Hartwell runs at a 9.2% margin on $8.42M in revenue — about $780,000 of operating profit — yet the cash that profit should produce keeps disappearing, and your line of credit sits stuck at 64% with under two weeks of headroom. A large part of the answer is on your shop floor: 68% of your maintenance is reactive, machines get fixed after they break, and that drove 264 hours of unplanned downtime last year. The bill you watch — the $148,000 maintenance budget — is the wrong number. At your two-shift production rate, those 264 lost hours are roughly $534,000 of product you never shipped, which makes the full cost of reactive maintenance about $563,000 a year. This engagement stands up a simple maintenance system that moves you from 68% reactive toward 40%, so the machines that make your money stop stealing it back — with the first measurable downtime reduction visible in 60–90 days.
This brief describes everything about this engagement. When you're ready to proceed, go to the Confirm tab.
Hartwell is a 22-year-old precision CNC machining and fabrication shop — 47 people, 31 of them on the production floor, running two shifts out of a single facility. The work is demanding: machined components and sub-assemblies for aerospace, defense, and industrial OEM customers, held to tolerances of ±0.0002". That is not commodity machining. It is the kind of work that earns repeat programs and long customer relationships, and roughly 60% of Hartwell's volume is exactly that — repeat work — with the other 40% new programs coming in the door.
In a job shop at this tolerance, the business competes on two things at once: the ability to hold spec, and the ability to deliver on time without blowing up cost. Both run straight through the machines. When a spindle is down or a job is stranded mid-process, it is not an inconvenience — it is margin walking out the door, because the revenue rate on a two-shift floor is high and the capacity is finite. That is what makes the maintenance finding matter: in this sector, machine availability *is* the product. Over 22 years Hartwell has built a real franchise — a repeat-customer base in regulated, hard-to-serve markets. The question this engagement answers is why a franchise that profitable is so cash-tight.
You described it as "tight but manageable," and that's fair — but your own words went further: cash is tighter than the revenue suggests it should be, you're drawing on the line of credit more often, and it isn't getting paid down between draws the way it used to. You weren't sure whether that was collections, inventory, or something on the floor. The honest answer is that it's partly all three — but underneath the stated reasons, the diagnostic found a structural driver you couldn't see from the reports you have. Your profit is real; it's just being converted into stranded work-in-process and lost machine hours before it ever becomes cash. The line of credit, stuck at 64% with $480,000 drawn against a $750,000 limit and 1.9 weeks of headroom, isn't a separate problem — it's where every operational leak comes to rest as a single number.
Your stated goal was clear: understand exactly where the cash is going, fix the right thing first, and get the LOC back to a tool you choose to use rather than something you depend on to make payroll. That outcome is achievable — because the problems the diagnostic found are operational and addressable, not market or pricing problems you can't control.
Reactive maintenance is costing Hartwell an estimated $46,924 every month it continues — not in repair bills, but in production you can't ship because a machine was down. Left alone for the next twelve months on the current trajectory, that's better than half a million dollars of capacity lost while the line of credit stays pinned and the headroom stays thin. The opportunity is that this is a recoverable problem with a clear path: a 22-year-old shop with loyal senior operators and 60% repeat work has exactly the stability a maintenance system needs to take hold. The diagnostic has named the problem and quantified it. The timing is right because every month of delay is measured in lost machine hours you don't get back.
The data establishes the starting point plainly: 68% of Hartwell's maintenance is reactive — machines fixed after they break, not before — and that produced 22 hours of unplanned downtime a month, 264 hours across the year. The maintenance budget is $148,000. But the budget is not the cost. The cost is what those 264 hours would have produced.
Here is the cascade. The primary direct cost is the premium you pay to fix things reactively instead of on a schedule: emergency repairs run about 40% more than planned ones, which on the reactive share of your spend is roughly $28,754 a year. That's the number that looks like "the cost of reactive maintenance" — and it's the small one. The first-order downstream cost is the production lost while the machine sat down: 264 hours at a two-shift revenue rate of about $2,024 an hour is $534,336 of product you couldn't ship. That's where the real money is. The second-order effects ripple outward from there — stranded jobs become work-in-process that piles up between operations (42% of your $1.085M inventory is WIP, unusually high for a job shop), and the scramble to catch up feeds the 38 hours of weekly overtime and the expediting that drains operating cash on top of the capacity it already cost you.
| Cost Category | Annual Amount | Notes |
|---|---|---|
| Reactive repair premium | $28,754 | 40% premium on the reactive share of $148,000 maintenance spend (68% reactive) |
| Lost production throughput | $534,336 | 264 downtime hrs/yr × $2,024/hr two-shift revenue rate (revenue ÷ 4,160 two-shift hrs) |
| Total annual cost of reactive maintenance | $563,090 | All figures from Hartwell operating data |
Set that $563,090 against your cash position. You have $480,000 drawn on the line of credit and 1.9 weeks of headroom. The annual cost of reactive maintenance alone is larger than your entire LOC draw. You are not short on profitability — you are short on the cash your profit should have become, and a meaningful share of it is being consumed by machines going down unpredictably and the catch-up scramble that follows.
