Service strikes rarely happen because nobody cared. They happen because the day got busy, the gang changed, the drawings weren’t the latest, and “it’s only a shallow dig” turned into a full-depth excavation with a breaker on standby. A solid permit to dig process is there to slow the job down at the right moments, so you can still keep programme without gambling with buried electricity cables, gas mains, water, fibre, drain runs and unknown legacy services.
On UK sites, a permit to dig should be treated as the working permission for any ground disturbance, not a piece of paper to satisfy someone in the office. Done properly, it ties together utility information, scanning, trial holes, supervision, exclusion zones, plant controls and stop-work triggers. Done poorly, it becomes a signature hunt and everybody assumes someone else has “sorted the services”.
How a permit to dig should control the risk, not just document it
A permit to dig system works when it forces three things to happen in the right order: information is gathered, the ground is physically verified, and the work is supervised so the conditions don’t drift. It also creates a clear line: no permit, no dig.
The “essentials” aren’t complicated. The details are where most strikes are either prevented or baked in.
# Key risk picture: why underground services are different to other hazards
Underground services combine hidden conditions with high-consequence outcomes. Unlike work at height or hot works, you can’t see the hazard once the ground is opened unless you deliberately expose and protect it. Even when you have utility drawings, they can be out of date, off line, or only show mains and not private runs, diversions, or temporary supplies added during earlier phases.
The permit to dig needs to recognise typical failure points:
– Multiple trades disturbing the same area across days (drainage, civils, landscapers, groundworkers, electricians).
– Plant arriving after the permit was issued (mini-digger swapped for 13T or a breaker introduced).
– Conditions changing (dewatering, slumping sides, reinstatement removed, surfacing stripped back).
– Interfaces (road opening permits, traffic management changes, live site/public boundary constraints).
# What “valid information” looks like on a real site
Good permits start with evidence, not assumptions. Before anyone allocates a gang or puts plant on the pad, the permit pack should pull together:
– Current utility plans for the area, including any updated sketches from previous excavations or as-built records.
– Known site temporary services and phase-specific diversions (power to welfare, temporary lighting, temporary water).
– A marked-up area plan showing the exact proposed dig footprint, depth intent, and access routes for plant.
– Records of previous strikes, near misses, or “unmapped service found” notes on that project.
If you can’t confidently say you’ve got the latest picture of what’s buried, that’s the reason to pause—not a reason to “dig carefully and hope”.
# Scanning and locating: making it more than a box-tick
A permit to dig that says “scan completed” but doesn’t state what equipment was used, who did it, what was found, and how it was marked is weak. In practice, a usable permit should capture:
– The locator type used and its condition (calibration/functional check as per site practice).
– The date/time of scan and whether it was repeated after any changes (rain, traffic loading, ground opened elsewhere, plant moved in).
– The markings method (paint/chalk/flags) and how they’ll be protected from weather and traffic.
– Any limitations (congested services, reinforced concrete, metallic clutter, depth uncertainty).
Scanning shouldn’t be treated as the answer. It’s one layer. When the scan results are unclear, or the risk is high, you move to hand-digging and controlled trial holes to prove the location before mechanised excavation.
How it plays out on site: a realistic near-miss chain
A civils gang is trenching for a new drainage run on a retail park refurb, working around a live car park that stays open until 10am. The permit to dig is signed at 7am based on plans pulled the previous week and a scan done at the compound entrance. By 9:30, the traffic marshal shifts the barrier line to keep customers moving, and the trench line is nudged half a metre to avoid the new route. The excavator operator swaps buckets and starts taking a deeper cut because the formation is softer than expected. The groundworker notices an old paint mark has been scraped away by a delivery wagon and the service route isn’t obvious anymore. Seconds later the bucket teeth snag something; a plastic duct lifts, and there’s a split second where everyone realises it could be live. Work stops, the area is taped off, and the supervisor discovers the shifted alignment moved them into a service corridor not included in the original permit sketch. The strike is avoided, but only because someone spoke up before the machine carried on.
