A pilot is a measurement, not a demo
It runs on your items, in your building, with your operators, against acceptance criteria written before it starts. If the numbers miss, the design changes before anything is rolled out.
If measured read rates or throughput miss the design targets during the pilot, Arrowhead revises the design before the rollout is committed.
This is Arrowhead's standard practice on RFID engagements, not a published customer outcome. It is the reason the measurement happens on your items rather than on a datasheet.
Five numbers, taken in your operation
Read rate is a physical result of tag choice, item surface, antenna placement, and the RF environment. It cannot be quoted from a specification sheet, which is why the pilot exists.
Read rate on your items
Measured on your own items, surfaces, and packaging — not on a vendor sample. Tag choice, surface, antenna placement, and the RF environment all move this number, which is why it is measured rather than quoted.
Count or capture time
How long the actual operating task takes with the designed process, run by your operators rather than by an engineer who built it.
Accuracy against a known baseline
The pilot result compared with a controlled manual count, so the discrepancy is a measurement and not an impression.
Behaviour under the real environment
Metal, liquid, cold, wash-down, forklift traffic, or whatever your site actually does to equipment — observed during the pilot rather than assumed away.
The data landing where it has to land
The integration contract from discovery, exercised from the capture point through to your system of record.
Scope, duration, and how it ends
- Scope
- One site, or one category, or one workflow. A pilot that tries to cover the estate is a rollout with a smaller budget and proves less than either.
- Duration
- Typically two to four weeks on site. This is a planning range for your calendar, not a commitment: the pilot ends on its acceptance criteria, not on a date.
- Who runs it
- Your operators, running the designed process. A pilot run only by the people who built it measures the engineers, not the operation.
- How it ends
- Acceptance criteria are met and the rollout is scoped, or they are missed and the design is revised. Both are successful pilots; only one leads straight to a rollout.
- What it costs
- Quoted against the design, never as a package. Most of it is the physical build — readers, antennas, tags, printers, and the survey and installation labour around them. The software side of a pilot is the smaller part. None of it can be priced before discovery produces a design, because the design is what decides how many read points and devices there are.
- What you keep
- The measurements, the design, and the acceptance record — whether or not you proceed with Arrowhead.