The structural cause is the trade you're making without choosing it: with two shifts running hard and preventive maintenance skipped "often" to keep machines available, small wear problems are allowed to grow into breakdowns. Because spare parts aren't staged ahead of need, each repair stretches longer than it should — you wait on parts while a revenue machine sits idle. In effect the shop is swapping scheduled 30-minute PM stops for unscheduled multi-hour failures, and every one of those failures pulls a machine off paying work. The data shows it, it doesn't just imply it: 68% reactive, PM skipped often, parts waited on "sometimes," 22 hours down a month.
Why hasn't it been fixed? Not for lack of capability — it's a structural bind common in shops running hot. Your maintenance person works from a radio call, not a schedule, because the moment a hot job comes through, the PM stop is the easiest thing to defer. There's no failure history to manage by, so you can't see which machines are actually eating your uptime — you treat symptoms as they appear. And your controller, who is good, is stretched too thin to build the system that would surface it. These forces keep the problem in place precisely because each individual decision to skip a PM and keep the machine running looks rational in the moment. Addressing the root cause — putting PM on a schedule and the failures on record — produces durable results where symptom-chasing never has, because it changes the trade you're making from "react and recover" to "prevent and plan."
The clearest thing this problem hides is its own size. Judged against the $148,000 maintenance line, reactive maintenance looks like a managed, ordinary cost. But the budget conceals the real number: 264 hours of downtime at your production rate is roughly $534,000 of unshipped product — more than seven times the premium you pay on the repairs themselves. If you've been deciding whether reactive maintenance is hurting you by looking at the maintenance budget, you've been watching the wrong number entirely. That's not an observation you could have made from the reports you have; it only appears when downtime hours are valued at what the floor actually produces.
The second thing it hides wears a different department's label. Your $520,000 of work-in-process reads like a purchasing or buying problem — too much inventory bought. It isn't. That WIP was manufactured and then stranded by downtime and the scheduling churn that follows it; half-finished jobs don't flow, they pile up between operations and freeze as cash on the balance sheet. So part of what looks like an inventory problem, and part of what looks like a collections-driven cash squeeze, is actually this maintenance problem expressing itself in a different ledger. The real cost of reactive maintenance is higher than the visible cost because the same instability shows up three times — as lost machine hours, as frozen WIP, and as the overtime you buy to recover — and only one of those three lands on the maintenance line.
The approach is to stand up a simple computerized maintenance management system — a CMMS — and to pre-stage the critical spare parts that today turn a quick fix into a multi-hour wait. In plain terms: instead of your maintenance person reacting to radio calls, the shop runs maintenance off a schedule and keeps a record of what broke, when, and why. The target is to move from 68% reactive down to 40% reactive. That's not perfection — it's the level where the machines that make your money stop being the things that most often stop your money.
The methodology is deliberately staged so it takes hold rather than becoming shelfware. We start with the five machines that eat the most uptime — not the whole floor — because a small dataset that's actually maintained beats a complete one that nobody keeps current. We build their PM schedules into the two-shift calendar as fixed blocks, scheduled the same way you'd schedule a customer job, so the PM stop stops being the first thing sacrificed when a hot job lands. Each machine's checklist is owned by the operator who runs it most, and completion is reviewed at the shift handoff you already hold — no new meeting. Critical spares are staged only for the failure modes that history shows caused the long waits, so you solve downtime without inflating the $1.085M of inventory you already carry.
Within a few weeks the CMMS starts producing something you've never had: a failure history. That's the hinge of the whole approach. Once you can see which machines fail, how often, and from what cause, you stop treating symptoms and start attacking your worst offenders by name. The system also survives turnover — which matters at Hartwell, where the people most likely to leave are the newer hires, and where institutional knowledge currently lives in a few senior heads rather than on a record.
What makes this different from "we'll do better on maintenance" — the resolution that fades the first busy week — is that it changes the structure, not the intention. PM windows are scheduled commitments, not aspirations. Failures are recorded, not remembered. The discipline is embedded in routines you already run. That's why it produces durable recovery where past good intentions haven't.
| Option A | Option B — Selected | Option C | |
|---|---|---|---|
| Approach | Conservative — PM checklists + staged spares | Balanced — CMMS + critical spares ← | Aggressive — add condition-based sensor monitoring |
| Goal | Start immediately, no capital, prove the approach first | Build a maintenance system that survives turnover and gives data to manage by | Recover maximum production fast; ready to invest in monitoring |
| Reactive target | 68% → 50% | 68% → 40% | 68% → 25% |
| Annual recovery | $168,927 | $309,700 | $450,472 |
| First cash | Within 90 days | 60–90 days | ~120 days |
| Change required | Low | Moderate | Significant |
| Risk level | Low | Moderate | Higher |
Option A is insufficient for Hartwell at this stage for a specific reason: checklists alone give you discipline but no data. You told us you "can't prove it with the reports you have" — and A leaves that exactly where it is. It also doesn't survive turnover, and with 22% production turnover concentrated among newer hires, a system that lives only in checklists and senior memory is fragile. Option C is premature, not wrong: condition-based sensors require capital and pull your maintenance lead off the floor during rollout, and with the LOC at 64%, 1.9 weeks of headroom, and $185,000 of capex already planned for the next twelve months, adding sensor capital now is more disruption and more cash strain than this moment can absorb. Option B is the fit because it builds the failure-history system the controller needs and you've been missing, targets a 40% reactive level your current staff can actually hold, and asks for modest cost and moderate change — the right amount of system for a shop that needs the data more than it needs the sensors, right now.
