# Predictive Maintenance Data Readiness Assessment > Source: https://netray.co/tools/predictive-maintenance-data-readiness-assessment > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Free Tools ## Short answer Ten questions that show whether your asset data, failure history and sensor coverage can support predictive maintenance, and what to fix first. ## Category maintenance ## Industry Discrete Manufacturing ## Audience - Reliability engineer - Maintenance manager - CIO - Plant manager ## Tool **Kind:** assessment **Questions:** #### How complete and accurate is your asset register? **Help:** Unique tags, hierarchy, make and model, criticality **Options:** ##### Partial, many duplicates or missing assets **Score:** 0 ##### Most assets listed, hierarchy is patchy **Score:** 1 ##### Complete with a clear hierarchy **Score:** 2 ##### Complete, hierarchy and criticality maintained and audited **Score:** 3 #### How are failures recorded in work orders? **Options:** ##### Free text only **Score:** 0 ##### Codes exist but are rarely used correctly **Score:** 1 ##### Problem, cause and remedy codes used on most orders **Score:** 2 ##### Structured codes used and reviewed for quality every month **Score:** 3 #### How much failure history do you have in usable form? **Options:** ##### Less than 6 months **Score:** 0 ##### 6 to 12 months **Score:** 1 ##### 1 to 3 years **Score:** 2 ##### More than 3 years with consistent coding **Score:** 3 #### What sensor or inspection data do you capture on critical assets? **Options:** ##### None, we rely on operator rounds **Score:** 0 ##### Some manual readings on paper **Score:** 1 ##### Route-based vibration, oil or thermography stored electronically **Score:** 2 ##### Route data plus online sensors on critical assets **Score:** 3 #### Do you link sensor readings to asset records and work orders? **Options:** ##### No link at all **Score:** 0 ##### Manual cross reference in spreadsheets **Score:** 1 ##### Readings stored against the asset tag **Score:** 2 ##### Readings, alarms and work orders linked through the asset tag **Score:** 3 #### Are baselines and alarm limits set for monitored points? **Options:** ##### No baselines **Score:** 0 ##### Generic limits for all machines **Score:** 1 ##### Limits per machine class with some baselines **Score:** 2 ##### Per-point baselines and limits reviewed annually **Score:** 3 #### Do you know the P-F interval for your key failure modes? **Options:** ##### No **Score:** 0 ##### Rough guess for a few **Score:** 1 ##### Estimated for most critical failure modes **Score:** 2 ##### Estimated and checked against real catches and misses **Score:** 3 #### Who owns turning a finding into a work request? **Options:** ##### Nobody in particular **Score:** 0 ##### The analyst emails someone **Score:** 1 ##### A named planner, but no tracking **Score:** 2 ##### A named planner with a tracked list and monthly review **Score:** 3 #### How reliable are your timestamps, runtime and operating context data? **Help:** Running hours, load, product, shift, speed **Options:** ##### Not recorded **Score:** 0 ##### Recorded by hand and inconsistent **Score:** 1 ##### Recorded from the control system for some assets **Score:** 2 ##### Recorded automatically and available alongside sensor data **Score:** 3 #### Can you measure the result of the program? **Options:** ##### No metrics **Score:** 0 ##### Downtime only **Score:** 1 ##### Downtime, backlog and planned work percent **Score:** 2 ##### Also avoided failures, false alarms and savings tracked **Score:** 3 **Bands:** #### Foundations needed **Min:** 0 **Max:** 34 **Verdict:** Your data and processes are not yet ready for predictive models. Time spent on the basics will pay back faster than buying sensors. **Recommendations:** - Clean up the asset register and set criticality for every asset - Introduce simple failure codes and enforce them on work orders - Start a manual route with a baseline on the top 10 critical assets - Name an owner for turning findings into work requests #### Building **Min:** 35 **Max:** 59 **Verdict:** You have pieces in place, but gaps in data linking and ownership would limit what predictive analytics can do. **Recommendations:** - Link readings to asset tags in one system - Set per-point baselines and alarm limits - Estimate P-F intervals for the top failure modes - Review failure code quality monthly #### Ready for a pilot **Min:** 60 **Max:** 79 **Verdict:** You can run a focused pilot on a few critical assets and measure the results. **Recommendations:** - Pick two or three assets with a known failure mode and enough history - Define success: catches, false alarms, avoided downtime - Keep models simple at first and compare with the analyst call - Connect alarms to work requests with human approval #### Scale up **Min:** 80 **Max:** 100 **Verdict:** Your data and process maturity support wider rollout. Focus on tuning, cost control and governance. **Recommendations:** - Extend to the next tier of assets by criticality - Track false alarm rate and tune limits - Feed repair findings back to improve diagnosis - Report avoided failures and program cost every quarter ## Content **Intro:** Predictive maintenance works only on top of reliable asset data, failure history and a process for acting on findings. This ten-question assessment shows how ready you are and what to fix first. **Sections:** #### How to use it **Body:** Answer each question based on what you can show, not on what is planned. Ask the planner and a technician to check your answers, since asset registers and failure codes are often better on paper than in practice. The score is converted to a percentage and matched to a band with a verdict and recommended next steps. Export the result to share with your maintenance manager or IT team. #### What the questions measure **Body:** Questions cover four areas: asset data (register, hierarchy, criticality), history (failure codes and years of records), sensing (route and online data, linkage to the asset), and process (limits, ownership, metrics). Weakness in any one area limits the others. For example, perfect sensors with free-text failure records cannot tell you what a signal predicted, and good analytics with no owner will produce alerts nobody acts on. #### Reading the result **Body:** A low score is not a reason to delay all work. It tells you where the cheapest improvements are, often code discipline and asset tagging. A middle score fits a pilot on a narrow scope. A high score means the main risk is not data but adoption, so invest in review meetings and measurement. Re-take the assessment every six months to see progress. #### Common mistakes **Body:** Buying a platform before fixing the asset register is the most common. Starting with too many assets, expecting models to learn from data with no coded failures, and ignoring operating context such as load and speed also lead to disappointing results. **Cta:** If the gaps are in how your ERP or CMMS holds asset and failure data, we can review your setup and suggest practical fixes. Ask for an expert review. **Faq:** #### How much failure history do I need for predictive models? **Answer:** Machine learning models need many examples of each failure mode, which most plants do not have. Simpler approaches such as limits, trends and rules need far less data and often deliver most of the value for the first year. #### Is predictive maintenance only for large plants? **Answer:** No. Small plants often benefit from basic condition monitoring on a few critical assets. The key is a clear failure mode with warning time and a process to act on it, not the size of the site. #### Do I need a CMMS first? **Answer:** You need a reliable way to record assets and work, whether in a CMMS or an ERP maintenance module. Without it, there is no place to attach findings or to measure results. #### How often should I re-run this assessment? **Answer:** Every six to twelve months, or after major changes such as a new CMMS, a new sensor program or a reorganisation of the maintenance team. Compare scores to see where progress has stalled. ## Search terms predictive maintenance readiness assessment, is my plant ready for predictive maintenance, predictive maintenance data quality, cmms data quality assessment, failure code quality maintenance, predictive maintenance maturity, maintenance data readiness, predictive maintenance prerequisites --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json