Unscoped servicing, line-side pallets and unfunded internal moves — three patterns from this week's articles, and why Q4 volume turns each one into a cost.
It is the middle of August, which means you already know what your December looks like. Volumes up, agency cover thin, and the same three or four bottlenecks that hurt you last peak still sitting there unfixed. Here is the thing that catches most UK warehouse directors out: it is almost never the picking that fails. Picking is the bit that got measured, funded and automated. What fails is the movement either side of it — the pallet that sits at the end of the line, the cage that crosses the yard, the truck that goes off the road for a service nobody scheduled. Those moves have no owner, no line in the budget and no number on the board. They still get paid for, in overtime and missed cut-offs.
The through-line across this week's articles is simple: the work that breaks your automation case is the work nobody scoped, because unowned movement does not disappear — it gets absorbed by people until the volume arrives that people cannot absorb.
---
1. The evacuation leg
Ask a production manager why the line stopped and you will get a story about the line. Ask the shift that was actually standing there and you will get a different one: the end-of-line buffer was full. Finished pallets were not cleared fast enough, the buffer backed up into the palletiser, and the palletiser stopped a machine that was running perfectly well.
This is the evacuation leg, and it is the least glamorous metre in the building. It is also the one that decides whether your output number holds at changeover. Every UK FMCG plant has a version of it: a fixed number of pallet positions between the line and the racking, cleared by whoever is free, at a cadence set by how busy that person happens to be. At steady state it works. At shift changeover — when the person who was clearing it is at the handover board — the buffer fills in minutes and the line pays for it.
The safety case runs in the same direction as the throughput case, which is unusual and worth using. HSE records around 5,000 workplace transport accidents a year in Great Britain, roughly 50 of them fatal — which is exactly why UK plants have stopped treating line-side pallet movement as a signage-and-segregation problem and started automating the move itself. Segregation asks a person to behave correctly in a pinch point at the moment they are most rushed. Taking the move off people removes the pinch point.
An automated truck earns its keep here on evacuation, not on picking. It does not get faster than a good driver over a single move. It gets more boring — the same cadence at 06:00, at 14:00, and across the changeover when nobody is watching, because there is no handover for it to attend.
Ask your next vendor: at shift changeover, how many pallet moves per hour does this configuration sustain — and what happens to that number when the buffer is already at capacity?
---
2. The internal move nobody funded
Multi-client 3PL sites have a quieter version of the same problem. Cages, totes and trolleys have to cross the building — from goods-in to the pick face, from the pick face to despatch, back again empty. That movement is real work, it takes real hours, and in most contracts it is not funded anywhere. It gets absorbed by pick labour, because the pickers are the people standing nearest to it.
That absorption is invisible until peak. At normal volumes a picker who spends part of the shift walking a cage across the building still hits rate, so nothing shows on the report. At peak, the same absorption is the reason a client SLA slips — and because the movement was never costed as movement, the conversation with the client becomes an argument about pick rates rather than a discussion about a transport layer that was never funded.
The lever is naming it. A lifting automated robot — a driverless machine that raises, carries and sets down a load and navigates a live building on its own sensing rather than on a fixed track — makes that movement a measurable, costable service line rather than a tax on pick labour. Once the moves have a count and a cadence, they can be priced into the contract like any other activity. That reframing is often worth more than the machine, because it changes what you can bill for.
The multi-client complication is that no two clients want the same flow, and contracts turn over. A fleet locked to one client's layout is a fleet you cannot redeploy when that contract ends — which is the practical argument for keeping the orchestration layer above any single manufacturer's stack, so machines can be moved between sites and clients without a re-integration project each time.
Ask your next vendor: if this contract ends in eighteen months, what does it cost to redeploy these machines to a different site with a different WMS?
---
3. The seven-year line
Capex committees approve fleets on truck price. It is the number on the quote, it is comparable across bidders, and it fits the approval template. What it leaves out is the seven or so years of servicing, spares, software support and — in regulated sectors — requalification that follow, which frequently go uncosted and, worse, unowned.
Unowned is the sharper problem. Under Regulation 5 of PUWER 1998, every UK employer must keep work equipment in efficient working order and in good repair. That makes servicing an autonomous fleet a legal duty sitting with the operator, not a discretionary line that can be deferred to next year's budget when things get tight. A maintenance plan you cannot evidence is not a cost saving; it is an exposure, and the person holding it is you rather than the supplier who sold you the truck.
