[ Pilot ] One campaign rehearsal
One campaign. Two weeks.
Bring the campaign you're planning. We first run one of your past campaigns blind, show you how close we got, then rehearse the next one in 3-5 versions. Fixed price, shared when we invite you from the waitlist.
01Days 1-2: Kick-off call, data check, you choose the past campaign for the blind test.
02Days 3-6: Blind run on the past campaign: prediction locked, then compared with the real result.
03Days 7-11: Rehearsal of your next campaign: 3-5 versions.
04Days 12-14: Decision memo and review call.
Day 1 → 14
[ 01 ] What you get
A blind test first, then the rehearsal.
- 01
A blind test on your own past campaign.
You choose one of your last 2-3 campaigns. We only see data from before it started, lock our prediction, then compare it with what really happened. You see the miss, whichever way it goes.
- 02
A rehearsal of your next campaign.
3-5 versions (e.g. 20% off vs. "2 for x" vs. a smaller discount), each with units, revenue and gross profit, including the weeks after the campaign.
- 03
The why.
Which effects drive the difference: promo lift, pull-forward, post-campaign dip, cannibalization within your own range.
- 04
A written decision memo.
Plus a 45-minute review call.
[ 02 ] What we need
A CSV export is enough.
No integration, no access to your seller account.
- Product × day (or week) sales for at least the last 12 weeks: date, SKU, units, net price, list price, campaign flag and mechanic, unit cost.
- Channel per row if you sell on more than one marketplace.
- Optional: platform-funded discount, stockouts, which products substitute each other.
- Not needed: any customer data. No names, addresses, phones or order numbers.
[ 03 ] Timeline
Two weeks, four steps.
- Days 1-201Kick-off call, data check, you choose the past campaign for the blind test.
- Days 3-602Blind run on the past campaign: prediction locked, then compared with the real result.
- Days 7-1103Rehearsal of your next campaign: 3-5 versions.
- Days 12-1404Decision memo and review call.
[ 04 ] Honest limits
What to know before a pilot.
- Accuracy is not yet backtested at scale. Your pilot's blind test is part of how we measure it.
- Free-shipping thresholds and cross-product bundles are not modelled yet.
- Jonbar recommends; you decide and run the campaign. We don't touch your store.
- Your data is deleted on request and kept no longer than 90 days.