Checklist

Predictive maintenance pilot checklist & success criteria

An interactive checklist to scope, run, evaluate, and record decisions for a focused sensor-driven predictive maintenance pilot. Includes fields for baseline metrics, candidate assets, sensor and signal requirements, data quality ownership, objective success metrics, safety/governance checks, and a clear rollout trigger.

Interactive Tool

Predictive maintenance pilot checklist & success criteria

This interactive checklist helps teams scope a focused, low-risk predictive maintenance (PdM) pilot that proves value. Fill the fields with cross-functional input from reliability, maintenance, controls/automation, IT, and safety before buying sensors or building models. Save the form to record commitments, owners, and acceptance criteria that will determine whether you scale the pilot.

Describe the measurable business outcome this pilot must demonstrate (for example: reduce unplanned downtime for Asset X by N hours per month, or reduce emergency repairs by Y%). Link to cost, safety, quality, or throughput impact.
Number of days of historical data used to calculate baseline metrics.
Total unplanned downtime for the baseline period for the selected assets.
Mean time to repair for the target asset(s) during the baseline period.
Mean time between failures for the target assets during the baseline period.
List assets included in the pilot and the top failure modes you expect to detect (e.g., Motor A — bearing wear; Pump 3 — seal leak). Separate items with semicolons.
Specify sensor types (vibration, temperature, current, pressure, acoustic, etc.), approximate placement, and target sample rates (e.g., 5 kHz for vibration envelope). Note any constraints (explosive atmosphere, mounting access).
Who will be responsible for collecting, storing, and granting access to raw signals during the pilot? Include system (e.g., historian, cloud bucket) and retention policy.
Define the minimum data quality checks (signal continuity, missing samples, timestamp alignment, SNR thresholds) and who will own verification.
Has controls/automation and IT signed off on connectivity, network security, and data pathways?
How many weeks will you run the pilot to collect required signal diversity and failure events (or to validate predictions)?
Estimate how many true failure or degraded events you need to validate models. If none expected, state how you will simulate or use historical replay.
The percent reduction in unplanned downtime you must achieve to consider the pilot successful.
Maximum acceptable percent of alerts that are false positives, or operationally tolerable alarm load.
If you use precision/recall metrics, specify target precision to ensure alerts are actionable.
If capturing most failures is important, specify a minimum recall target.
Select the groups that have provided sign-off for the pilot.
Describe how operators and maintenance will receive alerts, what actions they will take, and any training or documentation needed.
If yes, specify CMMS and how alerts map to work order priorities.
List the specific metrics to report during pilot (e.g., alerts per week, true positives, false positives, downtime hours avoided, MTTR change, cost avoided).
High-level steps, estimated timeline, required budgets, and responsibilities to scale from pilot to production.
Call out risks such as sensor failure, data leakage, regulatory constraints, or operator overload and how you will mitigate them.
Name and role of the person or governance body that will approve scaling.
List the next practical steps (e.g., purchase sensor X by date, configure historian ingestion, run signal validation script) and owner names.
You can explore this tool now. Sign in or create an account to save your responses and return to them later.
Sign In to Save Create an Account
Make this tool part of your work

Save a personal copy, bring it to your team, or tailor the questions and workflow to fit what you are hungry to improve.

Member customization and team collaboration are coming soon.
Discuss Ask the Librarian

Discussion

Comments and conversation will live here.