Pilot playbook
Exactly how a Gaincraft pilot runs.
Thirty days of setup that are not counted, then sixty counted days, then a decision. Published in full because a procurement team should be able to evaluate the programme before speaking to us, and because a vendor unwilling to write this down is telling you something.
Everything on this page is what we commit to. Where we cannot commit, it says so.
Shape
Ninety days, only sixty of which count.
Scoring a store while hardware placement, taxonomy tuning and the floor announcement are still settling produces a baseline nobody should trust. So the first thirty days are excluded from the readout by design, not as a concession.
Days 1–30
Setup, not counted
Legal scoping, installation, taxonomy calibration, floor announcement, cohort assignment written and fixed.
Days 31–90
Sixty counted days
Coaching live for the coached cohort, capture-only for the control cohort, groups swapped at the midpoint.
Day 90
Decision review
A readout against point-of-sale, presented whether or not it supports a renewal.
Week by week
What happens, and when.
Written for a procurement team. Each week states what we do and what you have in your hands at the end of it.
Week 1
Setup
Legal scoping and store selection
What we do
We run the scoping session with your legal and compliance team before any demonstration: capture zones, default exclusions, retention window, access policy, signage placement. The DPA is executed.
What you have at the end of it
A signed zone map per door, a configured retention window, and a written record of who may open a recording.
Week 2
Setup
Installation and signage
What we do
One installation visit per door. Counter devices mounted and powered, indicator verified, signage kit placed at entrance and at each capturing counter, QR notice live.
What you have at the end of it
Capture armed in the agreed zones, and a walkthrough photo record of signage in each store.
Week 3
Setup
Taxonomy calibration on your own floor
What we do
We score early conversations, compare against human double-scoring, and tune the behaviour definitions to how your customers actually talk. Definitions that produce disagreement are rewritten now, not later.
What you have at the end of it
A behaviour library for your vertical, adjusted to your assortment and your market, with inclusion, exclusion and eligibility rules written down.
Week 4
Setup
Floor announcement and cohort assignment
What we do
We brief your store managers to announce it, and we do not announce it ourselves. Cohorts are assigned and frozen in writing: who is coached, who is capture-only, and the midpoint swap date.
What you have at the end of it
A written cohort assignment, a manager briefing pack, and the first coverage report so you can see capture is working before anything counts.
Weeks 5–8
Counted days 1–30
Coaching live, first behaviour movement
What we do
Daily prompts to the coached cohort. Weekly scorecards to store and district leadership. Weekly thirty-minute call, same slot, same people.
What you have at the end of it
Weekly coverage and behaviour scorecards, the first clips from your own top quartile, and the dispute log as it opens.
Week 9
Counted days · midpoint
The cohort swap
What we do
Coaching moves to the previously capture-only group and pauses for the first. This is the test of whether the effect follows the coaching rather than the people.
What you have at the end of it
A midpoint readout: behaviour movement in cohort one across thirty counted days, against a control that shared the same door.
Weeks 10–13
Counted days 31–60
Second half, second cohort
What we do
Same cadence, reversed groups. Point-of-sale reconciliation runs in parallel so the day-90 numbers are not assembled in the final week.
What you have at the end of it
Continued weekly reporting, plus early sight of the point-of-sale comparison rather than a surprise at the end.
Day 90
Decision
The decision review
What we do
We present behaviour movement and the point-of-sale comparison for both cohorts, name any store excluded for low coverage, and set out what did not work.
What you have at the end of it
A readout you can take to finance, and a decision: proceed, extend, or stop with the clip library and taxonomy retained.
What we need from you
Seven things, and one of them ends the conversation if it is missing.
We would rather decline a pilot than run one that cannot produce an auditable number.
- Point-of-sale or CRM access at ticket level
- Non-negotiable. Without transaction-level data there is nothing to measure behaviour against, and any result we produce would be unauditable. A pilot without it does not start.
- A named executive owner
- One person with authority over the district in scope, who attends the weekly call. Programmes sponsored by a committee stall in week three.
- A defined store set
- A single district or an explicitly listed set of doors, chosen with us rather than for us, so that cohort assignment is defensible.
- Roster and shift data
- Attribution is by capturing seat and roster shift. Without accurate rosters, attribution degrades and disputes rise.
- Legal and compliance participation from week one
- Scoping happens before the product demonstration. Zone maps, retention window, access policy and signage placement are decided by you, and where your position is narrower than ours we take yours.
- Floor announcement by store leadership
- Announced as a coaching programme with commission evidence attached, by your managers, not by us. This single decision predicts acceptance more than anything in the product.
- Physical access for installation
- Counter mounting points and power. Devices sit on the counter, take no store LAN credentials by default, and are installed in a single visit per door.
What you get, every week
Five deliverables, on a fixed cadence.
None of these arrives only at the end. A pilot where the evidence appears in week twelve is a pilot you cannot steer.
- A weekly coverage report
- What proportion of eligible conversations were captured in each door. Published whether or not it flatters us, because coverage is the denominator behind everything else.
- A weekly behaviour scorecard
- Per store and per associate, with the weakest behaviour identified and coaching clips attached.
- A thirty-minute weekly call
- Same slot every week, with the same people. Agenda is coverage, behaviour movement, and anything the floor is telling us.
- A clip library from your own floor
- Built from your top quartile, yours to keep and reuse in your own training regardless of what you decide at day 90.
- The dispute log
- Every challenged score, the outcome, and which definitions we rewrote as a result.
Day 90
What the decision review actually contains.
A behaviour movement readout per cohort, per store, across the sixty counted days, with capture coverage stated for every week including the weak ones. Then the same comparison against point-of-sale at ticket level: close rate, average ticket and attach rate for the coached cohort against the capture-only cohort, including the effect of the midpoint swap.
Where a store’s coverage was too low for the cohort comparison to mean anything, that store is excluded from the readout and named as excluded. It is not averaged in quietly.
We also present what did not work: definitions we rewrote, behaviours the floor rejected, and any place where the taxonomy imported from another market failed to describe your customers.
If the readout does not justify the spend, that is what it will say. We would rather lose a renewal on an honest number than defend a weak programme for three years.
Three outcomes
Proceed
Chain deployment, scoped from what the pilot taught us about your floor rather than from the original proposal.
Extend
Where coverage or seasonality made sixty days too thin to decide on, we extend at the pilot arrangement rather than converting on a weak number.
Stop
You keep the clip library and the behaviour taxonomy built from your own top performers. Audio is deleted on the retention schedule already configured, and derived data is exported or destroyed on your instruction.
Scope a pilot.
Bring your district, your quality assurance spend and your counsel. The first hour is scope, consent and retention, not a product tour.