River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Proof of Concept Plan Template

A pilot plan, a criteria agreement both sides sign, and the sizing sheet that shows which criteria your window can actually decide.

Free download  ·  No account needed

Success Criteria Agreement

[Buyer] and [Vendor] — Pilot Success Criteria

Window: [start] to [end], [n] calendar days
Scope: [team], [n] named users, about [n] [units] a day
Observations in window: about [n], against [n] in the baseline period
Decision maker: [name, role]    Decision date: [date]

C-01. [Criterion, in one sentence somebody can quote]

UnitOne [quote / ticket / onboarding / named user]
Observations
Baseline—  source and period
Passes if—  absolute, not only relative
Smallest difference this window can decide
Verdict available
Evidence

What is not a criterion

Anything the pilot going well would not change. One line each, with a named owner and a date, carried into the Pilot Plan.

What happens on [decision date]

On a pass. On a partial. On a fail. All three written before the pilot starts.

Every proof of concept template in circulation lands on the same instruction: make the success criteria specific, measurable and dated. None of them checks the second half of that promise. A criterion with a number and a deadline in it can still be undecidable, because settling it takes more observations than the pilot will produce. Nothing about the wording gives that away. So the pilot runs, the data arrives, and the last week goes on arguing about what the criterion meant.

Take a thirty-day pilot on one nine-person quoting desk that handles fourteen quotes a day, which is 420 quotes inside the window. Cutting the rework rate from 12 percent to 8 percent sounds like the most concrete criterion on the list. The count it needs depends on where the rate already sits, and here that is about 882 quotes on each side, or 63 days of this desk. Six criteria came out of that discovery, and two were decidable as the buyer wrote them.

This pack sizes each one before anybody signs it, then rewrites the ones the window cannot settle. The rework criterion becomes a back-test over 96 cases that already exist, decidable in week one instead of week nine. Open it in River and the agent reads your discovery notes, computes the counts and drafts the agreement. Or download the pack as Word documents and CSV sheets and size the criteria yourself.

Every document in the pack

The Criteria Sizing sheet, the Success Criteria Agreement, and the Exit Note.

Criteria Sizing

Calder Freight, Chicago quoting desk. 30 days, 9 named users, 14 quotes a day, so about 420 quotes in the window against 420 in the baseline.

IDCriterion as statedObservations in windowSmallest difference this window can decideVerdict
C-01Cut average quote turnaround by 25%420 vs 4200.81 h, which is 49 min or 12.5% of a 6.5 h baselineDecidable
C-02Cut the live rework rate from 12% to 8%420 vs 420Needs 882 per side, so 63 days of this desk. 420 separates 12% from 6.4%Rewrite
C-03Zero data-entry errors on carrier onboarding6 onboardingsWindow expects 0.75 faults. Zero of 6 leaves the rate up to 39%Rewrite
C-04Exchanges loads with our transport management system3 round trips, day 3One observation decides a capabilityDecidable
C-05Passes our security reviewNot applicableBounded by the reviewer’s calendar, not by pilot volumeDependency
C-0690% user satisfaction9 usersReachable values are 6/9, 7/9, 8/9, 9/9. None is 90%Rewrite
 6 criteria stated 2 decidable as written, 3 rewrites, 1 dependency5 of 6 after sizing

Success Criteria Agreement

Calder Freight — Pilot Success Criteria

Window: 14 April to 13 May, 30 calendar days
Scope: Chicago quoting desk, 9 named users, about 14 quotes a day
Decision maker: Priya Raghunathan, VP Operations   Decision date: 20 May

C-02. At least 62 of the 96 known rework cases no longer rework

UnitOne historic rework case
Observations96, all of them known failures, from Q3 and Q4
Passes if62 or more complete without rework, which is 64.6%
Why this replaced the originalCalder first asked for the live rate to fall from 12% to 8%. That needs about 882 quotes on each side, which is 63 days of this desk. The 96 cases exist today, so the back-test settles the same question in week one.
Verdict available24 April, week two
EvidenceCase-by-case result table, both sides present at the run-through

What is not a criterion

Security review. Bounded by the reviewer’s calendar, so it sits in the Pilot Plan dependency list with Marcus Adeyemi named and a target of 28 April. The pilot going well would not change the answer, which is the test.

Exit Note

Calder Freight — Pilot Result, 14 April to 13 May

Headline: Four criteria met, one not met. The one not met is C-01 and it missed on a definition, not on performance.

