Warehouse Automation for Mid-Sized Operations: Robots-as-a-Service or Buy?

CV-Library reports that logistics and warehousing vacancies rose 9% in Q2 2025 compared with Q1, from 137,542 jobs posted to 150,578. Demand for warehouse workers is therefore rising, not easing, at a time when employers compete for the same labour pool. At the same time, Make UK and PwC note the main National Living Wage rate rose 6.7% to £12.21 per hour from April 2025, lifting the floor on entry-level pay.

Make UK and PwC estimate that UK manufacturing loses approximately £7 billion every year owing to unfilled vacancies. The figure refers to manufacturing, not warehousing, but the dynamic is shared across the supply chain: unfilled roles cause lost output, overtime premiums and service failures. For a distribution centre, the equivalent is missed dispatch cut-offs, agency dependence and rising error rates.

UK retailers surveyed by IT Brief UK reflect the same strain: 58% say their current fulfilment infrastructure is holding them back, and 56% say warehouse limits directly affect growth. For the COO of a mid-sized operation, the pressure is double: labour cost is rising and labour availability is falling, while customers ask for faster and more accurate fulfilment. Automation is a response to that squeeze, not a technology project for its own sake.

What to do: frame the case in operational terms before financial ones. Quantify the cost of vacancies, agency premiums and errors in your own site, by shift and by task. Identify the tasks that are repetitive, physically demanding and hard to recruit for. These are the candidates for the first deployment, because the benefit is visible to the workforce as well as to the finance team.

There is a workforce dimension that the financial case tends to omit. Operators who have deployed robots often find that the jobs that remain are different rather than fewer: exception handling, replenishment, supervision and maintenance. Planning for this early, with training and clear communication about what will change, reduces resistance and protects performance. Sites that present automation as a way to remove the hardest tasks, while keeping and developing people, usually recruit and retain better than those that present it as a cost programme.

CBRE reports that approximately 60% of warehouses use entry-level or base automation and only 1-2% operate an advanced fully automated specification. It also found that 80% of occupiers felt warehouse robotics were the top technology disruptor to logistics supply chains. The gap between belief and adoption is wide: operators see the shift coming, but few have moved beyond conveyors, scanning and basic software.

The retail survey reported by IT Brief UK points the same way on expected benefit. Some 88% of respondents believe robotics and automation will help them cope with demand surges, and almost 7 in 10, at 69%, already see tangible benefits. The main benefits reported were operational efficiency (41%), scalability (28%) and accuracy (26%). These are practical gains, which suggests mid-sized operators do not need to wait for fully automated sites to see results.

The same survey identifies the barriers. About 34% of respondents highlighted concern about the upkeep of systems over time, and 32% cited difficulty connecting new technology with existing systems. Both are operating problems rather than technology problems. They are relevant to the buy-or-subscribe choice because they determine who carries the maintenance and integration burden, and how quickly problems can be fixed.

What to do: place your site on the adoption ladder, from manual to entry-level mechanisation, to goods-to-person and mobile robots, to fully automated. Choose the next rung, not the top one. For most mid-sized operations, the next rung is mobile robots or goods-to-person picking in a defined zone, where the benefit is measurable and the disruption is contained.

A second consideration is the customer contract. Many mid-sized operators are third-party logistics providers whose revenue depends on contracts of limited duration. Investing in assets sized for one customer's profile carries the risk of stranded capital if the contract ends. Subscription aligns cost with the contract life, and in some cases the cost can be passed through to the customer explicitly. Operators should discuss automation with key customers early, because shared benefit can support longer terms and better pricing.

Benchmarking your own site against these figures is straightforward. Ask whether pick, put-away and replenishment each have a defined level of mechanisation, whether inventory and order data flow automatically between systems, and whether anyone is accountable for continuous improvement. Sites that answer no to most of these are probably in the 60% at entry level, and the next rung is likely to deliver more than a leap to the top of the ladder.

The International Federation of Robotics reports 102,900 transportation and logistics service robots sold in 2024, up 14%, within professional service robot sales of almost 200,000 units, up 9%. It also reports the robot-as-a-service fleet grew by 31% and that RaaS had a growth rate of 42% in 2024. The IFR cautions that its figures come from a sample of 294 suppliers, so the direction is firmer than the detail.

The market outlook is positive but uneven. Interact Analysis predicts mobile robot revenue rising from just under $5 billion in 2024 to $14 billion in 2030, below its earlier $15.6 billion estimate. It forecasts warehouse automation order intake to rise 7% in 2025 and 6% in 2026, after 1% revenue growth in 2024. Demand is growing, but at a slower pace than was once expected.

