Causal inference analytics for multi-site retail
Find out what your stores are capable of selling.
Your current reports show what your stores sold, and traditional benchmarking ranks them against each other. We compare each store with its twin to show what it should be selling, which of the initiatives running there at the same time are driving sales, and which are costing you margin.
The big chains build digital twins with cameras and sensors. Seurat builds a twin of every store from the sales data you already have.
Where your own history can't answer the question, we bring the answer key from chains like yours that have already made the change.
Two stores your reports get wrong
Both stores are illustrations.
HEARTLAND MARKETSLever BoardReads updated weekly
Store #217Maple & 9th, Cedar Rapids · fuel + kitchen + 2 EV chargersMeasured at 14 stores like this one. Ranked in order of opportunity for Store #217.
Store index · 0–100 · measured at 14 stores like this one
On the board at #217$154K/yr10 levers fit this store
Recommended to start now5 leversworth $95K/yr, no capex
Read maturingWeek 3 of 8: Eliminate loyalty clutter · verdict lands in ~5 weeks
Validated at #217$4.1K/yrCold vault reset (planogram): measured $4.1K vs. the $3.4–5.1K estimate. Inside the range ✓
Ranked by value delivered × confidence. All dollar figures are incremental inside sales, not gross profit. Reads draw on stores like this one plus transported findings from beyond the chain, and each row cites the source of each finding.
#LeverWhere it's provenWorth at this storeConf.Status
1Make the signature item identical, every visitStore-to-store drift is your #1 repeat-visit riskMeasured16 matched-store readsFieldShanghai store walksSurveyn=1,200 consumers$31Kper yearAStart now
2Cut the menu to what you're great atFewer items, tighter execution, faster lineMeasured9 matched-store readsFieldFarmingdale store walk$26Kper yearBStart now
3Two-Strike winback after a second bad visitHalf the shoppers who quit a store were gone within 2 bad visitsPOS12 chains, second-miss patternSurveyn=1,200 consumers$19Kper yearBStart now
4Keep serving the entire menu until closeLate-night crowds come specifically for pizza and friesFieldWarsaw · Kraków, 2am queues3 matched-store reads under way$17Kper yearBQueued
5Eliminate loyalty clutterPump screens + door posters run four competing offers todayMeasured11 matched-store reads$14Kper yearALive at #217wk 3 of 8 · verdict in ~5 weeks
Ruled out for #217 — measured elsewhere, doesn't pay here
—Go 24 hours at this locationWorks in nightlife zones. Maple & 9th goes quiet at 11.Measured13 stores that tried it$0skip it — save the laborARuled out
Built for your world
Convenience & FuelThe answer key for c-store decisions: foodservice, remodels, loyalty, the forecourt.Restaurant & QSRMenu, daypart, remodel and drive-thru levers, priced per store.EV ChargingWhat chargers actually do to the store, measured at sites that installed first.Private EquityCausal inference pinpoints what's wrong at a chain you already own. We direct the turnaround and monitor whether it's working.Ad Platforms & Retail MediaFor media sellers: Don't let your channel get miscounted in the analyses that set ad budgets. The correction can't come from you.Store Design & SuppliersWhat your remodels and installs actually do to the store, measured on projects you've already finished.
What we sell
Store-to-Store ReadStatistical inference on every review at every store, separating real differences between locations from noise and finding what drives them. Built from the public record, so we need nothing from you.Calibration ReadCausal impact analysis that isolates and quantifies the value of changes you've already made. We run 3 to 5 reads on your own sales data and grade each one on what it contributed.The Lever BoardOur flagship offering: the proven levers for stores like yours, ranked and priced. Every lever was measured at matched stores and valued for each of your stores, with a range, a grade and a status. After you pull a lever, we check what it actually did against our estimate and show you whether it landed inside the range.Remodel Sequence ReadYou already have next year's remodel list. We tell you what order to build it in, which entries shouldn't be on it, and what would have to change for the sites below the cut line.Predictive Sales MonitoringBetter than a benchmark: each site is read weekly against its twin, built for that site alone. The gap between the store and its twin is the part that's about how the store is run. $50 per site per month.
For vendors who sell into stores
Proof
Conexxus MemberThe Mid-Market Retail Intelligence Briefing
Subscribe for first access to the Causal Convenience Index, our latest consumer research, and new reads as the boards update. Free, monthly.
No spam. Unsubscribe anytime.