The practical failure mode is not a missed service. It is discovering, in year three, that spares and engineer availability for your fleet run through a single channel with no alternative, and that the whole-life cost of the estate is now whatever that channel says it is. The capex decision was made on a number that turned out to be the smallest part of the total.
This is where an open, multi-manufacturer approach does unglamorous financial work. When the machines, the fleet management layer and the servicing arrangements are not welded to one supplier, the seven-year line stays contestable — you can re-tender maintenance, source spares from more than one route, and add a different manufacturer's machine to the same fleet without discarding what you already own.
Ask your next vendor: show me the year-one to year-seven cost of ownership, itemised — and tell me which of those lines I can competitively re-tender without replacing the machines.
---
4. Who owns the fleet standard
The three patterns above share a root cause: someone else decided where the boundaries of your operation sit. The buffer was sized by the line supplier. The internal moves were shaped by the client's contract. The maintenance cost was set by whoever holds the parts. In each case the operator inherited a standard rather than setting one.
A fleet standard is the set of things you insist on before any machine arrives: how a robot is told what to do, how it reports what it did, who can read that data, what happens at a handover point, and what has to be true before a second manufacturer's machine can join the same fleet. Write it once, and every subsequent procurement is a comparison rather than a negotiation. Skip it, and each purchase quietly redefines your operation to match the machine that just turned up.
This is the whole argument for vendor neutrality, and it is an operational argument rather than an ideological one. FlyWei is an independent, vendor-neutral UK systems integrator of autonomous forklifts and AMRs — it integrates the best robots across multiple manufacturers; it is not an OEM, reseller or distributor. The reason that structure matters to you is not purity. It is that the party recommending the machine should not be the party whose margin depends on which machine you pick, and the standard your fleet runs to should belong to your operation rather than to a closed bundled stack.
BSI standards and HSE guidance give you the safety floor. The fleet standard is the commercial ceiling — the thing that determines whether, in year four, you are still choosing your suppliers or they are choosing for you.
Ask your next vendor: which parts of this system would still work if I bought my next ten machines from someone else?
---
The arithmetic
- Servicing is not optional spend. Regulation 5 of PUWER 1998 requires every UK employer to keep work equipment in efficient working order and in good repair — so fleet maintenance is a legal duty on the operator, not a line that can be cut in a tight year.
- The safety case and the throughput case point the same way. HSE records around 5,000 workplace transport accidents a year in Great Britain, roughly 50 of them fatal. Removing people from the repetitive pallet-move pinch points addresses both at once.
- Unfunded movement is still paid for. Cage, tote and trolley moves absorbed by pick labour never appear as a transport cost; they appear as a pick-rate problem at peak, in the one week you can least afford to argue about it.
- Illustrative worked example — the changeover gap. Say a site runs two shifts with a 20-minute handover at each changeover, and the end-of-line buffer holds 30 minutes of production. Two changeovers a day means the buffer spends roughly 40 minutes daily being cleared by nobody in particular. Multiply by your own cost of a stopped line and you have the number that should sit next to the truck price. These figures are illustrative — put your real buffer depth and handover length into the same shape.
- Illustrative worked example — the whole-life ratio. Take a hypothetical seven-year fleet where the purchase price is the only figure in the approval paper. If servicing, spares, software support and requalification together match the purchase price over that period, the committee approved roughly half the decision. Run your own numbers; the point is the shape, not the specific ratio.
---
What to do on Monday morning
- Count one unowned move. Pick a single flow — end-of-line evacuation, or empty-cage return — and count it for one shift: how many moves, how long each takes, and who did them. You will almost certainly find the work sitting with someone whose job description does not mention it.
- Pull the maintenance evidence for your existing MHE, not the new stuff. If you cannot produce a current, evidenced maintenance record for the trucks you already run, that gap will not improve when the fleet gets more complex. Fix it on the estate you have before you scale it.
- Write one page of fleet standard before the next quote arrives. Data access, handover definitions, orchestration interface, and what has to be true for a second manufacturer's machine to join. One page, agreed internally, changes every vendor conversation you have for the next three years.
---
If any of this is uncomfortably familiar, reply to this edition or leave a comment. Happy to walk through what an open-fleet design would look like for your specific flows — a quiet read of your operation, no deck, no pitch.