regulatory pain, single buyer type, shared channels.
2. Dominability (the "Minimum" part)
"The segment must be small enough to dominate it."
This is the counterintuitive criterion. The right segment is not the biggest one you can plausibly serve — it is the smallest one you can own. Once dominated, the segment becomes a referenceable beachhead from which adjacent segments can be entered.
The test: can you become the obvious choice for this segment? If the answer is "we'd be one option among many," the segment is too big.
3. Viability (the "Viable" part) — your MVP can actually serve it
"Which segment will have the least product requirements?"
A segment is only viable if your current product, with today's support capacity, can serve it without significant customization. The right segment is the one that needs the fewest features to convert.
This is the alignment point with Lincoln Murphy's "Able" dimension and Mark Roberge's LIR formula: the MVS is the segment where the current product can hit the activation event.
Additional factors that strengthen a segment
Beyond the three required criteria, Underscore VC names four practical reinforcers:
- Customer interconnectedness. Members of the segment talk to each other — forums, trade associations, conferences, peer networks. Once one buys, others hear about it. Referenceability compounds.
- Channel reachability. All members are reachable through one consistent channel. A LinkedIn-discoverable segment is a coherent campaign target; a "spread across LinkedIn + traditional press + niche Reddit" segment is three campaigns.
- Geographical proximity. For high-touch or in-person motions, the segment is concentrated enough to visit, host an event for, or sponsor.
- Shared vocabulary. Members use the same language for their pain. If the segment uses three different phrases for the same problem, the cold email is fighting the segment's incoherence.
Disqualifiers — what kicks a segment out of MVS
| Disqualifier | What it produces |
|---|---|
| Too-broad vertical with no shared specific need | Outreach reads as generic; reply data is mixed and uninterpretable. |
| Divergent needs across sub-groups | MVP gets pulled in multiple directions; cold outreach offer cannot match all sub-groups. |
| Channels and geographies require expensive, scattered outreach | Acquisition cost per qualified meeting is too high to support unit economics. |
| Customization needs exceed MVP scope | "Win" customers become a service load that prevents serving the next ten. |
| Members do not talk to each other | No word-of-mouth compounding; every customer is a fresh cold start. |
What this contributes to the BuildOS ICP and Signal Design child skill
- Three-test segment gate. Every segment definition the skill produces must pass the three MVS tests explicitly. The output contract has a
mvs_check: { common_needs: pass | fail, dominability: pass | fail, viability: pass | fail }field. A single fail blocks the segment from outreach. - The "size" inversion. The skill defaults to recommending the smallest dominable segment, not the largest serviceable segment. The compiler rejects "broader is better" framings.
- Channel reachability as a required field. Segment definitions require a named, consistent reach channel (a LinkedIn search, an industry list, a trade publication, a specific community). "Mixed channels" is treated as a soft-fail and triggers a sub-segment proposal.
- Shared-vocabulary requirement. The skill asks for the exact language the segment uses for its primary pain. If the segment uses multiple incompatible phrasings, the segment is incoherent and gets split or recut.
- Reconciles the MVS-vs-TAM tension. The skill explicitly separates "ICP for outreach this quarter" (must pass MVS) from "TAM over time" (the segments you graduate into after dominating MVS). The first is the skill's scope; the second is out of scope.
How MVS interacts with the other source layers
- Maurya (switching triggers): Trigger types describe when a segment is in-market. MVS is the gate before trigger analysis — a segment that fails MVS is not worth monitoring triggers for.
- Murphy (seven dimensions): Murphy's "Able" dimension overlaps with MVS Viability. Murphy's "Acquisition Efficiency" overlaps with MVS Channel Reachability. MVS is the cleaner first-pass gate; Murphy is the deeper quality scorer.
- Roberge (Green/Yellow/Red): A Green segment is by definition an MVS-passing segment that has passed LIR validation. MVS is necessary but not sufficient for Green status.
Caveats
- Practitioner framework, no peer-reviewed evidence (Level 4). Derived from Michael Skok's Harvard Startup Secrets curriculum and Underscore VC's portfolio observations. Widely adopted but not formally validated.
- Built for early-stage, MVP-shaped products. A post-PMF company with a mature product has more segments that satisfy Viability; the MVS criteria still apply but the gate is less restrictive.
- Strongest for B2B SaaS and infrastructure plays. Consumer products with low-AOV impulse buys do not need the Dominability criterion — the equivalent is "owning a specific use case in the algorithm," which is a different mechanic.
Source
- Framework page: Minimum Viable Segment (MVS) Framework — Underscore VC
- Original curriculum: Michael Skok, Startup Secrets, Harvard
- Authors of the Underscore VC version: Lily Lyman, Gabrielle McCarthy