How to Design and Execute a Pilot Program That Generates Investable Evidence
A pilot is not a demo – it’s a structured experiment designed backwards from what an investor or Investment Committee needs to see before writing a check. Every element of the pilot, from the hypothesis to the KPIs to participant selection, must trace back to a specific question the investment memo needs answered.
Why Most Pilots Fail to Produce Evidence
Teams commonly design the pilot before writing the investment hypothesis, starting from “what can we test” instead of “what does an investor need to see.” KPIs get defined operationally rather than as investable thresholds, the corporate-side stakeholder is an innovation champion rather than a P&L owner, and data gets collected without a plan to turn it into charts an investor can read in five minutes. Consequently, most corporate pilots fail to produce investable evidence even when the product itself is good.
Write the Pre-Pilot Investment Hypothesis
Before touching any execution detail, write a one-page document specifying four evidence types: Problem Evidence (a monetizable pain owned by a budget holder), Traction Evidence (real users getting real value with a “good enough to fund” threshold), Model Evidence (a testable pricing model and unit economics), and Strategic Evidence (specific corporate synergies the pilot can demonstrate). Apply the spec-sheet test: if the pilot as scoped can’t generate all four evidence types, redesign it before recruiting anyone.
Complete the Pilot Design Canvas
The canvas standardizes every pilot across the portfolio and must be signed by both the venture lead and the corporate sponsor before any budget is spent. It covers six elements: a falsifiable objective with a specific capital decision attached, named stakeholders including a BU P&L owner with real budget authority, tightly scoped constraints (one geography, one segment, one use case), three to five hypotheses each paired with a KPI and threshold, a data plan specifying exportable investment-memo visualizations, and a timeline with hard gates at week 4 and week 8.
Select the Right Pilot Type
Three core types fit different venture contexts. Type A, Internal Efficiency, tests a workflow tool inside one business unit with a process owner as the key stakeholder. Type B, Corporate-Startup B2B, recruits external customers through the corporate’s own distribution channel, with the BU head who controls those relationships as the critical stakeholder. Type C, New Market or Business Model, tests liquidity and take-rate for a marketplace or platform concept in one tight vertical. For CVC tracks with a strategic mandate, the synergy being pursued – distribution, data, regulatory positioning, or brand – determines which type and which BU co-design requirement applies.
Recruit Against the Target Persona
The most common source of bias is selecting participants who are already supportive rather than participants who match the target customer persona. Recruit five to eight participants per iteration cycle, screen every candidate against role, company size, current behavior, and buying authority, and – for Type B pilots – recruit through the corporate’s own distribution channel rather than the venture team’s personal network, since only the former counts as investment-grade evidence.
Track the Eight Core Pilot Metrics
Set up the dashboard before the first session, with a pre-defined threshold for every metric: task completion rate, time on task, NPS, willingness to pay, activation rate, retention proxy, issue frequency, and the composite Demand Signal Score. A metric without a threshold is just an observation; a metric with a threshold is evidence that maps directly to a section of the investment memo.
Run Iteration Cycles with Hard Governance Gates
A pilot runs a minimum of three Build-Measure-Learn cycles over 12 weeks, with hard gates at week 4 and week 8 that produce a documented Go, Correct, or Stop decision – never an optional check-in. A Stop, triggered by task completion under 40%, zero willingness to pay, or NPS below 6, preserves capital for a better opportunity and should be escalated to the Executive Sponsor within 24 hours rather than delayed with additional cycles.
Produce the Pilot Evidence Report and Convert to Customer Zero
Within five business days of the final session, produce a 10-section report that answers the four Pre-Pilot Investment Hypothesis questions with traceable data, feeding directly into the Tranche 2 capital request. Then run the commercial escalation conversation in the final debrief itself: every participant who reaches Round 3 with a positive NPS is a Customer Zero candidate, and the pilot isn’t closed until at least one signed LOI, reference letter, or peer introduction is secured.
Frequently Asked Questions
What makes pilot evidence "investable" rather than just positive feedback?
It’s tied to a pre-defined threshold that triggers a specific capital decision – for example, “40%+ three-month retention triggers seed-level funding” rather than a vague goal like “we will track retention.”
What are the three pilot types?
Type A (Internal Efficiency, tested inside one business unit), Type B (Corporate-Startup B2B, recruiting external customers through the corporate’s distribution channel), and Type C (New Market or Business Model, testing liquidity and take-rate for a marketplace concept).
How many participants should a pilot recruit per cycle?
Five to eight participants per iteration cycle, screened against the target persona’s role, company size, current behavior, and buying authority – not selected for enthusiasm.
What happens at the week 4 and week 8 hard gates?
Each gate produces a documented binary decision: Go (continue), Correct (adjust scope or KPIs with a specific documented change), or Stop (terminate and document three learnings). A gate without a documented decision is a meeting, not governance.
What does it take to convert a pilot participant to Customer Zero?
At least one of three commercial outcomes: a signed LOI or paid contract, a named reference letter or case study right, or a peer introduction – secured in the final debrief session, not weeks later once the relationship has cooled.
