
Price gets attention, but a supplier’s after-sales service decides whether your automatic assembly machine keeps running. Our Wenzhou engineers have seen lines sit idle for weeks. Measurable commitments prevent that.
Evaluate a supplier’s after-sales service by testing one question: after a breakdown, can they respond fast, diagnose accurately, deliver a fix, and restore production within a defined time? Verify warranty terms, response commitments, spare parts availability, training, and documented customer references before you sign anything.
That one question is the whole article in miniature. Below, I break it into four practical checks. Each one turns a sales promise into something you can write into a contract and hold a supplier to.
A procurement manager in Germany once asked me what one-year warranty actually covered in our quote. That question deserves a longer answer than most suppliers give.
Before signing, check when warranty coverage begins, whether parts, labor, and travel are all included, which wear parts and third-party components are excluded, how software is covered, and what response and repair obligations the supplier accepts in writing. Duration alone tells you very little.
Most quotes say twelve months. The real question is twelve months from when. If coverage begins at shipment, sea freight to the United States or Mexico, customs clearance 1, and installation can eat a large slice of that period before the machine makes its first good part. We prefer to tie the clock to a signed acceptance report, and I encourage buyers to ask every supplier for the same.
Acceptance itself needs a definition. Factory acceptance testing 2 at the supplier's plant proves the machine works with sample parts. Site acceptance proves it works with your parts, your utilities, and your operators. The contract should list measurable targets: cycle time, output rate, first-pass yield, changeover time, and the range of product variants the machine must handle. Do not release final payment because the machine arrived and powered on. Release it when the commissioning report and punch list are closed.
Here is how I separate weak warranty language from strong language. The difference is usually hidden in what is not mentioned.
| Coverage item | Weak contract wording | Strong contract wording |
|---|---|---|
| Parts and labor | Parts only | Parts, labor, and technician time |
| Travel and accommodation | Excluded or silent | Included for production-stopping faults |
| Software and controls | Not mentioned | PLC, HMI, and robot programs covered, backups delivered |
| Wear parts | Vague "consumables excluded" | Listed by part number with expected life |
| Third-party components | "Contact the component OEM" | Supplier handles the claim on your behalf |
| Replacement parts | No renewed coverage | Renewed warranty on each replaced item |
| Response obligations | None | Written Service Level Agreement with time targets |
Ask this exact question: if the machine stops because of a failed servo drive, sensor, PLC module, or robot controller, what will you pay for, and within how many days? A supplier who answers with specifics has thought about it. A supplier who answers with "we will support you" has not.
A warranty without remedies is a brochure. Ask what happens when a target is missed. Service credits, extended coverage, or escalation to a named manager all work. Also check the fine print on force majeure and geographic conditions. A response target that only applies "within the supplier's home country" is not useful to a buyer in Italy or South Korea.
The hardest lesson our team learned supporting machines in India was that answering a WhatsApp message in ten minutes is not the same as restoring output.
Expect a defined technical support response time, not a vague 24/7 promise. A credible supplier commits in writing to acknowledge a call, start remote diagnosis within hours, ship critical parts, and dispatch a technician within a stated window. Measure resolution time separately from first response.
A supplier can pick up the phone fast and still leave your line down for a week. I ask buyers to split the recovery chain into stages and demand a target for each one: acknowledging the request, beginning remote diagnosis, issuing a corrective-action plan, dispatching a technician, arriving on site, supplying the critical part, restoring production, and completing the permanent fix. The sum of those stages is your real Mean Time to Repair 3. That number drives machine downtime costs, not the speed of the first reply.
The numbers below are a starting template, not an industry average. Adjust them to your shift pattern and your distance from the supplier.
| Severity | Definition | Acknowledge | Remote support starts | On-site arrival | Target restore |
|---|---|---|---|---|---|
| Critical | Line stopped, no workaround | Within 1 hour | Within 2 hours | Within 24–72 hours depending on region | Agreed per case, parts permitting |
| Major | Reduced capacity or quality | Within 4 hours | Within 8 hours | Within 5 working days | Within 10 working days |
| Minor | No production impact | Next business day | Within 2 business days | Next scheduled visit | Next scheduled visit |
Ask whether support covers nights, weekends, and holidays in your time zone. Ask whether you reach an engineer or a call queue. Ask whether support is available in your language. Then ask for the escalation path to the original machine designer.
Buyers often tell me a local integrator must be faster than a team in China. Sometimes that is true. Often it is not. A large local integrator may route your call through a service desk and schedule a visit for next week. A smaller overseas builder who designed your machine can log in, read the alarm history, and point to the failed subsystem in one session. Our team is small, so we design our PLC-controlled lines for secure remote access from day one. The right comparison is measurable coverage: on-site technician availability, actual arrival commitments, and parts logistics, not the supplier's address.
Remote access cuts travel cost and shortens diagnosis, but it opens a door into your factory network. Write the controls into the contract: customer approval for every session, named user accounts, time-limited access, session logs, a secure VPN 4 or equivalent, and a requirement to return the machine to a documented baseline afterward. Video and augmented-reality guidance help too, but only when the machine design lets the remote engineer see real diagnostic data instead of guessing.
We weigh one trade-off on every automation line we ship: which spares go in the crate now, and which ones the customer orders later.
Ask upfront for a recommended spare parts list with criticality ratings, unit prices, and lead times, plus local stock levels, obsolescence notices, and a preventive maintenance schedule based on cycles or hours. Confirm who stocks critical items after warranty and whether components are proprietary or industry-standard.
