For twenty-five years the assay itself got better and cheaper. Sequencing chemistry, PCR, dry-lab automation, all faster every cycle, and that progress isn’t slowing down. What never kept pace is everything it takes to actually run it at scale: the prep, the handoffs, the integrations, the people, the plumbing that has to line up before a single test runs. Almost none of that is a capital purchase you can point to. It’s operating cost, the unglamorous kind, so it gets starved while the capital goes to the next instrument. And it’s exactly where the economics of the whole operation get decided. Treat it on par with the instrument itself as critical infrastructure, and that gap becomes the biggest untapped opportunity for growth in molecular testing.
Over my career I saw the same thing from three very different seats on the same workflow: the bench of a clinical genomics lab, the clinical applications supporting diagnostics platforms, and the software that ties a lab together at scale. People picture a molecular workflow as a straight line, sample in, result out. It isn’t. The work loops back on itself constantly with branching and parallel order flows. Flow cells re-queue, samples fail QC, instruments drift out of spec, samples get processed together in trios, a reflex test throws work back upstream and resets what everything downstream was expecting. The instrument works, the software runs, the assay is validated, and the thing that actually breaks is the space between them. The most expensive failures I saw were almost never a broken part. They were broken handoffs nobody owned.
And when you rely on roll-up P&L and educated guesses, you get the wrong number. Ask what a test costs and you’ll hear the assay, or the sequencing run, net of reimbursement. It’s a real figure, and it’s the one everyone compares on, because it’s clean and lines up next to the next option. It just answers a smaller question than the one that matters. Think of a listing for a house. $300K, fair for the neighborhood, so you buy it and move in, and then you start opening walls: the wiring’s not to code, the pipes are shot, and all the work behind them has to come out and go back in. None of it was in the listing. A hundred grand later the place finally works, a third more than you signed up for, and you may never get it back when you sell. A test runs the same way. The quoted cost is the list price. Everything it takes to actually run it at scale, the handoffs, the integrations, the people, the rework, is the part no list price includes, and it’s where costs add up fast.
The reason that renovation never makes it into the quote is a split that’s specific to our field. The bench, the business, and the software that runs them are three different worlds: different language, different training, different data, different metrics. The people who run the assay and the people who hold the P&L often speak past each other, and translating between them isn’t anybody’s job. So the lab gets held to economics it can’t see, and the business makes six- and seven-figure platform calls on a workflow it doesn’t really understand. The number that should decide the whole thing is the one no one is set up to build.
Labs running live clinical work on custom code that nobody still employed had written, nobody could fully explain, and nobody could fix without paying someone from outside to come figure it out. Single points of failure sitting inside regulated workflows, held up by institutional hope instead of institutional knowledge. Read more on laboratory technical debt →
Four patterns that repeat everywhere
Throughput scales linearly. Complexity doesn’t. Two labs in the same segment, same modality, same volume, can be completely different operations underneath, because no two labs build a workflow the same way. There’s no universal SOP for NGS. Maybe there should be. Until there is, benchmarking yourself against an industry average just hides the problem. The only honest place to start is your own current state.
Most of the delay in standing up a new molecular line, NGS or a new PCR panel, was avoidable. A real pre-implementation assessment, one that looks at IT, LIS/EHR integration, the workflow itself, and staffing, finds the gaps while they’re still cheap to fix, instead of at go-live when they’re the thing killing the business case.
A lab that can’t put a number on its own value looks like a cost center, however well it’s run. The figures that go upstairs are all COGS, with nothing tying them to profit. Siloed data and broken-up order flows make the financial case harder to build than it has any right to be. So the investment doesn’t come, the capability erodes, and the same conversation happens again next year.
Someone owns the LIMS. Someone owns the instruments. Someone owns the LIS integration. Someone owns the business case. Nobody owns what happens between them, and that’s exactly where it breaks: in the handoffs, the dependencies, the assumptions one system quietly makes about another. That missing whole-system ownership is the most consistent thing I’ve seen in every lab I’ve been in.
The one thing I’m sure of is that you can’t fix what you haven’t measured, and your own current state is the place to start. That’s what the Quick-Check is for.
Take the free Lab Readiness Quick-Check →Built for the problems nobody owns
I built HelixWrks for the people running the lab, not the people selling to it. Labs don’t fail because they bought bad technology. They fail because nobody scoped the real workflow, nobody put a real number on it before the PO went out, and the gap got filled with improvisation until the costs ran well past every estimate.
That’s the gap HelixWrks closes, and it starts with a clear look at your current state. The Lab Readiness Assessment maps the hidden operational gaps, the integrations, the handoffs, the readiness nobody pressure-tested, across the ten domains that decide whether a change succeeds. It’s live today, and it’s the place to start.
Then Fulcrix costs your whole workflow end to end, so you can model a change before the PO is cut instead of hoping it pencils out after. It meets you at the detail you have: start high-level, sharpen as you go. In early access now; every lab in early access is a Founding Lab.