How to Read a Tender Specification — 3 Questions

A bid manager with more than a handful of active files doesn’t read every tender specification that lands in the alerts. There isn’t time. Anyone who fully read every incoming specification would have no time left to write the bids that actually matter. The question isn’t how to read faster. The question is how you know which specifications to read, and once you read, what to look for.

The answer is a fixed set of three questions, always in the same order:

— Are we allowed to bid? — Are we able to bid? — Do we want to bid?

Skip them and you read a specification linearly and run out of attention before page thirty. Use them and you read with focus, knowing on every page what you’re looking for and why. Three yeses: on to the bid. One no along the way: stop and move to the next file.

How many pages, and what that means

A typical service specification runs to twenty pages. A construction specification with technical specifications, drawings, bills of quantities and the special specification document easily runs past a hundred. A complex IT tender with extensive functional specifications, similarly. It isn’t the page count that’s the bottleneck — it’s the attention. Every page you read costs focus you no longer have elsewhere.

Reading every incoming specification seriously is paid for in time taken from elsewhere — less room for the files that actually matter, or evenings that aren’t evenings. Filtering, then, isn’t administrative overhead but a productivity question. And filtering starts before you open the specification.

The Quickscan: before reading begins

Before you decide to work through a specification, there’s a filter ahead of it: the Quickscan. It summarises the tender, extracts the basic facts — object, contract value, deadline, award criteria — and gives you in five minutes enough to decide: do I add this tender to my workflow, or not?

The Quickscan doesn’t answer qualification questions. It answers one much rougher question: “Is this worth spending time on?”. That’s a filter at the level of a quick glance at an email in your inbox: keep, archive or delete. At TenderWolf, this function is included in every plan, free included.

Only when a tender turns green and lands in your workflow does the real work begin. And there the role shifts: in the workflow, the AI reads the specification for you and analyses it against the three qualification questions. You judge what it puts forward.

Are we allowed?

Are we allowed — formally? The answer sits in the first ten pages of most specifications: grounds for exclusion, selection criteria, quality system requirements, rules on reliance on third parties.

Concretely:

Grounds for exclusion — social and tax debts, criminal convictions, bankruptcy status, serious professional misconduct. Evidence via the ESPD.

Accreditations or categorical capacity requirements — for construction a specific class and category, for services professional qualifications or certifications.

Financial thresholds — a minimum turnover over recent financial years, sometimes tied to a specific activity.

References — minimum number of comparable projects, minimum amount per reference, defined period. A ground for exclusion often hides here: having to demonstrate a three-million reference when your largest project tops out at two-million-four-hundred is a NO, not a “we’ll sort it out with partners if we get in”.

In the TenderWolf workflow, AI reads this part for you. Exclusion grounds, selection requirements and capacity demands are laid out in structured form next to your company profile; the gaps are visible before the specification is open. The final judgement stays with you — a recently completed reference isn’t always already in the profile, and director roles in related companies can count in ways no tool spontaneously sees.

A NO here is a real ending. What comes after no longer matters.

Are we able?

Can we execute it — if we win? The question shifts to the technical specifications, the bills of quantities, the execution conditions and the part of the specification dealing with the project team.

What’s at stake?

— The object precisely. Not the summary, not the first paragraph of article 1, but the technical definition as it sits in the technical clauses and annexes.

— The bill of quantities, item by item, with the unit description set against your internal calculation items. A unit mismatch (m² vs m³, linear metre vs piece) happens and costs money.

— Execution timeline, milestones, delivery and acceptance dates. Does it fit your planning, or does it fall in a window where you have three other live commitments?

— Required team capacity. How many key roles must we obligatorily put forward? What seniority, what experience, what certifications? Are those people free on the relevant dates?

— Place of performance and material obligations. Own equipment, certifications, product specifications, presence in a given region.

The question “can we do this on paper” isn’t the question. The question is: can we do this on time, with the right people, without putting other commitments at risk. A won tender that turns out to be unworkable isn’t a victory. It’s a damages claim, or in the worst case, an exclusion at a future tender.

The AI in the workflow extracts these elements — execution timeline, technical obligations, required capacities — and lays them out in a readable summary. Whether you can actually deliver, with the people and planning you have today, remains a human check.

Do we want to?

The third question is strategic, and therefore the hardest. You’re allowed, you’re able — do you want to?

What shifts in the reading posture?

— Award criteria and their weighting. A tender at eighty per cent price has a different DNA from a tender at sixty per cent quality. Are your strengths rewarded, or are you walking into an arena with a predictable outcome?

— The evaluation method. A linear price formula doesn’t read like a cut-off formula; a qualitative scoring method doesn’t read like a weighted sum.

— Execution conditions. Price-revision formula, payment terms, penalty clauses, surety bond, subcontracting, dispute clauses. A won contract under unacceptable execution conditions is a loss-making contract — and you don’t see that if you skip this step because the award criteria looked good.

— The contracting authority. Track record, payment behaviour, way of working on previous comparable contracts.

— Strategic value. Are you building a reference here? Testing a new market? Strengthening an existing relationship? Or are you signing mainly for the revenue, and is that enough?

For the first two questions, the answer is verifiable inside the specification. Here, you weigh. Bid managers who handle “do we want to?” loosely write files they’d rather not have written, six months later, halfway through execution.

The AI extracts here too — award criteria, weighting, execution conditions. The weighing itself stays human work. No tool knows whether you want to build a reference in this sector today, or just don’t anymore.

And only then: writing

Three yeses, and you start drafting the bid. Not a second sooner.

What comes after the three questions — the writing itself — is outside the scope of this article. TenderWolf will soon offer AI support for drafting bids, based on the specification and your internal knowledge base. But that’s another conversation, for another article.

What this changes in your working week

Bid managers who consistently outperform don’t write more files. They write fewer, but sharper. They start fewer tenders, but each file gets the attention it needs. That isn’t talent — it’s filtering.

The Quickscan filters coarsely. The three questions filter finely. What remains gets your best work. Anyone who masters this rhythm wins more often — not because they write faster, but because they no longer start files they shouldn’t have started.

Concretely. Build this habit in from Monday morning: new tender → Quickscan → decision to add to the workflow or not. In the workflow, the three questions, in order. Only with three yeses do you block calendar time for writing. Give it two weeks and watch what happens to your hit rate.

→ Create a free TenderWolf account — Quickscan, workflow with AI analysis on the three questions, and soon AI support for the writing itself.

→ How to Choose a Tender Platform — 7 Questions to Ask