The service contract is where a distributor makes its margin and where it quietly loses its reputation. A 3D printer that stops mid-build costs the customer a print farm a day of throughput and costs a service bureau a customer. The distributor who cannot put a part and a technician in front of that machine inside seventy-two hours discovers that the contract does not renew, that the renewal conversation happens at a competitor, and that the twenty machines behind that customer go with it. This guide is about the physical network that makes the seventy-two-hour promise keepable: where parts sit, how deep the stock goes, and how many technicians are needed to cover a territory.
The economics are unforgiving because the failure mode is not gradual. An installed base below roughly 150 machines in a territory cannot sustain a parts depot on its own; above 400, a single central depot becomes a bottleneck because too many parts are too far from too many machines. The coverage model has to change with the base, and most distributors set one up at 100 machines and never revisit it. Our critical spares kit design guide covers which parts belong in the kit. This article covers where those parts live and how fast they reach the machine.
Why Uptime Contract Value Collapses Without Parts Coverage
Customer tolerance for downtime is not linear in time. It collapses at specific thresholds, and the thresholds are predictable enough to design around.
The reason the seventy-two-hour line matters is that it is the last window where the customer still accepts the incident as an exception. Past a week, the conversation is no longer about the machine — it is about whether the distributor is competent. And because a 3D printer is not a single machine but part of a farm, the loss compounds: a print farm with twenty machines and one bad service experience will not replace twenty machines with the same brand.
The financial expression is that service revenue is small and the installed-base churn it prevents is large. A distributor with 300 machines in territory and an average selling price of $1,800 carries $540,000 of installed base. Even a 5% churn reduction is $27,000 of retained revenue per replacement cycle — against which a parts depot costing $15,000–40,000 a year in stock and space is easily justified, provided the coverage model is right-sized. Our spare parts aftermarket revenue guide covers the revenue side of the same investment.
The Four Coverage Models and Their Break-Even Volumes
There are four workable ways to put parts near machines. Each has an installed-base range where it is the cheapest correct answer.
Single central depot. One warehouse holding $40,000–120,000 of parts at cost, serving a territory with a 24–96 hour delivery window. This is the correct starting model and the one most distributors should keep until roughly 250 machines. Below 80 machines the stock cost per machine exceeds the service revenue per machine, and the honest answer is to run critical spares only and accept longer waits on everything else.
Depot plus forward stocking locations. The central depot keeps depth; a set of small forward locations — a technician's van, a partner reseller's back room, a customer's own store under consignment — keep the top forty fast-moving parts within a 24-hour ring. This is the model that earns the seventy-two-hour promise at scale. Forward locations cost $3,000–8,000 each in stock and are usually justified at 250 machines per location cluster.
Third-party parts logistics. A logistics provider holds your stock in a multi-client facility and ships same-day on your instruction. Cost is 18–25% of part value in handling and storage, which is expensive per line but avoids depot capex entirely. The right answer for a distributor entering a new territory before the installed base justifies a depot.
OEM drop-ship. Parts ship from the factory on demand. Cheap, slow, and the reason distributors lose contracts. Use it only for low-value, low-urgency items and never for the parts that stop a machine.
Fill Rate Math: What Stock Depth Actually Buys
Fill rate is the percentage of part requests satisfied from stock on the first attempt. It is the single number that predicts whether the seventy-two-hour promise holds, and it behaves in a way most distributors underestimate: the last few points of fill rate are extraordinarily expensive, and the first eighty are cheap.
The reason is failure distribution. Spare parts demand is not uniform. In a typical 3D printer installed base, the top ten part numbers — nozzles, hotend assemblies, thermistors, heater cartridges, PTFE tubing, build plates, extruder gears, belts, fans, and the mainboard — account for roughly 55–65% of all part requests. The next forty account for another 25–30%. The remaining several hundred account for under 10% between them.
The practical rule: stock A and B universally, stock C only for models with more than twenty units in the territory, and never stock D. The difference between 88% and 95% fill rate is roughly one urgent air-freight shipment per month per fifty technicians, and air freight on a $40 part routinely costs $60–120. At that rate, holding the C-class stock pays for itself within two quarters.
Classify by measured failure frequency, not by intuition. Pull twelve months of part-request data, count requests per part number, and sort. If you do not have twelve months, use the OEM's field failure data and revise quarterly for the first year. Our annual service contract pricing guide shows how fill-rate commitments get priced into the contract.
Forward Stocking Locations: Choosing Where Parts Live
A forward stocking location is a small cache of fast-moving parts placed close to where machines are. It is not a warehouse and should not be managed like one. The decision is entirely about drive time.
Draw isochrones, not radii. A 200 km radius that includes a mountain range and a border crossing is not a 200 km service area. The useful rings are: a 1-hour drive ring for the A-class parts that stop a machine immediately, a 4-hour drive ring for the B-class, and a next-day courier ring for the C-class. A territory where 70% of machines sit inside one 2-hour ring justifies a single forward location serving all of them, even if the geographic centre of the territory is elsewhere.
Four placement options, in order of cost. A technician van stock at $2,500–5,000 per van is the cheapest and fastest, and covers the machines that technician actually visits. A partner reseller back room under consignment puts stock at a second company's premises and costs the carrying cost of the stock only, but requires a written consignment agreement and periodic reconciliation. A customer-site consignment at a large print farm or manufacturer is the highest-value placement per dollar, because it converts a seventy-two-hour promise into a same-day one for the account that spends most. A third-party locker or 3PL micro-hub costs $300–800 per month and gives courier-speed access without staff.
