Insights from Martin Tobias (Incisive), Webinar for Founders, September 17, 2025.
TL;DR — The Baby Unicorn Formula:
Pick a “hair‑on‑fire” problem and attack it with a tiny, founder‑led team that ships weekly, grows 5–7% MoM, and proves distribution before you overbuild. Keep the original problem sacred (soft pivots only), insist on a technical co‑founder, and exploit a fresh technology unlock in a big market. Then move faster than everyone else—especially on customer work and go‑to‑market.
Why listen to this playbook?
Martin Tobias is Managing Partner at Incisive, a pre‑seed fund focused on B2B software. Across the past five years, his program has backed 82 companies with 6 unicorns, 35 at $100M+, and 67% of pre‑seed rounds leading to follow‑on capital. Incisive has also been recognized among top seed VCs by SignalRank and Business Insider—underscoring that this isn’t theory, it’s a repeatable pattern.
What the data says about unicorns
Team composition matters—deeply. A 20‑year analysis of unicorns (popularized by Aileen Lee) shows:
- 2–3 co‑founders is the sweet spot; solo wins are rare.
- ~80% had prior startup experience (successful or not).
- A technical co‑founder is essential; true outliers without one are vanishingly few.
Vision persistence is a predictor. Roughly 90% of winners stay close to their original problem. Soft pivots (re‑framing the wedge or target customer) are fine; hard pivots (e.g., fintech → supply chain) correlate with failure.
Execution cadence is non‑negotiable. Unicorn teams ship weekly, target 5–7% monthly growth (the YC standard), and do hands‑on early customer work—demos, interviews, and manual workflows—long before automation.
What Martin looks for at pre‑seed
- 2–3 founders who’ve worked together 5+ years. At minimum: one technical + one business founder. If you can’t attract a co‑founder, the opportunity or your credibility needs work.
- Rapid execution receipts. Think 200 customer discovery calls in a month and visible week‑over‑week progress, even before there’s a product.
- Distribution > code (at first). Identify a credible wedge and the first 100 customers before you scale the build.
- A fresh technology unlock. New AI infra, better models, or platform shifts that make yesterday’s “impossible” now trivial (e.g., recent leaps in PDF understanding for accounting).
- A market with tailwinds. Directionally large and growing; $100–200M TAM is too small for venture scale. Blue ocean preferred; ruthless execution can win in red oceans too.
“Distribution is harder than code—and now code is easier than ever. Earn your wedge first.”
Martin Tobias
Case studies I have invested in: the wedge, the timing, the win
Deel (Employer of Record).
TriNet dominated the market with a ~70% market share, but its NPS was abysmal. Deel avoided head‑on U.S. combat and instead built international hiring across 180 countries, partnering with incumbents early and riding the 2019–2021 remote‑work wave. Only then did they expand into the U.S.—from a position of strength. Multiple founders with startup experience.
Jeeves (LatAm corporate cards).
A regionalized Brex‑style play: underwrite companies, not individuals, with a better UI and regulatory strategy tuned for Latin America. Timing aligned with a regional fintech funding boom and a second‑time founding team.
LMNT (Electrolytes, not software).
A CPG counterpunch anchored on a counterintuitive promise—drink more salt—and science‑backed formulation (roughly 1000mg sodium / 400mg magnesium / 200mg potassium). The distribution wedge was a co‑founder with real audience reach; stick packs cut logistics costs and avoided shelf wars. They stuck to one product until ~$500M revenue.
Pattern across all three: a sharp wedge + founder credibility + market timing, executed fast.
The 90‑Day Startup Blueprint
1) Sharpen the wedge around a “hair‑on‑fire” problem.
Be specific: a job title, a workflow, a format (e.g., AP invoices from PDFs). Your first 100 customers should be nameable before the build.
2) Validate product‑market fit with the Sean Ellis test.
Survey early users weekly/bi‑weekly: “How disappointed would you be if you couldn’t use this?” Push until ≥40% answer “very disappointed.”
3) Ship weekly; test distribution weekly.
Every week: release an improvement and run one channel experiment (e.g., LinkedIn outbound, community posts, paid search, TikTok, cold emails). Compare 5–6 channels over a few sprints; double down on what moves activation and retention.
4) Keep it founder‑led and tiny until PMF.
The CEO should personally handle onboarding (à la Superhuman), support, and key sales calls. Don’t hire a sales team until there’s a pull you can measure.
Traps that quietly kill promising companies
- Hard pivots. When you abandon the core problem, you’re usually starting from zero. If you must, start a new company rather than torching the cap table.
- “First to market” obsession. Plenty of giants were fast seconds or thirds (Jeeves vs. Brex; Google vs. earlier search incumbents). Be the fastest and best, not the first.
- Premature hiring and burn. The thing that kills startups is running out of money; the thing that causes that is hiring too soon. A $600K pre‑seed shouldn’t become $10K MRR and 18 months of drift. Delay sales hires until genuine PMF signals emerge.
The state of play (right now)
- AI‑native upstarts are pressuring legacy SaaS (marketing automation was the example), and development cycles are compressing with better tooling.
- Services firms with more demand than supply are perfect customers for workflow‑level AI.
- Even outside software, LMNT’s focus, brand clarity, and distribution wedge translate directly to B2B: stay narrow, keep promises, and own your channel.
Your 2‑week action plan (bookmark this)
Week 1
- Draft a one‑sentence wedge and list 100 named prospects (titles + companies).
- Line up 20 discovery calls; book them before you write code.
- Define a single North Star metric (e.g., weekly active teams) and 2–3 input metrics.
Week 2
- Ship MVP v0 or a concierge workflow.
- Run one channel test (e.g., LinkedIn founder‑led outreach with a tight message).
- Launch the Sean Ellis survey; start weekly cadence.
- Publish a weekly changelog (customers notice).
Rinse and repeat for three months. If you’re not at ≥40% “very disappointed,” you haven’t nailed PMF yet—iterate the wedge, not the vision.
Final word
If you remember only three things:
- Keep the problem constant; pivot the wedge, not the vision.
- Win distribution early; code is cheaper than credibility.
- Stay tiny and ship weekly until the market pulls you forward.