Subscription growth fits this environment. When forecast growth is uncertain, buyers prefer arrangements that let them scale up and down, and suppliers prefer recurring revenue to lumpy project sales. For a mid-sized operator, a subscription also moves the supplier's incentives. A vendor paid per robot per month has a reason to keep the fleet running and productive; a vendor paid at installation has less reason to care once the cheque has cleared.

What to do: ask each supplier for both a purchase quote and a subscription quote for the same scope, with the same service levels. Compare them on total cost over three and five years, including maintenance, software, upgrades and exit costs. Do not accept a subscription price that excludes services you would need to buy separately under ownership.

The gap between belief and adoption also tells operators something about timing. When most of the market is at an early stage, the first movers in a region or sector can gain a service advantage, but they also carry the learning costs. For a mid-sized operator without large engineering resources, a measured second-wave approach, using proven applications with references, is often more sensible than a pioneering one. The question is not whether to automate, but which proven application to adopt first.

McKinsey's analysis of automation in logistics offers a useful caution. It notes that most logistics contracts run for about three years and are often not profitable for two, leaving only about a year of profit before renegotiation. It also observes that a lot of equipment sits idle much of the year, because it is used for only one or two shifts a day. Capital tied up in idle equipment is where ownership economics fail.

The same McKinsey article records that purchase prices for the same equipment can vary by as much as 50 percent between buyers. That variation reflects negotiation, volume and timing, and it means a poor purchasing process can erase the advantage of ownership. A subscription, by contrast, is priced by the vendor against its own fleet economics, so the price risk is lower, though the margin is embedded in the monthly charge.

The decision therefore turns on three variables. The first is utilisation: how many hours a day the system will run. The second is stability: how confident you are that the process and volumes will persist for the life of the asset. The third is contract length: whether your customer contracts cover the payback period. Where all three are favourable, ownership usually wins on cost. Where any one is uncertain, subscription protects against being wrong.

What to do: build a simple model with three scenarios for volume, one conservative, one central and one optimistic, and compute total cost under ownership and subscription for each. If ownership only wins in the optimistic scenario, subscribe. If it wins in the central one and your customer contracts run longer than the payback period, consider buying. Revisit the decision each year as the data improves.

Mid-sized operators should also read the market forecasts with care. Interact Analysis has revised its outlook as conditions changed, which is a sign that even specialist forecasters find the pace of adoption hard to predict. A business case that depends on the market growing at a particular rate, or on prices falling on a particular schedule, is fragile. The more robust case rests on your own site data: measured throughput, measured labour hours and measured cost per order.

It is worth testing the model against the worst plausible year rather than the average. A subscription that costs more in a strong year but can be scaled down in a weak one may be cheaper over a cycle than an owned system that runs below capacity for months. The reverse also holds: a steady, high-volume process punished by subscription premiums for years is a clear sign that ownership should have been chosen.

Make UK and PwC found that 68% of manufacturers see enhanced productivity as critical to mitigating cost pressure, citing energy efficiency (46%) and increased investment in automation (31%) as ways to achieve it. The numbers show automation is important but not the first response for most; many firms start with cheaper efficiency steps. In a warehouse, the equivalent is improving layout, slotting and processes before buying robots.

Integration is the point at which many projects stumble. Connecting robots to a warehouse management system, to order feeds and to the physical layout requires work from both vendor and customer, and it takes longer than demonstrations suggest. The 32% of retail respondents who cited difficulty connecting new technology with existing systems are an indication of how common the problem is. Subscription does not remove this, but it can limit the capital at risk.

Maintenance and change management are the other hidden costs. Who services the fleet, how quickly, and with what spare parts policy? How are operators trained, and how do shift patterns change? The 34% of respondents worried about upkeep were identifying a real operating risk. Under ownership, the operator carries it, perhaps through a service contract. Under RaaS, it is usually bundled in the fee, which is part of what you are paying for.

What to do: write a responsibility matrix before signing. For each of integration, installation, maintenance, software updates, training and spare parts, name who is responsible, the service level and the penalty for failure. Ask for references from sites of similar size and complexity, and speak to their operations managers, not only to their sponsors. The answers will tell you more than the specification sheet.

