Key takeaway
Fund the next question only when its answer can change the decision. Costs already spent do not authorize the next stage.
Make the stopping condition observable
A licensing pilot drifts when every small task feels like progress: count records, export them, clean them, answer another questionnaire. The sponsor needs a narrower objective. What uncertainty will the next stage resolve, who can answer it, and when will the business stop paying to wait?
Handshake’s public licensing disclaimer makes its indicative estimate conditional rather than a purchase or license offer. A public program page can justify a question about possible fit; it cannot make the calculator output committed income. Keep a recipient’s actual written response separate from a general market description.
GAO’s cost guidance specifies scope, schedule, assumptions and actual-cost updates. The stage method below is an original business-planning application of those principles. GAO does not endorse this pilot, its rates or seller earnings.
Worked example: an October decision record
This hypothetical pilot begins on 7 October 2026. It asks whether a bounded operational archive merits a named receiving-route discussion. No raw records leave the business. The sponsor assigns a $750 first-stage cap and 21 October review date. Only a passed evidence test can release the next stage.
Amounts are USD internal estimates. There is no entered offer, probability-weighted revenue or recurring delivery commitment. External license fees and VOID’s upfront seller referral fee are entered as zero. Specialist effort is illustrative, not a compliance-service quote.
| Stage | Entered work and cost | Evidence before further spending |
|---|---|---|
| 1: Describe/triage | Metadata 6 hours × $75 +review 2 hours × $150 = $750 | Package description, accountable owner and named-introduction permission |
| 2: Answer bounded question | Schema 8 hours × $80 +review 6 hours × $150 = $1,540 | Written recipient question plus a documented path to resolve material rights gaps |
| 3: Approved evaluation | Engineering 20 hours × $90 +external review$1,600 +QA 12 hours × $80 = $4,360 | Separate evaluation permission, defined scope and acceptance criteria; currently unfunded |
Record what actually happened at the gate
By 21 October the hypothetical review identifies the period, fields and system owner, but finds no signed historical contractor agreement for attached reports. A fictional recipient has asked only for a high-level description, without a defined evaluation question. The completed decision is pause: keep the $750 cap, exclude reports from the description and ask the business owner to locate the agreement. The scenario records $525 actually spent by that review: five metadata hours at $75 and one review hour at $150. The remaining $225 covers at most one additional metadata hour and one review hour.
Pause does not authorize an export while procurement searches. The sponsor sets 28 October as the next decision date. Staff may locate the agreement only within the remaining $225 allowance and its specified hours. Any extra work requires a changed budget or a stop; the plan does not assume additional staff time is free.
In this illustration, the remaining $225 is spent checking the retained project records by 28 October, without finding the required agreement or a concrete recipient question. The completed stop decision closes the pilot at $750. Keep the approved metadata and unresolved issue so a later genuine question can reopen it; do not continue preparation indefinitely under the label of business development.
Give go, pause and stop different consequences
Go funds only the next specified stage. Here stage two needs a concrete recipient question answerable by a scoped schema, plus a credible rights-resolution plan. Any actual disclosure still needs permission. A question alone cannot authorize a sample or license.
Pause requires an identified dependency, a named owner, a cost limit and a date. Stop applies when evidence will not arrive, the package cannot be authorized, the cap is reached or the use no longer fits the business. Stopping can be good execution even if staff discover useful internal records.
Do not average an unresolved material right into a readiness score. Record the blocker separately. Before the next review, the owner can confirm that an excluded attachment really stays out of the description and that no staff member has begun an unfunded export. A stop rule needs those operational consequences, not just a colored status.
Know the cost at each stopping point
If only stage one is spent, no-deal cost is $750. Completing an authorized second stage makes cumulative cost $750 + $1,540 = $2,290. Funding all three would make entered preparation cost $2,290 + $4,360 = $6,650. No total creates revenue or proves a payable offer.
Update actual hours before extending the plan. If metadata work takes nine hours instead of six, approving the next stage against the old total conceals an overrun. A future real offer needs its own comparison of payment triggers, deductions and continuing duties.
Use readiness to identify unanswered issues and the introduction brief for approved metadata. VOID’s initial role is possible fit, a permissioned named introduction and handoff coordination. Your sponsor retains the spending decision. A useful pilot ends with recorded evidence and a clear consequence, rather than a growing folder awaiting someone else’s enthusiasm.