After four years of dissatisfaction the contract goes back out to tender. The new contractor is cheaper, makes a professional impression and starts in January. By March the situation is worse than before, by May the complaints are mounting, and in June somebody in committee asks whether the change was a good idea.
The cause may lie with the new contractor. In the cases we work through it more often lies in what went unresolved between the decision to terminate and the start of service: incomplete tender documents, missing handover data, an unsettled staffing question, an unfavourable start date. A change of contractor is a project, and it fails at the same points projects fail: too little lead time, unclear responsibilities, no handover.
The lead time
The lead time consists of two parts, both of which need time and which cannot be set off against each other.
The procurement procedure has deadlines of its own. In the open procedure the tender period is in principle at least 35 days, and to that come the standstill period after the award notification and the time for preparing, examining and evaluating. Shortening is possible under certain conditions, though not at will, and a faster tender does not solve a single operational handover problem.
After that comes mobilisation, and that is where the bottleneck sits. A contractor taking on a contract with fifteen full-time equivalents has to recruit or take over staff, register them, equip them, familiarise them and fit them into a roster. In a labour market where both sectors are recruiting, the hiring alone takes weeks. Fail to allow that time and on day one you get a team of floats and transfers from other buildings, and the deficit from that phase is carried along for months.
Plan backwards from the desired start of service, therefore: mobilisation time, the procurement procedure including the standstill period and any challenges, preparation of the documents, and before all that the analysis of causes. For a large, labour-intensive EU-wide procurement, experience puts you at nine to twelve months. That is a planning corridor for that case and not a general minimum; a small standard building manages with considerably less.
The date
A change on 1 January has fiscal appeal and several operational drawbacks: leave, public holidays, higher sickness levels, and in schools and administrative buildings little activity and therefore little feedback in the first weeks.
Rather than a calendar date, a short checklist is worth having. Are contacts on both sides reachable during the first weeks? Is staff available in the market? Is the building in normal operation, so that feedback arises? Can access be tested beforehand? Have special services such as a deep clean been completed before the start?
For schools that often means: mobilisation and deep cleaning during the summer holidays, with service starting shortly before term. Starting on the first day of term meets full operations with no run-up.
Transfer of undertaking
This is the legally and financially most important point of the change, and it is frequently neither mentioned nor settled in tenders.
The background in brief: simply losing a contract to a competitor is not a transfer of undertaking. A transfer occurs where an economic entity passes across while retaining its identity, and that is assessed in an overall appraisal. In activities essentially characterised by human labour rather than by significant operating assets, taking on a substantial part of the existing workforce in terms of number and skills may be decisive. Building cleaning is frequently classified in this way. For guarding contracts it is less clear-cut, because technical systems, connection to a receiving centre and organisation can carry considerable weight there.
And here lies the point we ourselves initially got wrong: a transfer of undertaking is not a decision anyone makes. It occurs by operation of law where the conditions are met. The existing employment relationships then pass to the new owner, the employees are to be informed in advance in text form of the date, the reason and the consequences, and they may object to the transfer. A client declaring in the tender documents that a transfer of undertaking is excluded binds neither the employees nor a labour court.
Whether a transfer applies in the particular case is a legal assessment, and we do not make it. What can be said is why the question has to be settled before the tender goes out: it determines every tenderer's costing. Leave it open and you receive tenders based on differing assumptions that are not comparable, and you learn of the problem when it is acute.
The practical recommendation therefore runs differently from what you might first think. Have the question assessed legally and disclose in the documents the facts on which everyone is to cost: the number of staff deployed, their functions, working time volume, length of service, qualifications and the key collective agreement data. Aggregated and without names, in a protected area of the procurement platform. Together with your organisation's legally assessed view of the risk. What you cannot give is a guarantee about a legal consequence you do not control.
Handing over the knowledge
An outgoing contractor owes the return of what the client made available to it, and little beyond that. What it takes with it in practice is considerably more.
Four things should therefore be settled in the original contract, though ownership is not always the right word. For the client's own data such as floor plans and key registers, it is a matter of surrender. For building-specific work product that the contractor owes under the contract, it is a matter of usage rights and data format. And in its general methods, its internal costing and its company-wide templates the client has no rights, even where these were used in the building. The old contract should therefore settle, for each category, what is handed over in what format, within what period and with what usage rights.