Total cost comparisons should be built to withstand challenge from the finance director and the vendor. Include the cost of capital, the expected life of the assets, the residual value, the cost of integration and downtime, and the cost of labour avoided at realistic rates. For subscriptions, include price escalators and the cost of exit. Present the result as a range rather than a single figure, and show which assumptions drive the difference, so that decision-makers understand what they are betting on.

Energy and space are also part of the cost picture. Make UK and PwC found 46% of manufacturers pointing to energy efficiency as a response to cost pressure, and a robotic system changes a building's power profile, charging needs and layout. Ask suppliers to document the energy consumption, charging infrastructure and floor-space requirements, and include them in the comparison. They are modest items individually, but they can shift the answer in a close decision.

A subscription can protect the buyer or trap them, depending on the terms. The key clauses are the minimum term, the minimum fleet size, the price uplift on renewal, the process for scaling up and down, the uptime commitment and the remedy if it is missed, ownership of operational data and the exit conditions, including removal costs and any buy-out price. A low monthly fee can conceal an inflexible commitment.

Scaling terms deserve particular attention. A contract that permits adding robots for peak season, but not reducing the fleet afterward, offers flexibility in only one direction. Similarly, a contract that lets the vendor change the price at renewal without a cap exposes the operator to the cost of switching. Uptime commitments should be measurable, with clear definitions and credits, and aligned to the hours the operation actually runs.

Data ownership is increasingly important. The operational data from robots, covering pick rates, travel distances, faults and congestion, is valuable for optimising the site. Operators should insist on access to their data in a usable format and on the right to take it with them if they change vendor. Without this, the subscription becomes a lock-in not through price but through the loss of the performance history that proves the case.

What to do: have the draft contract reviewed against these clauses before commercial terms are agreed, and negotiate the exit conditions at the outset, when the vendor most wants the deal. Include a buy-out option with a pre-agreed formula, so that the move from subscription to ownership is a choice you control and not a negotiation held when the vendor has the upper hand.

One practical protection is to separate the technology from the commercial wrapper. Ask suppliers whether the same robots can be offered under purchase, lease and subscription, and whether a subscription can convert to ownership with credit for fees paid. A vendor that offers all three is signalling confidence in its product and is more likely to be flexible as your needs change. A vendor that offers only one is asking you to accept its preferred risk allocation, which may not be yours.

The staged route has three phases. In the first, a bounded subscription deployment in a single zone, run for long enough to cover a peak and a trough. In the second, a review against pre-agreed metrics: throughput per hour, accuracy, uptime, labour hours saved and total cost per order. In the third, a decision to expand, buy, renew or exit, based on the data. The structure keeps the capital at risk small while the learning is large.

Metrics should be agreed before the pilot starts, with baselines measured in the same period. Without a baseline, any improvement is anecdotal. Include operator feedback in the review, because a system that frustrates the workforce will not sustain its performance. Include the vendor in the review, because their ability to respond to problems is part of what you are buying, and the pilot is the best opportunity to test it.

As volumes stabilise and utilisation proves out, selectively buy. The assets that run most hours, in the most stable processes, are the ones where ownership gives the greatest return. Keep subscription for capacity that is seasonal, uncertain or tied to a customer contract that may not renew. The mix is likely to change over time, and the contracts should allow for it, with conversion rights and credits for subscription fees paid.

What to do: set a decision calendar. At month three, check integration and early performance. At month six, review the first full cycle of data. At month twelve, decide on expansion or ownership. Assign the COO as the accountable owner, with finance, IT and operations represented, and report to the executive team with the same discipline as any capital project. Treat the first year as a structured experiment with a decision at the end.

Finally, remember that automation is part of an operating model, not an overlay. Layout, slotting, replenishment rules, order profiles and shift patterns all interact with the robots, and performance depends on how well they are tuned together. The best results usually come from sites that treat the deployment as a continuous improvement programme, with weekly reviews, a named owner and a backlog of adjustments. Whether the robots are owned or subscribed, that discipline determines the return.

Governance keeps the staged route honest. Without a named owner and a calendar, pilots drift: extended without decisions, expanded without evidence or abandoned without learning. Record the decision criteria at the start, keep the pilot report short, and share it with the workforce as well as the board. A culture in which the pilot is allowed to fail cleanly is more likely to produce a successful rollout than one in which failure is hidden.

Karter McKinley's AI and Technology Implementation service helps operators build the baseline, the vendor comparison and the staged deployment plan described here, including contract review against the clauses that matter and measurement of pilot results. We are independent of any robot vendor.