RFID and RTLS feasibility study: what one should actually contain.

It is a decision document, not a technology demonstration. It should be capable of concluding no, or yes but not the way you were planning, and it should reach that conclusion before capital is committed.

Three questions define it:

• Is it physically possible in this building? Radio behaviour depends on steel, concrete, racking, liquid, ceiling height and what your product is made of.
• Does the operation change if it works? A system that produces accurate data nobody acts on has failed, however well it performs.
• Is the cost proportionate to the benefit? Including tags, infrastructure, integration, and the years after year one.

A study that cannot fail is not a study. If the output was always going to be a recommendation to proceed, you commissioned a proposal.

Asking a vendor to assess the feasibility of their own product.

This is not about dishonesty. A vendor genuinely knows their technology best, and they will design a study their technology passes, because that is the technology they understand. The awkward findings — that choke points would be enough, that half the assets do not need tagging, that the existing barcode process is fine — reduce their revenue and rarely surface.

If a vendor must run it, at minimum define the success criteria yourself, in writing, before they start.

1. Physical assessment. Construction, racking material and density, ceiling height, obstructions, liquid and metal near the tagged items, hazardous area classification, temperature range, and the electrical noise environment. This determines which technologies are viable before any of them are compared.

2. The requirement, stated as a decision. Not "track assets" but "know within X minutes that a pallet has left staging, so that Y can happen". Accuracy requirements derived from decisions, not chosen as a number.

3. Item and volume analysis. How many things, how often they move, how long they stay, what they are made of. Tag cost scales with item count, infrastructure with area and precision. Getting these the wrong way round is an easy and expensive budget error.

4. Technology shortlist with reasons for exclusion. Every family considered, with a stated reason each one is in or out. The exclusions are where the value sits.

5. Integration assessment. Which systems must change, who owns them, what identity the tag carries, and what happens when systems disagree.

6. Cost envelope. Order of magnitude by option, with the variables that move it, and an honest statement of what cannot be costed yet.

7. Baseline measurement specification. What to measure now, before anything changes. Skip this and you can never prove the benefit.

8. A recommendation, including the option of not proceeding.

A full RF site survey. That is a design activity and it belongs after the technology decision, not before — surveying for a technology you then reject is wasted money.

A hardware pilot. Pilots answer "does this product work", which is a later question than "is this approach right". Running one first tends to anchor the whole programme on whichever vendor turned up.

Vendor demonstrations. Useful, but they demonstrate the vendor's best case in the vendor's chosen environment. They are evidence about the vendor, not about your building.

For a single site with a defined scope, two to four weeks of effort is usually right. Less than that and the physical assessment is guesswork. Much more and it is drifting into design.

Whatever it costs, judge it against the programme it protects: the study exists to stop you spending the larger number badly.

If it is being offered free by a supplier, it is a sales activity. That may still be useful, but price it as what it is.

• What would have to be true for the recommendation to be do not proceed?
• Which technologies were excluded, and on what evidence?
• What is the accuracy requirement derived from, rather than asserted?
• What is the cost per tagged item at full rollout, not at pilot scale?
• What are we measuring today so we can prove the change?
• Who is paid more if this proceeds?

We do not sell hardware, hold no partner tier with any vendor and take no referral commission, so a study that concludes "do less than you planned" costs us nothing. When the evidence points that way, it is usually the most valuable outcome.

Related: readiness assessment, pilot success criteria, RF site survey checklist, TCO model.

This guide sets out a method based on general engineering and procurement practice; it quotes no vendor figures, standards data or statistics, so there are no external sources to list. Checked on 25 September 2026.

Last updated: 25 September 2026