That’s what “permit drift” looks like: small changes that collectively invalidate the original assumptions.
Supervisor interventions that prevent “permit drift”
Permits to dig fail most often after they’ve been issued. The supervisor’s job is to keep the permit conditions true as the task moves from briefing to breaking ground.
Practical interventions that work:
– Treat any change in line, level, plant, or method as a trigger to pause and re-validate the permit, not as an admin hassle.
– Keep the permit alive at the excavation, not in the site office. If it’s not with the gang, it’s not controlling anything.
– Push for demonstrated understanding: ask the operative to point out the marked services and explain the hand-dig zone, rather than asking “everyone happy?”
– Manage interfaces: stop other trades using the same corridor for storage, skips, or welfare routes that destroy markings or force alignment changes.
– Make “hand dig only” a controlled zone with barriers and clear limits, not a vague instruction in a RAMS.
The permit to dig checks that actually change behaviour
A permit to dig should drive a short, repeatable set of controls that everybody recognises. Use this on the ground as a quick confirmation, not a long form exercise.
– Confirm the exact dig footprint is set out and matches the permit sketch (including offsets for access and spoil).
– Ensure scan marks are visible, protected, and understood by the excavator operator and banksman.
– Establish a hand-dig/trial hole zone and keep mechanised excavation outside it until services are exposed and protected.
– Put exclusion controls in place: pedestrian segregation, plant slewing limits, and a clear stop signal agreed with the operator.
– Identify what happens if a service is exposed: support, protect, notify, and do not backfill until authorised in your site process.
– Verify emergency arrangements are known: who isolates, who calls, and how the area is made safe if a cable/pipe is damaged.
# Common mistakes
1) Issuing one permit for “the whole area” and letting gangs work wherever they like under the same signature. It removes the discipline of defining an exact dig footprint and encourages scope creep.
2) Relying on utility plans as proof of location. Plans are a starting point; the ground still needs to be verified with scanning and exposure where necessary.
3) Allowing mechanised excavation to “just take the top off” in uncertainty zones. Many serious incidents start with a shallow scrape that catches a duct, warning tape, or service marker.
4) Treating the permit as valid for days without re-confirming after weather, traffic, or site changes. Markings fade, barriers move, and the job subtly migrates away from the original safe line.
Keeping momentum without shortcuts: sequencing that holds up under pressure
Programme pressure usually arrives as “we just need a couple of metres today” or “can we get the trench in before the delivery?”. Good sequencing lets you keep moving without breaking the system.
Practical sequencing steps:
– Start with proving holes at the highest-risk crossings first (where services are most likely) so you don’t commit to a long trench you later have to redesign.
– Keep spoil and materials out of the service corridor. Overloaded edges and clutter force plant repositioning and increase accidental contact.
– Use clear physical protection once services are exposed: supports, covers, and distance controls so a later gang doesn’t undo the good work.
– Plan reinstatement so it doesn’t hide unresolved issues. If a service is uncovered unexpectedly, treat it as a design/interface issue, not something to “bury and forget”.
# A one-week dig-control tune-up
1) Walk the active and upcoming dig zones with the supervisors and physically mark where permits begin and end, so scope creep is harder to justify.
2) Swap the daily briefing question from “have you got a permit?” to “what would invalidate your permit today?”, and capture the answers.
3) Assign one competent person per shift to own locator use and markings integrity, including re-marking after rain or traffic.
4) Re-sequence the top two excavations this week to start with trial holes at crossings, even if it means a short delay in machine hours.
5) Agree a simple escalation rule: any uncertainty, any uncharted duct, any move in alignment equals stop and supervisor attendance before the bucket goes back in.
What to watch on the next dig
Underground service control is vulnerable to competence drift: new starters, agency labour, and changing gangs can turn a solid method into “we’ve always done it this way” in a week. Keep an eye on permit validity in the middle of the shift, not just at sign-on, and look for the small changes that quietly move work outside safe assumptions. The best question for your next briefing is: what has changed since the permit was issued, and who has authority to stop the dig when it does?
