Most travel programmes arrive at a tender the same way. Something is wrong, it has been wrong for a while, and the contract is coming up anyway. Going to market feels like the responsible thing to do, and often it is.
What is worth establishing first is which of your problems a different supplier could actually solve.
Why the incumbent starts ahead
This is not a conspiracy. It is arithmetic.
Your incumbent knows your programme. They have priced it before, so they know where the margin is and what they can afford to give away. They know which of your travellers complain and which of your markets are difficult. A challenger is pricing a business they have seen described in a document.
They also know what you are weighing up privately, which is whether you want to run an implementation. Changing TMC means data migration, profile rebuilds, new booking tools for thousands of people, and a service dip while everyone learns the account. That cost is real and it sits entirely on your side of the table.
And most RFPs produce similar answers, because most RFPs ask similar questions. When the written responses are hard to tell apart, the decision moves to price and to the relationship you already have.
None of that means you should not go to market. It means the market is not where your advantage lies. Your advantage is in knowing what to ask for.
What a tender is genuinely good at
Testing whether your commercial terms are still competitive. Finding out what capability now exists that did not when you last looked. Forcing an incumbent to engage with problems they have been managing around. Giving you a defensible process when somebody above you asks why you chose who you chose.
Those are real and they are worth having.
What it cannot touch
A tender replaces a supplier. It does not change your organisation.
If bookings are being made outside your programme because a business unit has an arrangement nobody has ever challenged, a new TMC will not fix that. If your card programme and your expense system and your TMC report three different totals, that is a definitional problem between Treasury, Finance and travel, and it survives any change of supplier intact. If your policy is clear but nobody enforces it, the policy will still be unenforced in January under a new logo.
These are the problems that come back eighteen months later, and because they come back after a tender, they get blamed on the tender.
Three questions worth answering first
What specifically is wrong, and how do you know? Not "service is poor". Which service, for whom, measured how, compared with what. A tender built on a feeling produces proposals built on the same feeling.
Which of these problems would survive a change of supplier? This is the one that saves you. Work it out now and you have a second workstream to run alongside the tender. Work it out in year two and you have an awkward conversation instead.
What did your last tender actually change? Not what was promised. What measurably changed, and is it still true.
If you are going to market anyway
Then the work is the specification.
Ask for evidence rather than assertion. A supplier telling you they provide excellent service is worth nothing. A supplier telling you the name and tenure of the people who would be on your account, and what their current clients' abandonment rates look like, is worth something.
Decide what you will measure afterwards before you start, and capture the baseline now while it is still uncontaminated. Otherwise the savings claimed in year one cannot be tested against anything, and neither can the problems.
And cap the number of questions. I built the platform that made it easy to ask two hundred of them, and the main thing I learned is that nobody reads two hundred answers properly, including the people scoring them.
Why I am the one saying this
In 2004 I built ETABID, the first online tendering platform for corporate travel. It did exactly what I built it to do. It made tendering faster, cheaper and far more transparent, and some of the largest companies in the world ran their sourcing through it.
What it did not do was ask whether the tender was aimed at the right problem. That was never the tool's job. It has become mine.