The ERP runs in the cloud, the MES servers are hosted in a data center, and network and server support comes from a service company. At a medium-sized automotive supplier, this picture is normal. When one of these services stops, shipping, label printing or traceability records stop too. From an IATF 16949 perspective, this means IT suppliers aren’t just a name on purchasing’s list; they must be controlled as outsourced processes that affect product conformity and uninterrupted supply. Below is a framework that works in the field, from classification to selection criteria and from SLA indicators to an exit plan.
IT services as outsourced processes
Section 8.4 (Control of externally provided processes, products and services) of IATF 16949 requires outsourced processes not to adversely affect the organization’s ability to deliver conforming product to the customer. Supplementary clause 8.4.1.1 extends the scope to all products and services that affect customer requirements; 8.4.1.2 covers the supplier selection process, 8.4.2.1 the type and extent of control, and 8.4.2.4 supplier monitoring.
The standard doesn’t list IT services separately in these clauses. The deciding question is: If this service is interrupted or malfunctions, are product conformity, delivery or obligations to the customer affected? If the answer is yes, the service is controlled under 8.4. MES hosting directly affects traceability records, the EDI service customer call-offs and shipping notices, and cloud ERP shipping documents. Maintenance of office printers usually doesn’t fall within this scope. But label printers do.
This view differs from assessing the same suppliers for information security risk. The information security side is explained in the guide supplier information security risk assessment; the focus here is supplier control from a quality management system perspective. The two assessments can and should be combined in the same supplier record; otherwise there are two separate files and two separate results for the same company.
Classifying suppliers by their effect on production
Applying the same intensity of control to every IT supplier both creates unnecessary workload and causes critical ones to be overlooked. A three-tier classification is enough for most organizations:
| Class | Definition | Examples | Control intensity |
|---|---|---|---|
| A – Direct effect on production | An outage stops production or shipping in a short time | MES hosting, cloud ERP, EDI, label system support | Written SLA, monthly monitoring, annual assessment, exit plan |
| B – Indirect effect | An outage makes production harder but it can continue with a temporary method | Network and server support company, backup service | SLA, quarterly monitoring, annual assessment |
| C – No effect | No effect on production and delivery | Corporate website hosting, office hardware maintenance | Standard purchasing control |
The classification must be consistent with the business impact analysis and IT contingency plans. When requiring contingency plans, IATF 16949 clause 6.1.2.3 also lists interruptions in externally provided products, processes and services among the risk sources. If a class A IT service isn’t in the contingency plan, the auditor writes it up as an inconsistency between the two clauses. This is a situation you may come across: the supplier list in one place, the contingency plan in another, unaware of each other. For continuity assurance from critical suppliers, the guide business continuity at critical suppliers is complementary.
Selection and approval criteria
IATF 16949 clause 8.4.1.2 requires a documented supplier selection process and lists the elements to consider: the supplier’s risk to product conformity and uninterrupted supply, quality and delivery performance, assessment of the quality management system, multidisciplinary decision-making and, where applicable, software development capabilities. The clause also mentions other criteria that may be considered, such as financial stability, the change management process, business continuity planning and customer service.
A selection checklist adapted to IT services and IATF 16949 IT requirements:
- Service continuity: Redundancy of the hosting infrastructure, disaster recovery capability and evidence that it has been tested
- Change management: Whether platform updates are notified in advance, and whether maintenance windows can be adjusted to the production calendar
- Support model: Support hours suited to shift-based production, priority definitions and an escalation path
- Data ownership and access: Who owns the data, and in what format and how quickly it can be exported
- Logs and audit trail: Whether the service can keep logs and records in a way that meets IATF record retention periods
- Information security assurance: Independent assurances such as an ISO/IEC 27001 certificate, a SOC 2 report or a TISAX label, and whether their scope really includes the service
- Financial and organizational stability: The supplier’s sustainability for services that create long-term dependency
Pay particular attention to the support model item. For a site running three shifts, a support company that answers the phone during office hours isn’t there when the label server goes down at 02:00. If this gap isn’t written into the contract, you find out on the first night-time outage.
The selection decision should be the result not of a single department but of a multidisciplinary assessment involving IT, quality, production and purchasing, and should be kept as a signed form.
