Five requirements a procurement team should place on an integrator
Most tenders check technical capability and financial standing. They almost always omit the five points that determine what the contract actually costs.
An integration tender is rarely decided on the technical competence of the bidders. Everyone who reaches the shortlist knows how to mount cameras, pull fibre and install a generator. What separates them comes down to five points most specifications never address — and which nonetheless determine the cost of the contract over its real life.
1. Who answers for the interface defect
A programme split into four packages awarded to four firms produces four suppliers, each compliant with its own package. On the day access control fails to open the barrier because the controller never receives the signal, none of the four is in breach: each delivered what it undertook to deliver.
The requirement to set is simple: the contractor answers for the working of the whole, including the interfaces between packages, and that responsibility is written down before execution. It changes the price; more importantly, it changes how many meetings it takes to clear a snag.
2. What becomes of the documentation
An installation without operating documentation is an installation whose supplier and whose settings cannot be changed. This is not an acceptance detail: it is the variable that determines the cost of the second tender.
Requiring documentation is not enough. What it contains must be specified — the schematic, the actual configurations, the operating procedures, the administration credentials and their rotation procedure — and its verification by a third party or by internal teams must be scheduled before final acceptance.
3. The exit plan
The question to put to the contractor is not “how will you support us”, but “what happens if we replace you in three years”. The answer must cover four points: who owns the code produced, in what formats the data can be exported, how long the return takes, and what support is owed to the incoming supplier.
A bidder with no written answer to those four questions has never addressed them. That is not a moral failing, it is a budget risk.
4. What is measured, and how
Service commitments are readily written as annual availability percentages. Those percentages are hard to challenge and easy to announce. Three quantities are more useful: time from incident to detection, time from detection to first intervention, and recovery time observed during an exercise.
The difference is verifiability. An annual rate is established at year end; a detection time is established at the first incident, and it can be tested before the system even goes into service.
5. Who will operate it, and with what skills
A system handed to a team that cannot operate it produces two effects: permanent reliance on the supplier’s support, and a gradual drift in configuration as workarounds accumulate.
Skills transfer must appear in the schedule of prices as a separate item, with its volume, its audience and its acceptance criterion. Training whose effect no one verifies is not training, it is a billing line.
What these requirements produce
They filter out bidders who sell an installation and retain those who sell an operation. They raise the apparent price of the initial offer and lower the cost over five years. Above all, they move the negotiation off the ground of equipment — where comparison is easy and uninformative — and onto the ground of responsibilities, where the programme is actually won or lost.