C-01. Quote turnaround falls by at least 25%

RequiredAt or below 4.88 h against a 6.5 h baseline
Observed417 quotes, average 5.31 h, an 18.3% reduction
VerdictNot met
WhySplit on request channel: portal went 5.4 h to 3.7 h, a fall of 31.5%. Email went 8.3 h to 7.95 h, a fall of 4.2%. Email is 38% of volume and is re-keyed by hand before it reaches the process at all.
EvidenceRows E-014 to E-021

The decision on 20 May

Amend C-01 to the portal channel and settle it from data already collected: 259 pilot quotes against 260 baseline, a 1.70 h reduction against a 1.35 h target and a 1.03 h floor. Met, today, with no extension. Or extend three weeks with email intake in scope. Or close.

What's in the pack

01

Criteria Sizing sheet

One row per criterion with its unit of observation, the count the window produces, the smallest difference that count can decide, and a verdict of decidable, rewrite or dependency.

02

Success Criteria Agreement

The document both sides countersign. Every criterion carries its unit, its count, the comparison that settles it and the date the verdict is available, because a standard travels with its method of assessment.

03

How a Criterion Is Sized

The four questions, the three rewrites, and a fully worked pilot where six stated criteria became five decidable ones and a dated dependency.

04

Pilot Plan

Scope with the volume that justifies it, named owners, the dependency table, data and access, and the week-by-week shape of the window.

05

Milestone Tracker sheet

Built backwards from the decision date, with days-before, what each milestone blocks, and the slack column that shows which four rows move the decision.

06

Criteria Evidence Log sheet

One row per observation with who saw it, who from the buyer was present, and where it can be found again, so a verdict is never a recollection.

07

Exit Note

One block per criterion with required, observed, verdict and evidence rows, plus the courses in front of the buyer on a partial result.

How to use it

  1. 1

    Open in River, or download it

    Open the pack in River and let the agent size the criteria, or download the blank Word and CSV files instantly with no account.

  2. 2

    Send the notes and the team's volume

    Discovery notes, the buyer's requirements list, the dates, and roughly how much work the team in scope gets through in a week. That last number is the one nobody writes down and the one everything else depends on.

  3. 3

    Size every criterion, then rewrite what fails

    Each criterion gets a unit, an observation count and a floor. Rate comparisons become back-tests over cases that already exist, rare-event targets become replays of the records where the fault already happened, and percentages on small teams become counts of people.

  4. 4

    Sign it, then log evidence as it arrives

    The agreement gets countersigned before kickoff and the Milestone Tracker runs backwards from the decision date. Every observation takes a row the day it happens, which is what the Exit Note is built from and what a close plan picks up.

Frequently asked questions

Is this template free?

Yes. The whole pack downloads as Word documents and CSV sheets with no signup and no credit card. Edit with AI is a separate, optional path for sellers who want the criteria sized and the agreement drafted from their own discovery notes. The template library holds the rest of the packs.

What does it mean to size a success criterion?

Name what one observation is, count how many the pilot window will produce at the team's real volume, then work out the smallest difference that count can separate from ordinary variation. The floor depends on the spread of what you are measuring and falls with the square root of the count, so halving it takes four times the volume.

What happens to a criterion the pilot cannot decide?

It gets rewritten, not dropped, and the original stays visible next to it. A rate comparison becomes a back-test over cases that already exist. A rare-event target becomes a replay of the records where the fault already happened. A percentage on nine users becomes a count of users, since nine people cannot produce 90 percent.

What format are the downloaded files?

Word documents (.docx) for the plan, the agreement, the sizing method and the exit note, and CSV (.csv) for the three sheets, zipped into one file. They open natively in Word, Pages, Google Docs, Excel, Numbers and Sheets, with nothing to convert. Inside River the same content opens as live Docs and Sheets.

How is this different from a generic POC plan template?

A generic template asks you to write measurable criteria and stops there. This one checks each criterion against the observations your window will actually produce, and hands back the count it needs in your own units: this many more days, or this many more teams. That is the conversation a champion can act on.

What does 'Edit with AI' actually do?

It creates a free River account and installs this exact pack as a private workspace, with the agent ready to read your discovery notes, the buyer's scored evaluation and the team's volume. Nothing is written until you send it something.

Where does this sit against the proposal and the close plan?

Before both. The pilot decides whether there is a deal, so the criteria come first. Then the proposal pack prices what the pilot proved, and the mutual action plan carries the dates from the decision through to signature.

Sign criteria the pilot can actually settle

Download the blank pack as Word and CSV files, or open this exact pack in River and have the criteria sized against your own numbers.

Edit with AI