A poorly run tender produces three proposals for three different projects and a decision made on which supplier seemed nicest. That is not a criticism of the decision — with incomparable information, instinct is a reasonable fallback.

The fix is in how the process is set up.

Specify requirements, not solutions

The single most important principle.

Describe what the system must do, what constraints apply, and what success looks like. Do not describe how it should be built, which platform to use, or what the screens should contain.

Specifying the solution removes the expertise you are buying. It also means every supplier proposes essentially the same thing, so you are comparing price alone and have learnt nothing about who is any good.

Constraints are legitimate and should be stated: systems it must integrate with, data that must remain in a particular place, a platform your team already supports, a hard deadline. State the constraint and the reason for it, so a supplier can tell you if it is unnecessary.

What the document should contain

  • What your business does, briefly, and who the users are
  • The problem being solved, in business terms
  • Requirements, separated into essential and desirable
  • Existing systems and any integration required
  • Volumes — records, users, transactions — so effort can be estimated
  • Constraints, with reasons
  • Budget range
  • Timescale and any fixed dates
  • How proposals will be evaluated, with weightings
  • The process and dates: questions by, proposals by, decision by

The budget range is the item most often omitted and the one that most improves proposals. Without it, suppliers guess, and you receive responses at three different scales that cannot be compared.

Publishing the evaluation criteria is also worth doing. It produces proposals that address what you care about, rather than what suppliers assume you care about.

Get the cost structure right

Require costs in a fixed structure so they can be compared:

One-off costs — build or licence, implementation, data migration, integration, training. Recurring costs — licences or hosting, support, maintenance — stated per year for five years. Day rates for work outside scope. And an explicit statement of what is excluded.

That last item is where proposals genuinely differ. A cheaper proposal excluding data migration and training is not cheaper.

How many suppliers

Three to five. Fewer gives no real comparison; more wastes effort, and the best suppliers decline processes where the odds are poor and the specification is vague.

Do a light shortlisting stage first — a short questionnaire about relevant experience and capacity — rather than asking ten suppliers for full proposals. It respects their time and improves the quality of what you receive.

Good suppliers choose their tenders. A vague specification with ten invitees will be declined by exactly the people you most wanted to hear from.

Ask the questions that reveal something

Feature lists do not distinguish suppliers, because everybody ticks every box. These do:

Who specifically would do this work? Named people, with their relevant experience. Agencies sometimes sell with a senior team and deliver with a junior one.

Tell us about a similar project that went wrong, and what happened. The most revealing question you can ask. Suppliers with real experience have one; the answer shows how they behave when things are difficult.

What would you do if this ran four weeks late? You learn whether they have a process or a hope.

What do you need from us, and what happens if we are slow? Good suppliers are specific about client dependencies, because they have been caught by them.

What is the exit position? Who owns the code, how data is extracted, what happens on termination. See avoiding lock-in.

What have we asked for that you would advise against? A supplier willing to push back is worth more than one who agrees with everything.

Score it deliberately

Agree weightings before proposals arrive, not after. Typical shape: capability against requirements, relevant experience, approach and plan, the people, support arrangements, commercial terms, and cost.

Cost should be weighted but not dominant. A process where cost is sixty per cent selects the supplier who understood the least, because they scoped the least.

Score independently and then discuss. Where scorers differ sharply on a supplier, that discussion is where the useful thinking happens.

Take references properly

Not the reference the supplier chose to give you — or not only that one. Ask for a client whose project was difficult.

Questions worth asking a reference: did it deliver on time and on budget; what went wrong and how was it handled; how responsive are they now, after the project; and would you use them again for something larger?

That last question produces the most honest answer.

Afterwards

Tell unsuccessful suppliers promptly and give a reason. It costs ten minutes and it is why good suppliers bid for your work again.

And carry the proposal into the contract. The commitments made during the process — timescales, named people, what is included — should appear in the agreement rather than being replaced by standard terms.

Our consultancy team helps UK businesses run procurement processes, including ones we are not bidding in. Start a conversation.