Lesson 3 of 5
Maintenance analytics: backlog, breakdowns and costs
10 min
Listen to this lesson
SAP's training describes maintenance analytics in four layers. Master data (the technical object structure) is the basis for evaluations. Document data from notifications and orders becomes the maintenance history. Embedded analytics in SAP S/4HANA use CDS views and Fiori apps on live data. Strategic analytics, with data warehousing and BI tools, consolidate several systems and usually require additional licenses. Contoso's conversion scope covers the embedded layer.
What SAP offers in the system
| Tool | What it shows | Typical question |
|---|---|---|
| Technical Object Breakdowns (Fiori) | Breakdowns by object, with drilldown to the notification | Which machines break down most, and for how long? |
| Technical Object Damages (Fiori) | Damage codes by object | Which damage keeps recurring on winding machines? |
| Maintenance Order Costs (Fiori) | Costs on maintenance orders | Where is spend going and how does it compare with plan? |
| Maintenance Planning Overview (Fiori) | Unprocessed notifications, unreleased orders, purchasing gaps | What is waiting for the planner today? |
| PMIS: MCI5, MCI7, MCI8 | Damage, breakdown and cost analysis on information structures | Trends by object class, manufacturer or location over months |
| Order and notification lists: IW38/IW39, IW28/IW29 | Open, completed and historical documents with filters | Detailed lists behind any figure |
Embedded maintenance analytics named in SAP's training.
Measures maintenance managers use
SAP supplies the data and the apps; the measures are a business choice. Contoso's Detroit maintenance manager watches five, each defined in writing so that everybody calculates it the same way:
- Backlog: planned hours of released and unreleased open orders divided by weekly work center capacity. Detroit: 640 open hours against 400 hours a week is 1.6 weeks.
- Preventive compliance: preventive orders completed within tolerance of their planned date, as a share of those due.
- Planned share: hours on preventive and planned corrective work against emergency work.
- Breakdown measures: number of breakdowns and mean time to repair per machine, from notification malfunction times.
- Cost against plan: actual against planned costs by value category, and settled cost per line.
Analytics are only as good as the data under them. Breakdown measures depend on operators setting the breakdown indicator and malfunction times; damage analysis depends on catalog codes; cost measures depend on settlement. That is why analytics belong in the test scope: not to test SAP's apps, but to prove that the process produces data the reports can use. Completed notifications and orders feed the history used for medium-term analysis; archived orders become historical orders for the long term.
SAP keeps extending the embedded layer. SAP's training notes that SAP S/4HANA 2025 and its first feature pack add and release CDS views for maintenance, for example for maintenance order cost determination status and maintenance process phases. Check which views and apps your release offers before building custom reports; a standard view that already exists is cheaper to test and maintain than a custom one.
Contoso converted its ECC maintenance reports as they were, and Kenji's team adapted two custom reports that read tables changed by the conversion. Both are compared against the standard apps in testing: if a custom backlog report and the Maintenance Planning Overview disagree about the number of unreleased orders, one of them is wrong, and the test must say which.
Where the numbers come from
Every maintenance measure can be traced to a small number of fields. Knowing them turns "the app looks wrong" into a precise defect. The notification carries the breakdown indicator and the malfunction start and end (date and time), which give breakdown counts and duration; the order carries planned work per operation, actual work from confirmations and the basic start and finish dates; settlement carries the cost per receiver.
| Measure | Source fields | Contoso definition | Test check |
|---|---|---|---|
| Breakdown count | Notification with breakdown indicator set | Notifications per machine per month with the indicator set | Count in the app equals the filtered notification list (IW29) |
| MTTR | Malfunction start and end on the notification | Sum of breakdown durations divided by number of breakdowns | Recalculate three samples by hand |
| Backlog (weeks) | Planned work of open order operations; work center capacity | Open planned hours / weekly available hours | Order list sum equals the numerator used in the report |
| Preventive compliance | Planned date of the call and actual completion date | Completed within ±3 days of plan / due in the month | Check one late and one on-time order |
| Cost against plan | Planned and actual costs by value category; settlement | Actual / planned per order type and line | Settled total equals cost center receipt |
Contoso's maintenance measures and how each is tested.
Beyond the embedded layer, SAP offers asset health and reliability tools on SAP Business Technology Platform, such as SAP Asset Performance Management, which use the same technical objects and maintenance history. Contoso has them on its post-go-live roadmap, not in the conversion scope, but the data habits tested now (breakdown flags, codes, malfunction times) are what those tools will need later.
- Embedded analytics use live data through Fiori apps, CDS views and PMIS; strategic analytics need separate tools.
- Breakdown, damage and cost apps answer the reliability and spend questions; the planning overview shows today's work.
- Define measures such as backlog and preventive compliance in writing before go-live.
- Test analytics by tracing figures back to the notifications, orders and settlements behind them.
Check your understanding · not graded
Lesson check
1. Which Fiori app shows breakdowns by technical object with drilldown to notifications?
2. Which KPI best measures maintenance workload against capacity?
3. Breakdown statistics look far too low after go-live. What is the most likely data cause?
4. Which notification data drives the mean time to repair figure?