The annual proposal-team retro template
The questions we ask at year-end and the format that surfaces what monthly retros miss. A 90-minute meeting that produces a durable artifact instead of a slide deck that gets archived and forgotten.
A monthly retro answers a month’s worth of questions. An annual retro answers a different set. The patterns that only become visible across a full year — themes that drifted, verticals that shifted, SME availability that deteriorated, templates that aged out — do not surface in a 30-minute monthly. They need their own meeting, their own format, and their own artifact.
This is the template we use and the one we think most small-and-mid proposal shops should use. It runs 90 minutes. It produces a written document that lives in the KB, not a slide deck that gets archived to a drive and never re-opened.
The prerequisite: the record has to be clean
The retro is only as good as the data it runs on. If the win/loss tracker is out of sync, if the bids marked “pending” are actually lost, if the response time data is inconsistent, the retro produces noise that sounds like insight. The week-between ritual includes a reconciliation pass; run that first. Leulu’s broader framing on why post-mortems fail to produce action applies here — if the inputs are wrong, the outputs are wrong, and the team stops trusting the retro format.
The format: 90 minutes, five sections
Schedule it for the 29th or 30th. Not the 28th (people are still off), not the 31st (people are already out). Full team on video. A designated notetaker who is not the proposal manager — the manager should be able to participate, not transcribe.
The five sections, in order:
Section 1 — the year’s numbers (15 minutes)
Someone on the team — ideally the proposal manager — presents the year’s record in under 15 minutes. Total bids submitted, win rate overall, win rate by vertical, win rate by deal size, average response time, average team-hours per bid. No commentary yet. The numbers are the floor the rest of the meeting reasons from.
If the team cannot produce these numbers because the tracking is incomplete, that is itself the output of section 1, and the retro becomes a meeting about why tracking is incomplete. Lohfeld’s process diagnosis names this pattern: teams that cannot measure their process cannot improve it.
Section 2 — win themes audit (20 minutes)
Which themes ran in the year’s bids? Which correlated with wins? Which correlated with losses? Which were generic enough to swap competitor names with and have them still read? The end-of-year win themes audit is the longer-form version; for the retro, the 20-minute pass is a ranking exercise. Promote, keep, rewrite, retire. The team argues about each theme. The argument is the point — themes that the team cannot defend out loud are themes that will not defend themselves in a proposal.
Section 3 — the top three process frictions (25 minutes)
Each team member names the single process friction they hit most often in the year. Not “SMEs are slow” — everyone knows that. Specifically which SME, in which type of question, under what conditions. Specificity makes the friction actionable. “Security engineers do not respond to DDQ questions that land after Thursday” is actionable. “SMEs are slow” is not.
Rank the frictions by frequency and severity. The top three go on the action list. The rest go on the backlog. The team will not fix everything; pretending otherwise produces a work queue that stalls the whole retro output.
Section 4 — the year’s losses, examined (20 minutes)
Not all of them. Pick three — the one that hurt the most, the one that was the biggest surprise, and the one that feels most winnable in hindsight. Ten minutes on each is too short; the instruction is to go wide, not deep. What did we learn from the post-mortem (if one happened)? If a post-mortem did not happen, why not?
Losses without post-mortems are the largest single source of compounding failure in a proposal function. Most teams do not hold them. The retro’s job is not to run the missed post-mortems; it is to name the losses as a pattern and commit to a post-mortem discipline for year two.
Section 5 — the commitment list (10 minutes)
The last 10 minutes produce the artifact: a written list of commitments for the coming year. Each item has an owner, a first milestone in January, and a review checkpoint (typically the March monthly retro). Items without owners are not items. Items without a first milestone become intentions instead of commitments.
Typical commitment categories: one template refresh, one win-theme retirement, one SME-collaboration change, one tooling change, one measurement we will add to the dashboard. Five commitments, not fifteen. A year’s worth of proposal work cannot absorb fifteen major changes; it can absorb five.
The artifact
The output of the retro is a document in the KB. Not an email, not a slide deck, not a drive folder — a living document with a fixed URL that the team reads in March, in June, in September, and again in December when the next year’s retro starts.
The document has five sections mirroring the meeting: the numbers, the themes audit, the top frictions, the losses examined, and the commitments. The commitments section has checkboxes. The monthly retro each month reviews the commitments and updates them. An un-met commitment by March is a commitment that needed a different owner or a different scope; an un-met commitment by September is a commitment that probably was not going to be met and should be formally dropped rather than carried indefinitely.
What the retro does not try to do
It does not relitigate individual bids. There are monthly retros for that. It does not resolve conflicts between team members; those belong in a different setting. It does not produce a score for the proposal function’s performance. A score that is not tied to a compensation mechanism is a score the team argues about and does not act on.
It produces two things: a clean view of the year’s pattern and a short list of commitments that map to specific action in January and February. That is enough.
Why this is the meeting that compounds
Because the document it produces is readable. A slide deck presented in a conference room and then archived is a meeting that leaves no trace. A document in the KB, with checkboxes, updated across the year, is an artifact that outlives the meeting. The next year’s retro reads the previous year’s commitments before it starts. Commitments that got met inform the team’s confidence in its own process. Commitments that did not get met inform the team’s honesty about its own process.
The saturday retro ritual is the quick-cadence equivalent — a short-form, per-bid version. The annual retro is its long-cadence counterpart. The two together are the feedback loop a proposal function needs to actually improve rather than repeat. Run the annual one on the 29th or the 30th this year. Do not skip it because the calendar is short; the calendar will always be short.