Staff Shortage in Commercial Cleaning: What the Numbers Actually Say

The Bundesinnungsverband des Gebäudereiniger-Handwerks — the German building cleaners’ trade association — publishes an industry report compiled by IW Consult, the consulting arm of the Cologne-based German Economic Institute. The June 2025 edition puts the sector at 34,824 businesses, 658,325 employees and 27.55 billion euros in revenue, with the trade census of the Federal Statistical Office as the underlying data basis (reporting period through 2023). The association describes the trade as Germany’s largest by employment, and the share backs that up: 11.2 percent of all employees in the skilled trades subject to social insurance contributions worked in building cleaning in 2023. For scale, the Federal Statistical Office counts around 564,000 skilled-trade businesses, 6.0 million people employed in them and 762 billion euros in revenue for 2024 across all trades.

What makes the staffing situation hard is not decline — it is growth. Over the 2008–2023 reporting period of the trade census, employment subject to social insurance grew by 51.7 percent in building cleaning against 14.9 percent across the trades as a whole. Per the same report, 29.6 percent of all additional jobs subject to social insurance created in the German skilled trades came from building cleaning alone. Revenue moved the same way: between 2015 and 2024 it rose 80.2 percent in building cleaning against 40.1 percent across the trades.

That is the shape of the problem, and it is worth stating plainly because it contradicts the usual framing. This is not a shrinking industry losing people. It is an expanding industry that has to recruit continuously just to stand still against its own order book — in a labour market where every other entry-level employer is recruiting from the same pool at the same time. Demand for cleaned square metres grows faster than the number of people willing to clean them at the hours the work has to happen.

One caveat belongs up front, because it is the single most misquoted figure in this field. The association’s own sector page reports roughly one million employment relationships for 2024, against the roughly 700,000 employees cited for the years before. That looks like explosive growth — or, told the other way around by someone with an agenda, like a sudden mass of precarious jobs. It is neither. The association states plainly that the increase is attributable solely to a change in statistical recording: previously a person holding several jobs was counted once, now each employment relationship is counted separately.

The association goes one step further and says a valid comparison of the employment statistics with the years before 2024 is no longer possible. That is an unusually honest thing for a trade body to publish about its own headline number, and it is the reason we anchor this article on the trade-census figures rather than on the newer series. If you see „eine Million Beschäftigte in der Gebäudereinigung" quoted anywhere as evidence of anything, the number is being used wrongly.

This matters beyond pedantry. Staffing decisions — and robot purchases — get justified with sector statistics in board papers. A figure that moved because of a counting rule should never end up as the premise of an investment case. Where the sources give a spread, we state the spread: roughly 658,000 to just under 700,000 employees on the trade-census basis through 2023, roughly one million employment relationships on the new basis from 2024, and no clean bridge between the two.

The core constraint is not pay and not prestige. It is the clock. Cleaning happens when the customer wants it to happen, which for most contracts means before the building fills up or after it empties. The association describes this itself: unlike in Scandinavian countries, where cleaning is often organised so that most people notice the service, in Germany it is frequently done by part-time staff in the early morning or late evening hours — and the part-time ratio is correspondingly high. A job that offers two hours at half past five in the morning and two more at eight in the evening is not a job most people can build a life around, however it is paid.

The employment structure shows the same fragmentation, although it has improved markedly. Per the association’s report, between 2014 and 2023 the number of employees subject to social insurance in the sector grew by 23.7 percent while the number of minijobs fell by 24.4 percent; the minijob share of employment dropped from 40.2 percent to 29.1 percent, and the share subject to social insurance rose from 59.8 percent to 70.9 percent. That is a real and deliberate structural improvement, driven by binding sector minimum wages that have been declared generally applicable. It also means that nearly three in ten employment relationships in 2023 were still minijobs — a lot of small, scattered shifts to coordinate, and a lot of doors through which people leave quietly.

The sector is also, by its own description, an entry gate into the German labour market, and the figures are stark. Per the report, citing Federal Employment Agency data for June 2024, 57.5 percent of employees subject to social insurance in building cleaning had no vocational qualification or an unknown one, against 21.2 percent across the economy; 47.4 percent held no German passport, against 16.0 percent economy-wide. That integrative role is a genuine strength and the association treats it as one. Operationally it also means continuous onboarding, language support and instruction effort — for a workforce that competes for the same people as logistics, gastronomy and retail.

