The OGC Blog

A Second Set of Eyes on Your Restaurant Tech Stack

Written by Matt Haselhoff | Aug 4, 2026, 3:26:51 PM

Sometimes we look at an operator's stack and tell them we've got nothing for them.

That happens more often than you'd expect, and it's the part worth saying first, because every "free assessment" in restaurant technology is a sales call in disguise.

The person doing the assessing has one product to sell, so the assessment finds a gap that product fills. You already know this. It's why the offer makes you tired.

So how is OGC different?

We represent a vetted portfolio of solutions instead of a single product, which changes what we can say. There's nothing we need you to buy. What we're good at is matching a real gap to something that fixes it and getting it stood up, and that shows up on the P&L.

Why does an outside read on your stack matter at all?

Because nobody owns the whole picture, and the numbers are unforgiving.

A fast-casual chain or QSR now commonly runs 20 or more systems, most of them bought at different times, by different people, for different reasons. Each one was the right call in isolation. Nobody is responsible for the shape they make together.

Meanwhile, the margin they're supposed to protect has thinned. Full-service net margin sits around 2.8% and limited-service around 4%, with prime cost eating 55% to 65% of revenue before rent, utilities, or insurance. At a 3% margin, a dollar you stop losing to overhead does the same work as about $33 of new sales. You don't have to sell your way out of a leak. You have to find it.

And the leaks aren't where the attention is. Food and labor get reviewed line by line every period, because that's what the P&L conversation is about. The categories underneath them mostly don't get reviewed at all.

What does a stack review look for?

Four things, and only one of them is about the tools you chose.

  • Categories you didn't know were categories. This is the big one, and it isn't a criticism of anybody's judgment. You can't shop for a solution you've never heard described. Most of what we find falls here: a cost bucket nobody assigned an owner to, because it never occurred to anyone that it was a bucket.

  • Money already earned and never collected. Delivery disputes are the cleanest example. Somewhere between 2.5% and 3% of an operator's revenue is tied up in disputes with marketplace providers, about a fifth of already-thin delivery profit. And more than 70% of restaurants never dispute at all, because the process is slow, manual, and per-order. That money is earned and uncollected.

  • Overhead that never gets the annual review food and labor get. Energy is 3% to 5% of sales, and restaurants are among the most energy-intensive commercial buildings in the country at roughly 4 times the average per square foot. Cooking, refrigeration, and lighting are about two-thirds of the bill. Telecom and connectivity across a multi-unit estate are the same story: real money, rarely reviewed, almost never renegotiated.

  • Renewal timing. Across industries, about 30% of software licenses go unused, and roughly half of the rest are used well below what's being paid for. Auto-renewal with a built-in price uplift is why that persists. A tool that stopped pulling its weight 8 months ago renews at a higher rate than it did last year, and nobody notices because nobody is watching the calendar.



What did we find for operators this year?

We've run this for dozens of brands. Here's a sample of what came back.

A 60-unit fast-casual chain recouped $39k in delivery chargebacks and refunds over the first half of 2026. That's about $6.5k a month, and they kept $29k of it after fees. The program cost nothing up front. It worked for a structural reason: their dispute rate was running 3.2% of orders against a 2.2% marketplace benchmark, and once someone was actually working the disputes, they won more than 55% of what they filed. Set that against the 70% of operators who never file, and luck stops being the explanation. The money was sitting there the whole time.

On energy, we cut operator costs 15% in July, measured against their own electric bills instead of a projection. And one of our partners saves about $10k per new build-out, which matters if you're opening at any pace at all.

The chargeback numbers did something we didn't plan for. We shared them with a CFO who wasn't in market for anything. Two corporate locations went into a pilot inside a month, and the paperwork was signed the following week. Nobody was sold. Somebody was shown a number they hadn't seen before.

What this isn't

It isn't rip-and-replace. Ripping out a working stack concentrates every risk into one quarter and delays payback on all of it. Our whole argument is the opposite: one proven pick at a time, each one's savings funding the next, which is the case we make in detail in our post on building margin one pick at a time.

We're also not here to grade your technology decisions. If you're a CTO or a VP of IT, you already know your stack better than we'll know it after one conversation. What we bring is the pattern across 150+ brands and 12+ technology verticals, which is a different thing from knowing your building.

Most reviews surface one or two things worth a conversation. Some surface none.

And this isn't an argument about AI. Around three-quarters of brands are investing in AI now or inside the year, and that spend is going to keep climbing. A review is how you make sure the plumbing underneath it can hold the weight. Integrations are where most of this breaks, and we wrote about that in our piece on integrations. Fundamentals and AI aren't a sequence. You can work both at once, and both are on the table.

What happens if you say yes?

You rattle off your stack. POS, ordering, delivery, kitchen, back office, connectivity, and whoever handles your energy and your telecom. It takes about 2 minutes, and you don't need to pull a report.

We come back with a short list of gaps worth looking at, sized against what they'd be worth to you, or we tell you there's nothing here at the moment. No data room, no NDA, no discovery deck.

FAQ

Is this an audit?

No, and we're careful about the word. An audit implies we're checking your work and expecting to find it wanting. This is a second set of eyes on a stack nobody has looked at as a whole, which is a different exercise with a different tone.

How long does it take?

The conversation is about 2 minutes of you listing systems. Getting back to you with something specific takes a few days, because we're checking your categories against what we've placed elsewhere and what it returned.

Do I have to share financials?

Not to start. A list of systems is enough to find category gaps. Sizing a specific opportunity eventually needs the relevant bill or statement, which is also the only way to measure savings against something real instead of a vendor's projection.

What if we're already covered?

Then we say so. That's the outcome about a third of the time, and saying so is worth more to us than a manufactured gap.

Who should be on the call?

Whoever can list the systems. Often that's a VP of Operations or an IT lead. It doesn't need to be a buying conversation, and it usually isn't one yet.

Where One Goal fits

We sit between enterprise operators and a portfolio of pre-vetted technology partners, and we make the introduction that fits instead of pitching one product. Most vendors' assessments only end one way. Ours can end with "you're covered."

Want to learn more? Book a short conversation, and we'll tell you what we see, including if what we see is nothing.

 

Matt Haselhoff is a co-founder of One Goal Consulting. He has spent 27 years in restaurant technology, including as CRO of Omnivore before its acquisition by Olo.