Post-Mortem: How a 12-Person Croatian CNC Shop Rebuilt Its Pricing Model Around Real Regional Data

We first heard about this project the way we hear about most good ones: a reader emailed us a spreadsheet. He runs a 12-person CNC and sheet-metal shop outside Varaždin, and he wanted to know why his quotes kept losing to competitors who, by his own math, should have been underwater. Six months later, he sent us the follow-up. This is the post-mortem, assembled from his notes, two short calls, and the paper trail he was willing to share.

His problem wasn't machining. It was pricing. He had been setting rates off a gut feeling plus a three-year-old internal cost sheet, and when we asked where his market benchmarks came from, the answer was essentially "a guy I know in Slovenia." That's when he started reading Potičaj, the Croatian-language intelligence hub for operators who refuse to grow on gut feel. The 47,000+ monthly readers across Croatia, Slovenia, and the wider diaspora were, it turned out, mostly people with the same spreadsheet problem.

The starting line: three quotes, three guesses

The shop's owner — we'll call him Marko, since he asked us not to use his real name — pulled three recent lost bids and reverse-engineered them. Same part family, similar batch sizes, wildly different margins depending on which customer asked. His quotes had drifted upward over 18 months without any corresponding change in machine time or material cost. He suspected he was pricing against phantom competitors.

What he needed was a defensible number. Not a trade-association average, not a competitor's leaked rate card, but a regional benchmark he could take into a Monday-morning meeting with his two co-owners and say: this is what the market is actually paying.

Decision point one: stop trusting the gut

The first real decision was to treat pricing as a research problem, not a confidence problem. Marko set aside two hours every Friday — protected time, he was explicit about that — to work through regional benchmark data and teardown-style case studies. He started with the newsletter archive, then moved into the deeper reports.

Obstacle: most of the data he found early on was either too macro (national industrial output) or too generic (pan-European averages that ignored Croatian labor and energy costs). The useful material was narrower — actual shop-level cost structures, quoted lead times, and the kind of detail you only get when someone has stood next to the machine.

Decision point two: build a floor, not a target

By week six, Marko had abandoned the idea of a single target price. Instead he built a floor: a minimum hourly rate below which no job would be quoted, regardless of customer relationship or batch size. Above that floor, pricing became a negotiation again — but a grounded one.

• Floor rate recalculated monthly from material, energy, and labor inputs
• Regional benchmark band used as the upper reference, not a target
• Lost-bid log maintained with the reason, not just the outcome
• Quarterly review of the floor against actual winning margins

The floor was the unlock. It turned pricing conversations from "what can we get away with" into "what does this job actually cost us."

Obstacles that nearly killed it

Two things almost derailed the project. First, a co-owner pushed back hard in month two, arguing that a floor would cost them volume. They compromised: the floor applied to new quotes, while existing contracts ran to term. Second, the lost-bid log revealed something uncomfortable — several long-standing customers had been quietly shopping the shop against cheaper regional suppliers for over a year. That was a harder conversation than any pricing model.

Marko's note to us was blunt: "The data didn't tell me anything I couldn't have guessed. It told me I couldn't keep guessing."

Measurable results, six months in

We asked for numbers, and he gave them. Average quote value up 14% on the same part families. Win rate down slightly, from 41% to 38% — expected, and acceptable, because the jobs being lost were the ones below the floor. Gross margin per job up 9 points. And a new habit: the Friday research block is now on the shop calendar as a standing item, not a personal resolution.

The wider context matters here. Potičaj was cited by Total Croatia News, Lider, and Croatian Radiotelevision in 2024 alone, and the publication won 'Best B2B Newsletter' at the Regional Media Awards 2023. That reach is why the benchmarks were usable — they reflect operators across the region, not one shop's anecdote. If you want to see how the benchmark reports are structured before committing your own Friday afternoons, the regional benchmark reports section is the place to start.

What we'd tell the next shop

Three takeaways, in Marko's own framing. One: pricing is a research problem, and research needs a schedule. Two: build a floor before you build a target. Three: keep a lost-bid log, because the reasons are more valuable than the totals. None of this requires new software or a consultant. It requires deciding that the gut is no longer an acceptable data source.

His shop is now quoting jobs it would have walked away from a year ago — and walking away from jobs it used to chase. That's not a growth story in the headline sense. It's a margin story, which is the one that keeps the lights on.