And then there is the piece operators cannot control: cost structure. The association puts the labour cost share in commercial cleaning at around 85 percent on average, against 31 percent for other business services per the Federal Statistical Office. When 85 cents of every euro is wages, there is no cushion. An unfilled shift is not an inconvenience to be absorbed somewhere else in the P&L — it is the service not happening.

The levers operators actually pull are unglamorous and mostly about the clock. Daytime cleaning is the biggest one, and the association argues for it explicitly on recruiting grounds: considerably more people would be interested in a job in building cleaning if contiguous, family-compatible working hours during the day were more widely possible. Beyond that: consolidating scattered shifts into blocks, training and qualification paths, and holding on to experienced staff rather than replacing them. The binding sector minimum wage, declared generally applicable, sets a floor under all of it. But the association also names the limit of this lever honestly — in the end, the customer decides how much is cleaned and when.

This is where robots have a real and narrow role, and it is worth being precise about it, because the marketing around this topic is bad. A cleaning robot does not fill a shift. It removes work from a shift — specifically the most repetitive, least skill-dependent share: large, open, obstacle-poor floor. That is the part of the job that scales with square metres rather than with judgement, and it is the part that makes a two-hour early shift feel like an unrewarding grind.

The best evidence for the limits of this comes from the industry’s own report, citing the Job-Futuromat of the Federal Employment Agency: only 3 of 10 tasks in commercial cleaning are automatable — where for a retail salesperson, for example, 7 of 8 core tasks are. Three in ten is a meaningful share of a workload, and it is nowhere near a job. The report’s own conclusion is that digitalisation will increasingly accompany and support people in building cleaning, but the sector will remain a „people’s business" for the foreseeable future. We would sign that sentence.

So the honest framing is a redistribution of scarce hours, not a headcount cut. The robot runs the open floor; the people do the work that actually needs people — edges and corners, sanitary areas, surfaces and touchpoints, glass, exception handling and quality control. If you are working out which machine fits which of those situations, our guide to choosing a cleaning robot walks through it by venue type, and the piece on what drives the cost of a cleaning robot covers the factors that decide whether a deployment carries itself.

In offices, the shortage shows up as the early shift nobody wants and the evening shift that keeps turning over. Offices are also the easiest case for machine support, because the building is genuinely empty at night: a vacuum robot such as the NR1700 can work the open floor and circulation areas without disturbing anyone. Two details matter in practice. Vacuum robots are loud — which is a non-issue in an empty office and a hard constraint anywhere people sleep. And the NR1700 navigates without any camera, which shortens the conversation with the works council and the data protection officer considerably.

In hotels, the same shortage is sharper because the building is never fully empty and the quality bar is visible to guests. Here the split runs by area rather than by hour. Hard floors in the lobby, restaurant and back-of-house are a good fit for a scrubber such as the CC1, run in off-hours — with the caveat that a scrubber leaves a thin film that needs a few minutes to dry, so the route has to be planned around guest traffic rather than through it. Corridor vacuuming on guest floors is the opposite case: never at night next to occupied rooms, which is precisely the kind of thing that pushes hotels toward daytime cleaning anyway.

In warehouses, the arithmetic is the most favourable of the three, because the ratio of open floor to detail work is the highest. This is largely dry work — sweeping rather than scrubbing — and the honest failure mode is not the machine but the site: pallets in aisles, wrap and strapping on the floor, and traffic that changes hourly. A robot handles a clean, predictable main aisle very well and a congested one badly. Our comparison of sweeping robots goes into which machine suits which floor and debris profile.

A robot is not a person and it is also not an appliance you switch on and forget. It needs water filled and waste water emptied, brushes and squeegees changed, filters and bags handled, maps maintained when the layout changes, and someone to walk over when it stops in front of something it does not understand. Every one of those is an hour, and if that hour lands on the same scarce early-shift person the machine was supposed to relieve, the deployment has moved the problem rather than solved it.

That is the whole reason we rent our machines as a full-service model rather than selling them: maintenance, wear parts, remote monitoring, exception handling and mapping changes sit with our service, not with your team. It is not a moral position — it is the only structure in which a robot reliably produces net relief for an operation that is already short of hours. A machine that needs a competent operator to be productive is a machine that fails in exactly the situation it was bought for.

The rest of the honest list: robots do not do sanitary areas, they do not do surfaces or touchpoints, they do not do glass, they do not do the judgement call about what is dirty enough to matter, and they do not do the reassuring presence that a guest or an employee reads as „someone is looking after this building". They handle the flat, open, repetitive share so that scarce people spend their hours on the rest. That is a smaller promise than the category usually makes, and it is one that survives contact with a real shift plan.