Win loss analysis that counts every reason and shows the buyer saying it
Turn every loss into a specific fix, in the buyer's own words.
See one case, screen by screen ↓Win-Loss Analysis reads your bids, buyer debriefs, interviews, call notes and CRM records to explain why deals were won or lost. Seven agents group the reasons into themes with their deal counts and the buyers' own words, test each theme against CRM data and the deals you won, and check every quote. The VP Sales approves each finding with a reason code and signs before anything is shared.
How to choose AI solutions for back-office and enterprise work →
Updated
The reason in the CRM is one pick-list word, typed by the seller on the day the deal closed — usually “Price”. The real reasons sit where nobody reads them together: the proposal you submitted, the buyer's debrief (a formal score sheet in the public sector), interview notes, call notes and the CRM's own numbers. Without reading them together, nobody can count how many deals a reason really cost.
So the win-loss readout is built by hand. The product marketer chases interviews, codes notes and writes it up over weeks; it rests on whichever deals got an interview, and no reader can check a quote. Nobody tests “we lost on price” against the deals you won at a higher price — so the fix is often the wrong one, and it rarely gets an owner or a measure.
Estimated split of the six or so weeks a win-loss readout takes by hand, and with the solution.
Seven agents read every bid, debrief and interview, test each theme against CRM data and the deals you won, and check every quote; the VP Sales approves and signs.
Kestrel & Vane's Technology Services business sells KV Assure, licensed per seat, to hospitals and health systems. The first half of FY27 closed with 14 healthcare deals lost, $18.6M of contract value. Here is what happens when the VP Sales asks why, screen by screen, in the working solution.
The H1 analysis finished overnight. Dana's home screen shows the healthcare win rate at 55% — 17 of 31, down from 64% in H2 FY26 — $18.6M of contract value lost across 14 deals, and 4 of 14 CRM loss reasons the buyer contradicts. Needs you lists four items: approve the H1 findings, due 21 Oct; one quote the checker removed; 4 CRM corrections; 4 interview proposals.
“9 of 14 losses partly on price per seat — mostly shared workstations billed as named seats.”
She asks in plain words. The theme agent sets the scope — healthcare, KV Assure, April to September 2026, lost: 14 deals, $18.60M. The intake agent reads 14 bids and 11 debriefs, two of them formal public-sector debriefs read as score sheets, then 9 interviews with consent checked on each, 38 call notes and 14 CRM records. Each agent's step shows as it runs, through to her review.
“Only deals you are allowed to see are read.”
Four themes, every loss under at least one: price per seat, 9 of 14 deals; integration took too long, 5; no reference our size, 4; security review late, 3 — marked Emerging, below the 4-deal minimum. Under each column are the buyers' own words, with the role, the deal, the source and the date. Of 19 quotes, 18 are verified and 1 removed.
“Your rate per seat was fine. What killed it was paying a full seat for every shared workstation on the floor.” — Director of Clinical Engineering, St. Brennan Health System
She opens the Northgate University Health quote. It comes from the written post-award debriefing, requested on 4 June within 3 days of the award notice: “Offeror's price score: 22 of 30. The successful offeror's evaluated price was 18% lower over five years.” The passage sits in its context, marked Verbatim match and In context, with the $3.2M deal beside it.
Under the answer, each agent reports what it did and asks for a decision. The quote checker removed Fairhaven Health's “Honestly, price was what decided it”: the sentence goes on to say the four-month go-live came first. It suggests moving the quote to integration, which would then rest on 6 deals; Dana moves it. She asks whether the seat tier would have saved St. Brennan. The challenger: closer, not certain — about 21% lower, within 3% of the winner, and the 16-week plan was the other reason.
The quote checker can remove a quote, never rewrite one. A person decides where it belongs.
The CRM check agent puts numbers on the price theme: 18–24% above the winner per seat, median 23%, deal by deal — $41 against $33 a seat a month at St. Brennan. The challenger tests it against the 17 wins: we were dearer than the runner-up in 6, and 5 of those 6 offered the shared-device seat tier. Verdict: holds, refined. The fix becomes a seat tier, not a discount.
