How a live crude-market feed, a third-party competitor pricing service, and two configurable variables replaced a manual, once- or twice-a-day price check with a system that watches, quantifies, and alerts — continuously.
Retail fuel margins move on small numbers — a few cents a gallon, multiplied across thousands of gallons a day, per site. Before this tool existed, the pricing team's read on the competitive landscape came from manually checking nearby stations, with no fixed definition of "nearby" and no consistent trigger for when a move was actually worth reacting to.
Nobody was doing this badly on purpose — there just wasn't a system watching continuously. A competitor could move their price at 6am, and the first anyone at Walmart Fuel & Convenience found out was whenever someone next happened to look.
Reacting correctly to a competitor's price change required answering two questions at once, and the manual process could only ever answer one of them:
The pricing team didn't need someone checking more often. It needed something that never stopped checking at all.
The engine replaced the manual scan with two inputs the pricing team controlled directly, checked continuously against a live competitor feed, with a market backdrop and a dollar estimate attached to every alert.
The competitive radius around a given site — defines which competitor stations count as "in the set" for that location.
The tolerance threshold, in cents (e.g. 8¢, 10¢, 12¢), a competitor's price had to move — relative to the site's current price — before it counted as a breach.
WTI + Brent (macro context) and Kalibrate competitor prices (local layer), continuously
Combines those feeds with Walmart's own fuel pricing, traffic & gallons-sold data into one dataset
For each site, check competitors within radius X against the site's current price
Has any competitor's gap exceeded threshold Y since the last check?
Estimate the $ cost of not repricing, using site traffic & gallons sold
Power BI dashboard + mobile push, naming the competitor and the delta
Every alert was self-explanatory: which competitor moved, by how much, in which direction, at which site, and roughly what leaving it unaddressed was costing per day — so the pricing team could act, and prioritize across sites, without first going to verify the situation themselves.
Below is a recreation of the shape of a real alert, so the mechanism is concrete rather than just described.
| Manual competitor check | Smart Pricing Engine | |
|---|---|---|
| Monitoring frequency | A few times a day, whenever someone checked | 24/7, continuous |
| Definition of "nearby" | Inconsistent — varied by person and day | Fixed, configurable radius (X) |
| Market context | Judged by feel | Live WTI + Brent backdrop |
| Urgency of an alert | All moves looked equally worth reacting to (or not) | Every breach comes with an estimated $ impact |
| Where the alert reaches you | Wherever you happened to be looking | Power BI dashboard + mobile push |
The bigger shift wasn't just automation for its own sake — it was turning "a competitor changed their price" into "here's what that's costing us today, at this site, and here's how it ranks against every other open alert." That's what let a small pricing team stay ahead of a market that was never going to wait for a manual check.