A 300-page invitation to tender lands on a sales director's desk with a week to decide whether to bid. Bidding costs a bid team's week and an engineer's day; not bidding costs the opportunity. The decision usually rests on a sales engineer's skim and a gut feeling about the product range.
The question that matters
Not "can we comply with most of it" — the answer is always yes. The question is: how many mandatory requirements have no compliance path, and can engineering close them in time? Four blocking items that engineering can close in 48 hours is a bid. Four that need a new certification is not.
A coarse matrix is enough
A go/no-go matrix does not need every citation polished. It needs the register extracted, every requirement classified, mandatory ones identified from the instructions to tenderers, and a status assigned at speed from the library. The output is a summary: 412 requirements, 208 mandatory, 201 comply, 3 partial, 4 open — and the four open items listed with their closest evidence.
That takes hours, not a week, and it costs one tender's worth of metering whether or not the bid proceeds. If it proceeds, the same register and matrix carry forward; nothing is redone.
What changes
Sales stops bidding on packages the product cannot win and stops passing on packages it could. Engineering sees the four questions on day one instead of a vague "can you look at this". And the win/loss tagging at the end feeds the next screen: the agent learns which requirement types this company tends to lose on, and flags them earlier.
Frameworks
For framework tenders the coarse matrix becomes the baseline. Each call-off is a diff against it — only new or changed requirements need an answer — which turns five years of call-offs into a few afternoons.