How to recover a failed RTLS pilot
Pilots run in a controlled corner with the best tags, readers and network. Rollouts don’t. Here is how to diagnose why a pilot stalled — and the steps that get it moving again.
Why pilots stall
- Accuracy that held in a clean corner collapses in production RF (metal, racking, forklifts)
- No integration to the systems people actually use, so the data is never adopted
- Tag/inlay or anchor density chosen for a demo, not for scale or cost
- No clear KPI or business case, so nobody will fund the rollout
- A vendor optimising for their hardware, not your outcome
The recovery sequence
- Diagnose: a fast, independent read of where accuracy, integration or economics actually broke
- Re-baseline the use case and the KPI that justifies spend
- Fix the physics: site survey, tag-on-material testing, read-zone or anchor redesign
- Wire it into the workflow so the data is used, not just displayed
- Re-pilot against the real environment, then scale through gates
Why independence helps here
When a pilot is failing, the incumbent vendor is rarely the right party to grade their own homework. An independent advisor with no reseller margin can tell you plainly whether to fix, re-scope or change approach — and we can drop in at any stage.
Independent recovery: diagnose physics, scope, integration and adoption separately
A stalled locating pilot is usually not “the technology failed.” It is one or more of: RF physics never validated under peak load; success criteria written after the vendor demo; integration deferred to phase 2; operators never changed how they work; or the business case used lab accuracy as if it were production accuracy. Recovery starts by separating those failure modes so you do not rip out a radio that was fine and leave the real problem untouched.
Recovery sequence that works. (1) Freeze further hardware spend. (2) Reconstruct the original KPI and baseline — if none exists, measure current search time, WIP accuracy, dwell or utilisation before changing anything. (3) Run an independent RF and process audit: multipath, tag mounting, locator/anchor geometry, update rate, and whether the pilot bay still matches the live layout. (4) Score integration: does the event reach WMS/MES/EMR/CMMS, or only a vendor dashboard? (5) Score adoption: who was supposed to act differently in week one? (6) Decide repair, re-scope, re-pilot, or replace with written gates.
Common repair patterns. Retune density and orientation instead of adding kit; change tag form factor or mount point; drop accuracy requirements to what the workflow truly needs; move from all-UWB to hybrid BLE + UWB hotspots; or keep the radio and rebuild the integration and SOP layer. Many “failed” pilots become successful programmes once the locating question is restated in operational language.
When to change vendors vs when not to. Change when the platform cannot export events, cannot meet a realistic accuracy band after proper design, or locks you into proprietary tags with no exit. Do not change merely because a sales engineer oversold a demo bay. Replacing kit without fixing scope and adoption repeats the failure at higher cost.
Re-pilot rules. Rental or existing kit only; 8–16 weeks under full shift load; one primary KPI; one core-system integration; named operations sponsor; go/no-go thresholds signed before work restarts. Independence matters here because the original vendor is structurally incentivised to defend the first design.
Communications and commercial reset during pilot recovery
Recovery is as much a stakeholder process as an RF process. Brief the steering group with a fault tree: physics, product fit, integration, adoption, economics. Avoid blame narratives that freeze learning. If the original success criteria never existed, write them now against a fresh baseline before spending more capital.
Commercially, open a structured conversation with the incumbent: remediation credits, rental extensions, density redesign without automatic hardware uplift, or an agreed exit if the platform cannot export events or meet a realistic accuracy band. Parallel evaluation of an alternative radio should be evidence-led, not punitive theatre.
Capture lessons in a short locating playbook: which assumptions failed, which zones need different accuracy classes, which SOPs operators ignored and why. That artefact is what makes the re-pilot cheaper than the first attempt and what protects the next site in a multi-site programme.
Ninety-day recovery plan template
Days 1–15: freeze spend, rebuild baseline, independent RF/process audit. Days 16–45: remediate density, tags, integration or SOPs per the fault tree; agree re-pilot gates. Days 46–90: production-load re-pilot with weekly steering. Exit with go, repair-again, or replace. Keep communications factual and KPI-led so the organisation can learn without relitigating the original vendor beauty contest.
Do not skip the baseline on recovery
Even in rescue mode, spend the week to re-baseline search time, dwell, utilisation or read-rate before changing architecture. Without that measurement, you cannot prove the recovery worked — and the steering group will not fund the next gate with confidence.
Frequently asked questions
Can a failed RTLS pilot be saved?
Often, yes. Most failures are fixable: RF design, tag selection, integration or scope — not the whole concept. The first step is an honest diagnosis.
How fast is a diagnosis?
Usually a short, fixed-fee review that pinpoints what broke and whether to fix, re-scope or switch approach.
Do we have to change vendor?
Not necessarily. Sometimes the kit is fine and the design or integration isn’t. We recommend what fits, with no reseller margin.
Can you take over a stalled programme?
Yes — you can drop us in at any stage, including rescuing a rollout that has lost momentum.
Last updated:
Advisory note
This guide is written for operators selecting and deploying location technology — not for selling a radio. Use it to frame questions for vendors and integrators; bring TRACIO in when you want those answers scored independently against your site.