“The rate alone does not lose deals; the seat model does.”
The CRM says Price for 12 of 14 losses; buyers named price in 9. At Penrose Regional, Westbrook Senior Living and Thornbury Medical Alliance the buyer named the integration plan; at Oakmere Cancer Institute it was the security review, not Timing. Dana asks the agent to select the four and sends them. Each now sits with its account owner to confirm.
“Corrections go to the account owner; nothing is written to the CRM until they confirm.”
Security review late rests on 3 deals: questionnaires went back after 19, 23 and 26 days against 10-day deadlines. To confirm or drop it, the interview agent proposes security questions for the two pending interviews at Oakmere and Kingsreach, and two wins for contrast — Cedarline Health Network, where security cleared in 6 days, and Bellhaven University Hospital. The questions are drafted; each account owner approves outreach first.
The interview agent never contacts a buyer without the account owner's approval.
Review and approve lists each finding with the agents' recommendation. With the open quote question answered, she applies them: price per seat, integration (now 6 of 14) and references approved with A1, Evidence sufficient; security review late held with W1, Await interviews. Any finding can be changed, edited with a reason such as E2, Narrow to a segment, or rejected with one such as R1, Too few deals.
She picks the audiences — sales leadership and product and pricing with names shown, the proposal team and enablement with names hidden — confirms she has read the findings and the quotes behind them, and signs. Held and rejected findings are not shared. Each approved finding becomes a fix with an owner and a measure: a shared-device seat tier (Leo Marchetti, 30 Nov), an 8-week integration plan with a named engineer (Hannah Brooks, 13 Nov), a reference bench of three systems over 800 beds (Tomasz Wilk, 15 Dec).
“Pilot since 1 Sep: 3 won, 1 lost, 2 open — 4 decided, too few to call.”
The same half-year of lost deals, seen by the four people who act on it — what their quarter looked like, and what it looks like now.
Seven agents read every bid, debrief and interview, test each theme against CRM data and the deals you won, and check every quote; the VP Sales approves and signs.
Gathers every source for each closed deal — the bid, the debrief, interviews, call notes and the CRM record — reads it, links it to the deal and checks consent. A password-protected file is not dropped: it stops and drafts a request for a person to send.
Sees which themes rest on too few deals and proposes exactly whom to interview, wins included for contrast; drafts the questions and runs async AI-led interviews with follow-up questions once the buyer consents.
Reads all the sources, groups the reasons into themes from your codebook, counts the deals behind each and picks the buyers' quotes.
Tests each theme against the numbers — price per seat, seat counts, plan weeks, reference sizes, questionnaire dates — and compares every closed-lost reason with the buyer's words, proposing corrections.
Confirms every quote is word for word, in context and allowed for each audience. A quote out of context is removed with the full passage shown.
Argues against every theme: looks for the counter-case in the deals you won and in the CRM data, and marks each theme holds, refined or weak.
Drafts the readout for each audience, proposes a fix for each finding with an owner and a measure, and tracks results with the number of deals they rest on.
Approves each finding with a reason code and signs before it is shared. The agents propose; a named person decides.
Ask about any theme, deal, quote or competitor — or tell the agents what to do. Every answer carries its sources and can open the screen it is about.
Summarise H1 in three sentences for the CRO
Healthcare win rate fell from 64% to 55% in H1; we lost 14 deals worth $18.6M. The biggest reason is how we count seats, not our rate: 9 losses, 7 of them over shared workstations. The fixes are a shared-device tier, an 8-week integration plan and a larger reference bench — the findings need approval before they go out.
Is the price problem getting worse?
Yes. Price-per-seat losses went from 2 of 6 losses in Q3 FY26 to 5 of 8 in Q2 FY27, while the win rate fell every quarter.
Would the seat tier have saved St. Brennan?