Six months in, the reactive share of your maintenance has moved from 68% toward 40%, and unplanned downtime has dropped from about 22 hours a month toward the low-teens — call it roughly nine fewer lost machine-hours every month on your highest-value equipment. That recovered capacity is worth on the order of $309,700 a year, and you can see it landing because the CMMS is producing a monthly report that shows downtime falling and tells you which machines improved. The floor feels different: PM happens on schedule instead of by emergency, the catch-up overtime eases, and jobs stop getting stranded mid-process, so WIP starts to flow instead of piling up. And you have something you flatly didn't have before — a record of what your machines actually do, which means the next decision about where to spend a maintenance dollar is made from data, not from the loudest breakdown. The cash that frees up goes where you said you wanted it: toward a line of credit that finally pays down.
| Deliverable | Description | Format |
|---|---|---|
| Machine-Criticality Ranking | Top revenue-critical machines ranked by downtime impact and failure history | Spreadsheet |
| Maintenance Audit Report | Current reactive/planned split, downtime baseline, and failure patterns | |
| Configured CMMS Instance | Live work-order and PM system configured for the top five machines | Hosted software |
| PM Schedule & Checklists | Fixed PM windows plus per-machine, operator-owned checklists | CMMS + printable |
| Critical-Spares Staging Plan | Failure-mode-justified spares list with reorder points | Spreadsheet |
| Monthly Uptime & Failure Dashboard | Recurring report of downtime, reactive %, and worst-performing machines | CMMS report |
| Shift-Handoff Integration Guide | Handoff checklist and work-order logging procedure | |
| Baseline & Recovery Tracker | Start-state metrics and monthly progress against Hartwell's own targets | Spreadsheet |
| Component | Amount |
|---|---|
| Phase 1–2 (Foundation + Implementation) | $TBD — to be confirmed |
| Phase 3–4 (Stabilization + Optimization) | $TBD — to be confirmed |
| Total engagement fee | $TBD |
Investment confirmed prior to commencement. Payment structure to be agreed.
| Conservative | Expected | Upside | |
|---|---|---|---|
| Engagement fee | $TBD | $TBD | $TBD |
| Year 1 cash recovery | $309,700 | $309,700 | $309,700 |
| Net Year 1 return | TBD — fee-dependent | TBD — fee-dependent | TBD — fee-dependent |
| ROI | TBD | TBD | TBD |
| First cash visible | 60–90 days | 60–90 days | Earlier |
Note: All projections use Hartwell's actual operating data. The diagnostic models Option B as a single-point annual recovery of $309,700 — it does not separately model conservative and upside recovery bounds, so the same figure governs all three columns; the engagement fee, net return, and ROI are confirmed prior to commencement (needs verification). The conservative scenario governs all commitments in this engagement. The full monthly cash model belongs in the implementation plan, not this brief.
Every month this engagement is deferred costs Hartwell an estimated $46,924 in production lost to unplanned downtime and reactive-repair premiums. Across the 60–90 day window before the first recovery would otherwise begin to show, that is roughly $140,772 of capacity gone — and on an annualized basis, the cost of not acting is $563,090. That is not a projection. It is what Hartwell's current operating data shows is already happening, month after month, while the line of credit stays pinned at 64%.
The engagement begins upon receipt of signed confirmation and the first installment.
All client financial, operational, and diagnostic data is held in strict confidence and is not shared with third parties. All materials are for the exclusive use of Hartwell Precision Parts.
Any change to scope is documented in a written scope amendment and agreed by both parties before that work proceeds. Verbal agreements are not scope changes.
The client provides timely access to the data, systems, personnel, and facilities the engagement requires. Delays in access may affect the timeline.
All deliverables become the property of Hartwell Precision Parts upon receipt of full payment. Kaufmann Consulting may reference this engagement as experience without identifying the client, unless the client requests otherwise in writing.
Either party may terminate with 14 days' written notice. Work completed to the point of termination is billed at the agreed rate. Deposits are non-refundable.
This engagement is governed by the laws of the State of [State — needs verification].
This brief describes the full scope, approach, deliverables, and expected outcomes of our engagement. Please review it carefully before confirming.
When you are ready to proceed:
Questions before confirming? Contact Bruce directly — no question is too small.