Outsourcing a temperature mapping study: what to specify

• Acceptance criteria from your products, and how uncertainty is treated.
• Scope: areas, zones, empty and/or loaded, seasonal runs, door-opening and power-failure tests.
• Sensor count and placement rationale, to be approved before the study.
• Logger calibration: certificates from an accredited laboratory, points spanning the study range, verification after the study.
• Sampling interval and duration.
• Report content: statistics per position, hot/cold spots, recommended monitoring positions, deviations, raw data files.
• Data ownership and retention: you keep the data.
• Contractor's quality system and staff competence.

ValiTrac's mapping engine is designed to analyse contractor-delivered data independently, so a report can be checked rather than accepted on trust.

What belongs in the specification

A contractor given a room and a date will produce a study that satisfies their own template. It may be excellent and still not answer your question, because your acceptance criteria, your seasons and your products were never stated. Everything below belongs in the request, before a price is agreed.

Insist on the raw data

This is the single most important line in the contract and the one most often left out. A PDF report is a conclusion; the raw export is the evidence. Without it you cannot check whether the stated maximum is really the highest reading, whether a sensor flat-lined, whether the interval was what was specified, or how many readings there actually were. You also cannot re-analyse the space against a new product's tighter limits in two years without paying for a new study.

Specify the format — CSV with declared units and ISO timestamps — and specify that it is delivered unchanged alongside the report. A contractor unwilling to hand over the data they charged you to collect is telling you something.

Checking what comes back

• Count the readings. Sensors times duration divided by interval gives the expected number; a shortfall locates itself in the data.
• Check the extremes against the per-sensor tables rather than the summary, and confirm the stated hot spot is the sensor with the highest maximum under the criterion the protocol named.
• Look for flatlines and gaps yourself, and check each one has a recorded decision beside it.
• Confirm the calibration certificates cover the range the study actually reached, not just the target band.
• Confirm the report names the decision rule if it makes a pass statement, and that the acceptance criteria are yours and not a generic default.
• Check the recommendations are actionable and owned. A recommendation to move a monitoring probe, filed and never implemented, documents that you knew.