Whatever the placement, the stock has to be visible to the central system or it becomes invisible inventory that gets re-ordered. Every forward location needs a defined par level, a replenishment trigger, and a monthly reconciliation. Consignment without reconciliation turns into shrinkage: expect 2–5% annual loss on consigned stock and price it in.
Technician Coverage: Mobile, Resident or Hybrid
Coverage is a people problem before it is a parts problem. The binding constraint is travel share: the proportion of a technician's paid hours spent driving rather than fixing.
A mobile technician in a dense urban territory with clustered customers spends 25–35% of the day driving and completes 3.5–5 service calls. In a dispersed rural territory with the same call volume, travel share rises to 50–65% and calls per day fall to 1.5–2.5. The cost per call therefore roughly doubles without any change in wage rate. This is the arithmetic that makes the resident model worth considering.
The hybrid model is the one most distributors settle on between 250 and 800 machines: mobile technicians covering the dense cluster, plus scheduled preventive maintenance visits to the periphery batched geographically, plus a local contractor on retainer for the far edge. The contractor is paid per call at a higher rate but carries no fixed cost, which is correct for a territory where a full-time technician would sit idle half the week.
Resident placement makes sense only when a single customer site holds more than roughly forty machines. Below that, the resident technician is underutilised and either the customer pays for idle capacity or you do. Our technician certification and upskilling guide covers the qualification path that makes any of these models work.
First-Time-Fix Rate: The Metric That Decides Renewals
First-time-fix rate is the share of service calls resolved on the first visit without a return trip or a waiting period. It is the metric customers actually experience, and it is the best single predictor of contract renewal.
Target bands by machine class: entry-level consumer-adjacent machines 85–92%, prosumer and small-farm machines 90–95%, industrial and production machines 93–97%. A distributor below the low end of its band has a coverage problem regardless of what the SLA says.
Three causes account for most first-time-fix failures, and only one of them is a parts problem. No part on the van — the technician arrives, diagnoses correctly, and cannot fix because the part is at the depot. This is a stock placement problem and is fixed by van stock depth. Wrong diagnosis — the technician replaces the hotend when the real fault is the extruder gear or the thermistor. This is a training and remote-diagnostics problem, and fixing it is cheaper than fixing stock. Missing tool or fixture — no heat-set insert tool, no nozzle torque wrench, no calibration shim. This is a van-kit standardisation problem and costs almost nothing to solve.
Measure first-time-fix per technician and per machine model, not just as a company average. Averages hide the pattern: a single model with a systematic fault will drag the rate down while every technician looks equally mediocre. Our service level agreement guide covers how to write the commitment and the exclusion list that keeps it honest.
Consignment and Vendor-Managed Parts at Customer Sites
Consignment is the highest-leverage tool available for a large account and the easiest to lose money on. The mechanic: you place a defined parts cache at the customer's site, the customer draws what they need, and you invoice monthly for what was consumed.
Three terms decide whether it works. The par level — a written list of part numbers and quantities that must be maintained, reviewed quarterly against actual consumption. The shrinkage clause — a stated acceptable loss (2–5% annually is normal) with the customer liable above it, and a physical count cadence. The ownership and exit clause — stock remains your property until consumed, and on termination the customer either buys the balance at cost or returns it within thirty days.
Price consignment stock into the service contract rather than absorbing it. A $12,000 consigned cache at a customer site carries a real carrying cost of roughly $1,200–1,800 a year in capital, insurance and obsolescence, and that belongs in the annual fee.
Depot Economics: Stock, Space, Staff and Carrying Cost
A depot is a small business with its own P&L, and it should be reviewed as one. Four cost lines dominate.
Stock. Budget $35,000–70,000 at cost for an A/B/C mix at 250 machines, rising with installed base. Our multi-site parts logistics guide covers the customs and freight treatment of that stock.
Space. 60–150 square metres of secure, dry, temperature-controlled storage. Parts degrade: belts and elastomers have shelf lives, mainboards are static-sensitive, and resin-adjacent components dislike heat. Expect $8,000–25,000 a year depending on market.
Staff. One parts coordinator per 250–400 machines, at 0.6–1.0 full-time equivalent including picking, packing, receiving and returns processing.
Carrying cost. The line most often left out. Capital tied in parts, insurance, obsolescence on parts for discontinued models, and the write-down when a model generation turns over. Budget 18–25% of average stock value annually. On $50,000 of stock that is $9,000–12,500 a year before a single part moves.
The test of a healthy depot is parts gross margin against total depot cost. A well-run depot at 250 machines turns stock 2.5–4x a year at 35–50% gross margin, which covers its full cost with 15–30% contribution left over. A depot turning 1.5x is a warehouse full of parts for machines the customer already replaced.
Bottom Line
Design the coverage model against the installed base you will have in two years, not the one you have today, and revisit it every time the base grows by 150 machines. Stock A and B parts universally and C parts per model; place forward stock by drive time rather than geography; and measure first-time-fix per technician and per model, because the company average will hide the failure that costs you the renewal.
The seventy-two-hour promise is not a marketing statement. It is the output of a stock decision made eighteen months earlier. Our depot repair and swap pool guide covers the complementary model for whole-unit failures, and our OEE and availability guide explains how customers measure the uptime you are contracted to deliver.
Reviewed by the Precise3D engineering & OEM team. Compliance files that accompany the catalog are auditable at the certification register.
Start Your Partnership
Ready to Build a Service Network That Renews?
Parts kits, depot stocking plans and technician training programmes are available to partners from the first container. Design the coverage model before the installed base outgrows it.