SLA and performance monitoring indicators
Clause 8.4.2.4 requires supplier performance to be monitored with defined indicators. For production suppliers, the standard lists indicators such as delivered product conformity, disruptions at the customer’s plant and delivery performance. For IT suppliers, the equivalent is service indicators that measure the effect on production:
| Indicator | Measurement method | Why it matters |
|---|---|---|
| Availability during production hours | From the monitoring system, according to the shift calendar | Night-time maintenance isn’t the same as a mid-shift outage |
| Critical incident response and resolution time | From support records, by priority | Directly determines line downtime |
| Number of unplanned changes | Updates made without notification | An unverified change is a production risk |
| Service-related production stoppage | Reason code in production downtime records | The real measure of quality impact |
| Restore test result | Periodic test report | Evidence that the backup works |
| Was reporting done on time | Delivery date of the monthly service report | Continuity of control |
The most common mistake is measuring availability in calendar time. A four-hour planned maintenance at midnight and a half-hour outage at the start of the morning shift don’t carry the same weight for production. The report is green, the line is red. That’s why the definition of “availability” in the contract must be tied to the production calendar, and service-related stoppages must also be tracked from production’s own downtime records. The difference between the supplier’s report and your own downtime records should be the first agenda item in the monthly review.
Target values are written into the contract; when a target isn’t met, a corrective action is requested, similar to the supplier development approach in clause 8.4.2.5, and its closure is tracked.
To decide in which cases supplier-driven updates require revalidation or customer notification, you can use the classification in the guide change management in production systems.
Supplier change and exit plan
Changing supplier for a class A IT service is as serious as moving a production process to another plant. The exit plan is prepared while the relationship is good. Once the relationship breaks down, negotiating over data and support becomes much harder. Minimum content:
- In what format, in what scope (including historical records, attachments and audit trails) and how quickly data will be handed over
- The current supplier’s support obligation and its duration during the transition period
- How quality records whose retention period is still running will stay readable in the new system or archive
- Confirmation that the data on the supplier’s side has been deleted
- Assessing whether the change requires customer notification
The last item is easy to miss. Some customers require, in their customer-specific requirements, notification of significant changes to systems where production records are kept or to EDI infrastructure. Reviewing the exit plan once a year and actually trying a data export at least once shows the plan hasn’t stayed on paper. Open the exported file and check whether it includes attachments and the audit trail; that’s the part most often missing in the field.
Records to show in the audit
- The approved supplier list including IT suppliers, with the justification for their classification
- Selection/assessment forms for class A and B suppliers
- Contract and SLA annexes; indicator definitions and targets
- Monthly or quarterly performance reports and review minutes
- Corrective actions opened when targets weren’t met, and evidence of their closure
- Contingency plan sections covering IT suppliers
- The exit plan and a record of the data export test
How to combine these records with the general IATF IT evidence set is explained in the guide IT evidence in the IATF 16949 audit. If you need to manage IT suppliers with a single assessment model from both a quality and an information security perspective, we combine these two views in our supplier and third-party risk management work. The primary source for the standard’s official interpretations and FAQs is the IATF Global Oversight website.
Frequently asked questions
Does a cloud ERP provider count as a supplier under IATF 16949?
Yes, if the service affects product conformity, delivery or obligations to the customer. If shipping documents, material movements and traceability data are kept on the cloud ERP, the provider is controlled as an outsourced process under clause 8.4.
Is an IT supplier’s ISO 27001 certificate enough for supplier assessment?
Not on its own. The certificate shows information security management; it doesn’t cover topics decisive for quality, such as service continuity, the support model, change notification and data export. Also check whether the certificate’s scope really includes the service you receive.
How should the availability target be defined in the SLA?
Availability should be tied to the production calendar rather than calendar time, planned maintenance windows should be defined separately, and service-related production stoppages should also be tracked from the organization’s own downtime records. The target value is chosen according to the tolerance time set in the business impact analysis.
Is a selection form needed for a small IT support company too?
If the company provides a service with a direct or indirect effect on production, a selection and assessment record is needed whatever its size. The intensity of control can be adjusted by class; for class C suppliers, standard purchasing control may be enough.