Closer, not certain. St. Brennan has about 600 shared workstations; billing them as devices at the tier rate takes the bid about 21% lower, within 3% of the winner. Integration was the other reason there: our plan said 16 weeks.
Which losses have no interview yet?
Five. Ashgrove, Calloway and Larkspur declined; Oakmere and Kingsreach are pending. Three of the five sit under price per seat, which already rests on 6 interviews and debriefs, so the gap that matters is security review late.
13 screens from the working solution, on its sample data. Pick one to see it large.
How themes are judged, who sees buyer names and the clocks the work runs to are settings. Threshold changes are proposed, approved and apply to the next analysis.
| Setting | Default | Choose from |
|---|---|---|
| Minimum deals for a finding | 4 | Any number; below it a theme shows as Emerging |
| Minimum decided deals before a fix result is reported | 10 | Any number; fewer is reported as too few to call |
| Quotes verified before sharing | 100% | Fixed |
| Default when a buyer gives no choice | Role only | Role only · Not quoted |
| Keep recordings for | 24 months | 24 months · 12 months |
| Buyer names shown to | Sales leadership, Product and pricing | Sales leadership, Product and pricing · Sales leadership only |
| Request a public debrief within | 3 days of the award notice | Set in days; the default follows FAR 15.506 |
| Approve a half-year readout within | 15 business days of close | Set in business days, with an approver reminder after 2 days |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (a 2026 RFP benchmark survey of 1,500+ teams, run by an RFP software vendor — 39% average win rate, 166 RFPs a year, 33 hours each; FAR 15.506 on post-award debriefings). People, companies, competitors and deals named on this page, including Kestrel & Vane Group, are fictional — characters and sample data in the working solution.
Win-Loss Analysis reads your bids, buyer debriefs, interviews, call notes and CRM records to explain why deals were won or lost. Seven agents group the reasons into themes with their deal counts and the buyers' own words, test each theme against CRM data and the deals you won, and check every quote. The VP Sales approves each finding with a reason code and signs before anything is shared.
It reads every source for every deal in scope, not only the deals that got an interview, so each theme shows how many deals it rests on. A challenger agent argues against each theme using the deals you won — in the sample, “price” became “shared workstations billed as named seats”, which changes the fix from a discount to a seat tier.
Yes. The CRM check agent compares every closed-lost reason with what the buyer said and cites both. In the sample half-year, 4 of 14 reasons were contradicted. Corrections go to the account owner, and nothing is written to the CRM until they confirm.
Yes. Formal written debriefings and assessment summaries are read as score sheets, so “we lost on price” is backed by the evaluator's own marks — in the sample, 22 of 30 on price, with the winner's evaluated price 18% lower over five years. Only what the buyer disclosed about you is kept, and a clock prompts the debrief request within 3 days of an award notice.
Where a theme rests on too few deals, the interview agent proposes whom to interview, drafts the questions and can run an async AI-led interview with follow-up questions once the buyer consents; human-led interviews are read the same way. No buyer is contacted until the account owner approves, and buyers choose whether they are quoted by name, by role only or not at all.
The quote checker matches every quote word for word to its passage and reads it in context. A quote out of context is removed, never rewritten, with the full passage shown, and a person decides whether to move it, keep it removed or restore it.
Yes. Agents never contact a buyer, write to the CRM, rewrite a quote or share a readout. The VP Sales approves, edits, rejects or holds each finding with a reason code and signs; buyer names are hidden for the audiences you choose; and every step is on an audit trail that cannot be edited.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your loss-reason pick-list, a sample of past bids and debriefs, your CRM fields and consent rules — and it goes live once every quality gate has passed. We will walk you through it on your own closed deals first.
Questionnaires back in days, not weeks, and every answer matches what customers were told.
1-min demoRFP Response AutomationHigher win rate, no promise you can't keep, and the bid in on time.
1-min demoDue Diligence Questionnaire (DDQ) ResponseEvery DDQ answered from approved sources, and no investor told two different stories.
1-min demoWe'll run Win-Loss Analysis on a sample of your own closed deals, bids and debriefs.