Methodology

Causal incrementality.
Not correlation.

Difference-in-difference is the same methodology economists use to measure real-world causal effects. PathX runs it at campaign scale, on actual retailer POS data, every week.

Difference-in-difference

The gap between actual and expected is the lift.

DiD compares a test group's actual sales against what those same stores would have done if they had only followed control-group trends. The gap between actual and expected is the incremental lift: not a guess, not a model output, but the measured difference.

Every result is evaluated at the SKU and store level, not rolled up to a single category number you can't break apart in the buyer's room.

  • Test line: test stores' actual weekly sales.
  • Expected line: test stores' sales projected from the control-group coefficient (what they'd do absent the campaign).
  • Shaded area: measured % lift during the campaign window.
Test & control setup

Control stores aren't picked at random.

A randomization algorithm selects control stores constrained on four dimensions simultaneously, so the matched group is as close as possible to the test group before the campaign starts.

Item distribution
matching criterion
Historical sales volume
matching criterion
Parallel trend
matching criterion
Geographic dispersion
matching criterion

Control group is typically 8-20% of total measured locations. Test/control groups are built and delivered within one week of campaign launch. Store-level measurement is the best practice; regional and channel-level reads are available.

Panel vs. point-of-sale

Panel data extrapolates. POS data measures.

Privacy-friendly, transparent, and retailer-aligned, because the number comes from the retailer's own register.

Typical competitorsPathformance
MethodologyMatch-market or panel-basedDifference-in-Difference on retailer POS
Data basisExtrapolated from a small sample1:1 transaction-level impact
SetupHigh minimums, rigid parametersQuick feasibility, flexible study design
StrategyPay-per-view data cutsPre-campaign insights included; in-flight optimization on 7+ week campaigns
How it works

Four steps from RFP to results.

01

RFP submission & campaign details

Submit an RFP approximately 4 weeks before launch. Feasibility is confirmed within 48 hours. Pathformance returns recommendations on budget, flight length, and store count.

02

Setup & launch

Test and control groups are identified, matched, and delivered approximately 5 business days before campaign launch. Best practice: store-level data, up to 3 retailers, minimum 4-week campaign.

03

In-campaign optimization

For campaigns running 7 or more weeks, PathX provides mid-flight visibility so underperforming stores can be cut and budget reallocated before the campaign ends.

04

Results & learning

Final reports delivered approximately 4-5 weeks post-campaign, including a client-success review with you and/or your client.

BEFORE YOU SPEND

Submit an RFP. Feasibility is confirmed within 48 hours.

Best-practice study design: store-level data, up to 3 retailers, minimum 4-week campaign. Pathformance returns recommendations on budget, flight length, and store count to set the campaign up to clear a defensible result before you commit the spend.

Request a feasibility check
What PathX delivers

Six outputs. All defensible.

Incremental dollarsiROASSales lift %Category performanceItem (hero/halo) performanceGeographic performance
Ready when you are

See what a PathX study design looks like for your campaign.

Tell us the retailer, the flight, and the meeting you're preparing for. We'll confirm feasibility within 48 hours.

Solutions

What we measure.

Sales Lift

Sales Lift Measurement

Week-over-week incremental sales lift at the SKU and store level, sourced from retailer POS data. The number your category buyer needs to see.

Lottery

Lottery Measurement

Causal incrementality measurement for lottery and gaming campaigns, isolating true lift from baseline ticket sales across retail locations.

Media

Media Measurement

Cross-channel media incrementality tied directly to in-store sales outcomes. No walled-garden estimates.

Data Sources

Where the data comes from.

Walmart
Point-of-sale scan data from Walmart store network.
Kroger
Retailer POS data across Kroger-owned banners.
Albertsons
Transaction-level data across Albertsons and affiliated banners.
Convenience & Gas
C&G channel data across major convenience networks.