The Long Tail of Removal: What a UK Smart Street Owes Once the Pilot Ends

A pilot that never formally ends is not a pilot

Local authority officers have become practised at the procurement of novelty. A lamppost-mounted air-quality node, a kerbside camera counting cyclists, a bin-fill sensor bolted to a litter bin outside a leisure centre — each arrives through a framework, a challenge fund or a partnership with a university, and each is announced as a step towards something larger. What is far less rehearsed is the moment the trial concludes and the street is expected to return to being a street.

Across the UK in 2026, a cohort of smart-street and connected-places pilots funded in the middle of the decade is reaching the end of its initial term. Many were designed around twelve- or twenty-four-month windows of insight. Fewer were designed around the question of what happens on day seven hundred and thirty-one. The result is a growing category of municipal problem that does not appear in portfolio dashboards: equipment that nobody has budgeted to remove, data that continues to accumulate in a supplier's tenancy, and an asset register that does not agree with what is actually screwed to the public highway.

Where the gap sits in the paperwork

The decommissioning gap is rarely the product of negligence. It emerges because the documents that govern a pilot are written to answer different questions. A business case addresses benefits and risk to delivery. A grant funding agreement addresses eligible spend and reporting. A supplier contract addresses performance, payment and, if the authority is fortunate, termination. None of these necessarily addresses physical removal, cryptographic erasure of derived datasets, or the transfer of a working integration into a successor system.

There is also a structural prompt in the wrong direction. Innovation funding rewards demonstrable delivery and discourages the appearance of failure, so the closure phase attracts little organisational energy. Officers who championed an installation may have moved on by the time it needs retiring. Contractual notice periods lapse quietly. By the time the question is asked in earnest — often prompted by an insurance renewal, a streetworks permit or a Freedom of Information request — the commercial leverage has gone.

Removal is not one act but four

Treating decommissioning as a single line item is the most common error. In practice it has at least four strands, each with a different owner, timeline and evidentiary requirement.

Physical recovery. Brackets, cabling, enclosures, power supplies and the fixings that held them. A surprising volume of pilot hardware is never recovered, because power and mounting were supplied by a third party — an estate, a bus shelter operator, a highways contractor — who was never party to the original agreement. Small cell sites left dormant on street furniture remain on UK streets long after the service behind them has ceased.

Data disposal. Deletion must be defined precisely: raw sensor streams, derived aggregates, trained models, backups, and any copies held by analytics sub-processors. A written certificate of destruction, naming the datasets and the date, is worth more than an assurance in an email.

Interface handover. Where a pilot produced something the authority wishes to keep — a calibrated baseline, a data schema, a dashboard configuration — the exit plan should specify the format in which it is transferred and whether any licence survives termination.

Liability settlement. Who indemnifies whom if the hardware fails after removal, if a data subject exercises rights over pilot-era records, or if a defect is discovered in a unit now in a skip. These questions have long tails, and they should not be answered by silence.

A sunset ledger worth the name

A practical exit schedule can be drafted in a few pages and appended to the main contract at award. It should contain, at minimum, the following.

• Named exit trigger. A calendar date, not a vague reference to project closure, together with the notice period required from either side.
• Removal specification. A schedule of equipment, mounting locations, responsible party, restoration standard for the highway or street furniture, and the permits required under the Traffic Management Act 2004.
• Financial provision. A retention, bond or priced removal option fixed at signature, so the cost is not renegotiated in a market where the authority holds no leverage.
• Data register and deletion protocol. Datasets identified by name, retention periods, sub-processor list, and the form of the destruction certificate.
• Asset and record transfer. The fields to be populated in the corporate asset register, including location, serial number, custodian and disposal route.
• Survival terms. Which clauses — confidentiality, liability, records retention under the Public Records Act — persist beyond termination, and for how long.
• Alumni briefing. A short handover note for the officer who inherits the residue, written by the officer who signed the contract.

The contractual culture that makes exits routine

None of this is novel contract law. It is closer to the discipline that already exists in facilities management, where equipment registers and lifecycle replacement are ordinary practice, or in property leases, where dilapidations are priced before occupation rather than after. Digital pilots have escaped that discipline because they were framed as experiments rather than estates.

The correction is largely cultural. Authorities that add a decommissioning section to the standard evaluation template, scored alongside data protection and social value, will find suppliers begin to price exit honestly. Buyers on Crown Commercial Service frameworks can push the same requirement through call-off terms rather than renegotiating bespoke contracts. Portfolio boards can ask for a retirement schedule as routinely as they ask for a delivery plan.

There is a defensible argument that some pilots deserve to become permanent services, and that a heavy exit burden discourages worthwhile experimentation. That is a real tension. The answer is not to make exit painful, but to make it legible — a known cost, agreed in advance, that a council can weigh against the value of continuing.

What the street is owed

A UK high street carries obligations to the people who use it. A sensor that no longer reports but still watches, a dataset nobody can locate, a bracket left to corrode above a bus queue — these are small failures individually, and a reputational problem collectively. Smart-city work earns its licence to continue from the visible competence of its endings.

Authorities preparing their 2027 pipelines would do well to write the removal clause before the press release. The question is not whether a pilot will end. It is whether anyone will be able to say, with evidence, that it did.