The situation the advice does not describe
A firm of forty people wins work by answering formal tenders. There is no bid department. When a tender arrives, a partner decides to go for it, two or three technical staff are pulled off delivery to write the sections they know, somebody's assistant chases certificates, and the whole thing is assembled in the last four days by whoever has the file open. Everybody involved has another job, and the proposal is done in the time left over from it. This is not an immature version of a bid team. It is a different operating model, and it is how a large share of professional-services work is actually won.
You do not have a proposal team. You have an engineer who also writes the technical approach, on a Tuesday evening, because the deadline is Thursday.
What the standard advice assumes
Published guidance on bid team structure typically names something like six roles — a bid manager, a proposal manager, a proposal writer, a coordinator, a graphics or desktop-publishing owner, and a knowledgebase owner — and then discusses how to divide work between them. Capacity guidance assumes you can model each role's hours across the response period. Workload-balancing guidance assumes there is a workload to balance across people whose job this is. All of it is sound for organizations that have those people. For the firm above, the first instruction — appoint a bid manager — is already impossible, and everything downstream of it depends on that instruction having been followed.
- Our reading, not a rule
- The observation that mainstream bid-team guidance presupposes a dedicated function is ProposalOS's reading of the published material, not a claim any of those sources makes about itself.
Three responsibilities that have to be named
Not roles, and not job titles — responsibilities that must belong to a specific person for a specific tender, and be known to belong to them. Everything else in the standard model can be absorbed.
- Someone owns the requirement list. One person is responsible for the fact that every mandatory item and every scored criterion is written down and has a state. Not for answering them — for the list being true. This is the single highest-value appointment, and it takes a couple of hours a week rather than a job.
- Someone owns the submission mechanics. Portal registration, file formats, signatures, certificates that must be requested from third parties, the upload itself. This work has lead times nobody controls and it is invisible until it fails, which is why it must be somebody's job rather than everybody's assumption.
- Someone can say no. The authority to decline an opportunity has to sit with a named person who is willing to use it. Without that, every tender that arrives is answered, and capacity is set by what the market sends you.
What you can do without
A dedicated writer, in most cases. Technical staff who did the work usually write a more credible technical approach than a professional writer working from an interview, provided somebody edits for consistency at the end. A graphics owner, unless your sector scores presentation. A separate bid manager and proposal manager — at this scale they are one person, and splitting them adds a handover rather than removing one. What you cannot do without is the three above, and the common failure is having none of them because everyone assumed the partner was handling it.
The constraint is retrieval, not writing
It is worth being precise about where the hours actually go, because it changes what you fix. Very little of the effort in a small-firm proposal is spent composing original argument. Most of it is spent locating things: which past project matches, where the completion certificate is, whose CV is current, what the methodology said last time, whether that certification has expired, what the client asked at clarification stage two years ago. That is a retrieval problem, and it is the reason proposal effort does not fall as a firm's track record grows — the archive gets larger, so finding the right item in it gets harder, not easier.
Declining well is the only free capacity
You cannot add people, and you cannot make the deadline later. The one lever fully within your control is which opportunities you answer. A short, honest qualification step — run before senior time is committed, with the authority to end it — is worth more than any productivity improvement available to a firm this size, because the hours it saves are senior delivery hours. The decision should turn on what you can evidence rather than on whether the work is appealing.
Standardise these, leave the rest alone
Standardisation programmes fail at this size because they are scoped as if there were someone to run them. Keep it to the things that are reused on literally every submission and that go stale silently.
- Company facts — legal name, registration numbers, headcount, turnover, insurance. One approved record, refreshed at each financial close.
- Certifications and registrations — one place, each with its expiry date, checked against the date rather than on a calendar.
- Project references — a short structured record per project rather than a folder, with the proof attached and the client contact confirmed annually.
- Key CVs — current, in the formats your main clients demand, with the owner responsible for their own.
When two land in the same week
This is the moment the model breaks, and it breaks predictably. Two tenders with overlapping deadlines need the same two technical people, and the honest options are to decline one, to submit both at reduced quality, or to move a deadline you do not control. Decide which before you are in it. A firm that has agreed in advance that it will decline the second unless it clears a higher bar makes that call in an hour; a firm that has not will spend three days trying to do both and submit two weak responses.