Buyers push back on spares lists for a fair reason: stocking every possible component ties up cash and creates obsolete inventory. I agree. The answer is not "stock everything" or "stock nothing". It is a matrix.
| Category | Failure effect | Lead time | Recommended action |
|---|---|---|---|
| A | Stops production | Long or uncertain | Stock on site before start-up |
| B | Reduces capacity, line keeps running | Medium | Stock regionally or with the supplier |
| C | Common, low-cost, fast delivery | Short | Order as needed |
| D | Obsolete or fully custom | Unknown | Negotiate redesign or lifecycle support |
Typical Category A items on a bowl-feeder assembly cell include the PLC and I/O modules 5, servo drives, encoders, safety relays, the HMI or industrial PC, and custom grippers or tooling. A relatively cheap drive sitting on a shelf is far cheaper than several lost shifts.
Walk the machine during factory acceptance testing and read the labels. Sensors, pneumatic valves 6, drives, and controllers from recognized industrial brands can be sourced locally in the United States, Japan, or France if the supplier is slow. Proprietary parts cannot. Our lines use widely available PLC platforms for exactly this reason, and I tell buyers to make openness a line item in their evaluation.
A supplier who only reacts to breakdowns is not protecting your uptime. Here is the sequence I expect a serious supplier to document:
Preventive maintenance runs on a schedule. Predictive maintenance reads condition data such as motor current, temperature, cycle counts, and alarm frequency to catch failures early. Ask which one the supplier offers and what data the machine exposes.
Operator training programs and maintenance training decide whether you stay dependent on the supplier forever. Ask how many training days are included, and ask for a second session after a few weeks of real production, when operators have seen real faults. Then confirm what you receive at handover: electrical schematics, pneumatic diagrams, parts lists, parameter backups, error-code explanations, and maintenance schedules. Negotiate access to PLC source code and HMI project files, or at least software escrow. A PDF manual plus a password-locked controller is not a complete handover.
One more objection I hear: service contracts are a waste for standardized machines. For a simple cell with a strong in-house maintenance team, that is correct. A parts program, a training package, and pay-as-needed engineering time cost less. For a highly customized line running around the clock, a guaranteed-response contract with lifecycle support services usually pays for itself after one avoided stoppage.
During a final inspection on a terminal crimping line, the buyer's engineer asked to call a customer we had supported for years. I handed over the number gladly.
Verify after-sales reputation by requesting references from customers with similar machines, volumes, and locations, including one installation several years old. Ask them specifically about breakdown response, spare parts availability, warranty disputes, surprise invoices, and whether they would reorder. Then check public reviews and video channels.
Most reference calls drift into how friendly the sales team was. Steer them back. These are the questions I would ask any supplier's past customer, including ours:
The age of the installation matters most. Launch support is easy to deliver. Support in year four, after the warranty has expired, is where reputations are made or lost. Ask for at least one reference several years old, and ask for anonymized examples of recent service cases in your own region.
Reputation is only useful when you can compare suppliers side by side. I suggest scoring each category from 0 to 5: 0 means no evidence, 1 means a verbal promise, 2 means documented but weak or geographically limited, 3 means a clear process with references, 4 means contractual service levels and proven local capability, and 5 means measurable performance data plus strong reference evidence.
| Evaluation category | Suggested weight | What counts as evidence |
|---|---|---|
| Response and escalation | 20% | Written SLA, named escalation contacts |
| Spare parts availability | 15% | Priced list with lead times, stock locations |
| Commissioning and acceptance support | 15% | Acceptance criteria, sample commissioning report |
| Warranty coverage terms | 15% | Full contract text, remedies for missed targets |
| Local or regional engineering coverage | 10% | Engineer locations, partner names, travel terms |
| Preventive maintenance capability | 10% | Sample checklist and report |
| Training and documentation | 10% | Training plan, document list, software access terms |
| Remote diagnostics and cybersecurity | 5% | Access procedure, audit trail, session approval |
Shift the weights to match your risk. A 24-hour operation should push more weight toward response and parts. A low-volume plant can favor training and flexible pay-as-needed support.
Walk away, or at least slow down, when a supplier refuses service references, routes support only through sales staff, starts the warranty at shipment, excludes travel and labor from repairs, cannot quote lead times on critical parts, cannot explain software backup and recovery, or discourages you from training your own maintenance team. Any one of these is a warning. Several together are a decision.
Finally, run the service terms through a simple model. Downtime cost per hour equals lost contribution margin plus idle labor plus recovery cost plus scrap or rework plus any delivery penalties. Expected annual service cost equals planned maintenance plus parts plus travel and labor plus expected downtime cost. Test four scenarios against each supplier: a minor fault, a major controls failure, a long-lead custom part, and a repeated intermittent fault. The cheapest service package rarely stays cheapest once the second scenario is on the table.
Cheap quotes hide expensive downtime. Choose the supplier who can respond, diagnose, fix, and restore production on a written timeline. Email [email protected] if you want our checklist.
1. WTO trade facilitation rules explain delays affecting warranty start timing. ↩︎
2. Defines the standard pre-shipment verification step referenced in warranty discussion. ↩︎
3. NIST reliability engineering handbook defines this downtime metric precisely. ↩︎
4. NIST framework guides secure remote access controls for industrial equipment. ↩︎
5. Explains the core control hardware listed as critical spare parts. ↩︎
6. ISO develops component interoperability standards relevant to sourcing industrial parts. ↩︎