Why Good Proposals
Lose

The frustrating losses are not the ones where you were unqualified. They are the ones where you were compliant, the writing was good, the price was competitive — and you came third. Those losses almost always trace back to a mismatch between how you wrote and how the proposal was actually read.

How your proposal is really read

It helps to picture the room. An evaluation panel of five people. Each has a day job. They have been handed six proposals averaging 90 pages, a scoring sheet, and about ten days. Do the arithmetic: that is roughly forty minutes per proposal per evaluator, and some of that goes to the pricing volume.

Nobody is reading your document the way you wrote it. They are working through a scoring matrix, section by section, hunting for the specific thing each row asks about. The experience is closer to marking exam papers than to reading a pitch.

That third point is the one most teams underestimate. Public-sector evaluators, and an increasing share of private ones, must be able to defend every score. An evaluator who personally believes you are the strongest bidder, but cannot point to a sentence supporting it, will score you at the level they can defend. Your job is not only to be good — it is to be easy to award points to.

Six patterns break that, in rough order of how much they cost.

1. Answering the question you wish they had asked

The most expensive failure and the least visible from the inside. A question asks something narrow; the response addresses the surrounding topic, because that is where your strength lies.

The self-check is mechanical. Take each question, underline every noun and verb that names a deliverable — escalation path, named roles, target response times, outside business hours — and confirm each one appears explicitly in your answer. Anything unmatched is a point you are choosing not to earn.

2. Claims without proof attached

"Extensive experience", "industry-leading", "proven methodology", "deep expertise". Every competitor writes these, which makes them worth exactly zero as differentiators. Worse, they consume the space where evidence should be.

The discipline is to treat every claim as requiring a receipt in the same paragraph. Numbers, named clients, dates, certifications, audit results. If you cannot attach one, consider whether the claim is worth the words.

3. Making the evaluator work to find things

A proposal that reorganises the buyer's structure into your preferred narrative is asking a tired person to do reconciliation work at 8pm. Sometimes they do it. Often they score what they can find and move on.

Mirror the RFP's structure and numbering exactly, even when it is illogical. If they ask in the order 4.1, 4.2, 4.3, answer in that order with those numbers, using their terminology rather than your internal product names. If your platform module is called "Atlas" and the RFP says "reporting module", write "reporting module" — the evaluator is scanning for their own words.

Where the same evidence serves several questions, repeat it rather than cross-reference. "As described in Section 3.2" costs the evaluator a page-turn and a context switch; some will simply not make it. Redundancy is cheap, and a proposal is not a novel.

4. Answering the requirement but not the worry behind it

Requirements are written by people with a specific anxiety. A question about data migration is usually a scar from a migration that went badly. A question about staff continuity often follows a vendor whose good people rotated off after month three.

Answering the literal requirement earns the compliance point. Recognising the underlying concern is what separates the top two scores. If a question is unusually detailed relative to its neighbours, that detail is a tell: something went wrong there before. Address the mechanism and what happens when it fails.

5. Pricing that arrives without a story

On most evaluations price is a scored criterion, not a tiebreak — and cheapest rarely wins outright. What loses is price the evaluator cannot explain to a finance colleague.

A number 30% above the nearest competitor is survivable if the proposal makes the difference legible: what is included that others exclude, which risks you are absorbing, what the total cost of ownership looks like over the term. A number 30% below is genuinely dangerous — panels read a low outlier as a misunderstanding of scope, and price the delivery risk accordingly.

Either way, the fix is the same: state your assumptions explicitly, show what is in and out, and give the evaluator the sentence they will use when someone asks why the expensive bidder won.

6. Inconsistency across sections

Under deadline, sections get split across authors and never reconciled. The technical section proposes a four-person team; the pricing volume prices three. The methodology says two-week sprints; the timeline shows monthly milestones. One section says data stays in-region; the security annex mentions an overseas support centre.

Evaluators notice, because they are cross-checking to justify scores. The damage is disproportionate: an inconsistency does not just cost the point in question, it makes every other claim feel provisional. It is the cheapest failure to fix and the most commonly shipped, because it is the last check and the deadline always eats it.

Reconcile numbers, names, dates, and team sizes across volumes as a distinct pass with a named owner — not as something the proposal manager does while also formatting the appendices.

The pre-submission audit

Run this after the content is final and before formatting. It takes a couple of hours and catches most of the six.

The hard case: when you lost and cannot tell why

Debriefs are frequently useless. "Your submission was strong but another bidder offered better value" tells you nothing actionable, and in many jurisdictions the buyer is constrained in what they may disclose.

Ask narrower questions instead. Rather than "why did we lose", ask which criteria you scored below the average of the field on. Ask whether any answer was marked non-responsive. Ask whether the panel found the information they needed where they expected it. Those are usually answerable within procurement rules, and they point at a specific failure mode rather than a vague verdict.

Where scores are published — common in public procurement — plot your per-criterion scores across the last ten bids. Individual losses are noisy; the pattern is not. A team consistently mid-range on technical criteria and bottom-range on management approach has a specific, fixable problem, and it is almost never the one they assumed.

One caution on the other side: do not over-correct from a single loss. Sometimes the incumbent was always going to win, or the panel had a preference formed months earlier. Change your process on patterns across a set of bids, not on the most recent disappointment.

What actually moves the number

None of the six failure modes are writing-talent problems. They are attention problems. Every one of them is what happens when the deadline arrives before the reconciliation pass, when the SME had ninety minutes for a question that needed three hours, when nobody had time to check that answer 4.7 still matched answer 2.3.

Which is why the durable fix is structural rather than motivational. Get the extraction, the reuse of already-approved content, and the compliance tracking off your people's plates, and the hours come back for the work that changes scores: matching the question, attaching the proof, naming the worry, reconciling the whole.