Tank outage monitoring for fuel marketers and carriers

What an outage costs

The call usually comes after the second runout in a month, not the first. One dry tank is an apology; two is a pattern, and patterns end up in contract reviews.

When a c-store tank runs dry, the store loses pump revenue for every hour it’s down, sells less inside the store because fuel is what brings cars in, and remembers exactly whose name is on the fuel contract. The carrier eats the blame regardless of whose forecast missed.

The frustrating part: outages almost never come from missing data. They come from data nobody was looking at.

Why tanks run dry in the age of telemetry

The warning signs exist. Stick readings get taken. Many sites have automatic tank gauges streaming levels. Orders, deliveries, and demand history all live somewhere. The problem is that “somewhere” is four different systems owned by three different parties, and no one of them sees the whole picture.

So the practice becomes reactive. The store calls when the tank is critical. Dispatch scrambles a load. Everyone treats it as bad luck instead of what it usually is: a site that had been trending toward empty for weeks, visibly, in data nobody had assembled.

What working outage monitoring looks like

When we sit down with a carrier’s operations team, the ask is never another report. It’s a standing view someone opens every morning, built on all the sources at once. The components that earn their place on that screen:

Outage counts and trends

The basic discipline: how many outages, where, and which way the line is moving. An outage count without a trend line is trivia; the trend is what tells you whether the operation is getting better or worse.

Stick-level distributions

The leading indicator. Across every site, what does the distribution of tank levels at reading time look like? A site whose average stick keeps drifting toward the bottom of the tank is a runout that hasn’t happened yet. Watching the distribution turns outages from surprises into a queue of sites to act on.

Views by the dimensions that drive decisions

By market sector, by store, by owner, by product. The same outage data answers different questions for different people: which owner’s portfolio is chronically tight, which product runs dry where, which market needs its delivery cadence rethought. If the view can’t be cut by the dimension a decision needs, the decision gets made on anecdote.

An accountability loop

The dashboards only matter if a person owns what they show. The sites trending wrong need a name attached: who’s adjusting the ordering pattern, who’s calling the owner, and who confirms at the next review that it moved.

The build reality

The dashboards are the easy part. The work is underneath: pulling stick readings, gauge feeds, orders, and deliveries out of the systems they live in and into one place where the joins are tested and trusted. That’s a data platform problem, and it’s solvable at mid-market scale; we built exactly this for a fuel carrier, unifying six source systems and putting outage counts, trends, and stick distributions across 5,000+ sites on one screen.

Once the foundation exists, outage monitoring is one dashboard among several: the same platform watches toll compliance and delivery-window accuracy from the same unified data. The foundation is the investment; every dashboard after it is cheap.

In every one of these engagements there’s a moment when the team pulls up a past runout and finds it was visible three weeks early. The data was there; the view wasn’t. AI & Automation is the practice that builds the view.