The room schedule or area listing. Without it you will next time be tendering on the basis of figures nobody can verify any more.
The cleaning or security concept. The determination of which surface is treated how, or which reporting routes apply, is building-specific and should stay with the building.
The key and access documentation. Complete, with a trail of who received what and when.
The history. When the last deep clean was, what damage is known, what maintenance is due, what incidents there have been. Care is needed on the last point: incident reports frequently contain names, health information or security-relevant detail about vulnerabilities. They have to be prepared for the purpose, reduced and handed over by role, rather than passed on as a folder.
Where these documents are missing, the new contractor starts with a stocktake lasting four to six weeks, which the client experiences during that time as teething trouble.
The condition on handover
Between the outgoing contractor's last day and the incoming contractor's first there is a moment at which the condition can be established. After that, not.
A documented record at that point protects everyone. It prevents damage the new contractor inherited from being attributed to it, and it equally prevents damage that genuinely arose later being dismissed by reference to the predecessor.
A joint walk-through with both contractors and the client is the cleanest form, and it is not always feasible. Where the change is contentious, in security areas, or where competitive information is touched, it does not work. A client-led walk-through then serves, with an identical record for both sides and the opportunity to comment separately.
On photographs, one point is easily overlooked. Pictures from offices catch screens, name plates, documents and people; in security areas they catch system details. Purpose, access, retention period and deletion have to be settled beforehand, and sensitive areas excluded or obscured.
One point is frequently underestimated here: recording the condition solely on the day of the change documents a snapshot and hands it to the successor as the starting position. A two-stage record is better. A picture of the condition some months before the contract ends, while normal operations are still running, and an update immediately at the changeover. The difference between the two is information in its own right, and in case of doubt it is worth more than any supposition about the outgoing contractor's conduct.
The overlap
An overlap phase, with both contractors in the building at the same time for a few days, secures knowledge of the building like almost nothing else. It is also the element that meets the most resistance, because it costs twice.
The extent need not be great. A few days with the new staff working alongside the existing ones replace weeks of searching and asking. How many days depends on the size of the building, the number of staff and the documentation available. The precondition is that the outgoing contractor cooperates, and that cooperation has to be paid for and agreed in the expiring contract. It cannot be enforced after the event.
The overlap is not unproblematic. Two employers in the same building mean unclear lines of instruction, duplicated access and costs on both sides, and where the change is contentious the model does not work in any case. Where it does not fit, other forms serve: a handover workshop, a documented joint walk of the building without parallel operation, time-limited professional support for individual roles, or simply better documents.
What users expect and what they should be told
The underestimated factor in any change is perception within the organisation. Somebody who does not know a change has taken place will attribute the first irregularity to the new contractor, even where things were exactly the same before.
A short note to all users, one or two weeks before the change, alters that surprisingly much. It should contain: from when the change takes effect, who the new contact is, what changes organisationally, and where feedback goes. What does not belong in it is any announcement that irregularities are to be expected. That shapes expectations negatively and additionally anticipates a deficiency that is not contractually owed.
The last point is the most important. Feedback running to one place in a bundle is usable. Feedback going individually to changing recipients creates atmosphere.
The last eight weeks: a checklist
Eight weeks before: schedule the record of the condition, take stock of the keys, request the handover of documents.
Six weeks before: name the contacts on both sides, apply for access rights, request the new contractor's mobilisation plan.
Four weeks before: schedule the periodic services for the first year, agree the overlap phase, follow up open points from the document handover.
Two weeks before: send the user note, test access, carry out the record of the condition.
First week: ensure availability, bundle the feedback. Critical and contractually relevant deviations are formally recorded from day one, unstaffed positions as much as uncleaned sensitive areas. What does not arise during this week is the unstructured collection of small individual observations.
Week eight: first full performance inspection, adjustment, transition to normal operation.
What is to be done in detail in the first eight weeks after the start is described separately.
The sentence that should come first
A change is usually decided because the service was not right. The expectation is then that a change will solve the problem.
In some cases it does. In others the cause was not the contractor but a costing that was too tight, an incomplete specification of services or a lack of steering within the client's own organisation. Fail to change those causes and you get the same result with a new contracting party, only a year later.
Establishing what it actually came down to therefore comes before the tender. It is less comfortable than a change of contractor and in many cases the cheaper solution.