Close the loop from won and lost deals back to positioning

01 Anchor the loop on a decision, not a deal-review ritual.

Before you pull a single transcript, decide which three downstream calls this loop has to inform in the next quarter. Typical examples: which loss reason to fix in positioning, which competitor to re-arm sellers against, which capability gap to escalate to product. Win/loss without a decision attached is data collection. Diagnose before executing, refuse the playbook ask

Win/loss without continuous synthesis stays ad hoc. The deals keep closing, the disposition codes keep getting typed, and nobody pulls the thread that connects last month's losses to this month's positioning. State the three decisions at the top of the synthesis doc and check the finished output against them. If a theme does not connect to one of the three, park it. It is real, but it is not this quarter's work.

02 Treat CRM disposition codes as a roster, not an answer.

Pull every in-scope closed-lost and closed-won deal from the CRM. The codes tell you who you lost to and how often. They do not tell you why, and the why is the entire point.

A disposition code is a guess a rep typed in thirty seconds while closing the record. "Lost to Competitor X" often hides a status-quo loss, a budget freeze, or a champion who left. Use the codes to size the cohort and to weight your interview sample by ARR. Do not use them as the finding. Frontline customer contact is the PMM substrate

Calling every lost deal is capacity-bound. Sample smartly: weight by deal size, cover both won and lost, and prioritize recent closes where the buyer still remembers the decision. Surveys are not a substitute. A survey averages out the one specific sentence that tells you what actually broke.

03 Run the closed-lost interview from a 7-question bank, one primary per theme.

The interview is a JTBD switch interview pointed at a deal you lost. You are tracing the timeline backward from the decision to find the language the buyer used before they knew your product existed, and the moment the deal turned. JTBD interviews surface the customer's actual language and the switch trigger

"JTBD interviews help you define a customer's language — what they actually mean by 'easy' — the root cause of why they switched, and how their story connects, all of which directly impact your product marketing."

· Bob Moesta, Intercom podcast, 2020 · “JTBD interviews surface the…”

Keep to seven questions, one primary per theme, so the call stays a conversation and not an interrogation:

The same discipline that governs a switch interview governs this one. Stay open-ended and resist the urge to ask about features. Let the buyer narrate their own path. Use the JTBD switch interview on recently-converted high-value customers, segments emerge from jobs, not demographics

"The cardinal sin is asking about specific features rather than letting customers describe their own journey."

· Claire Suellentrop (synthesized from operator's published work) · “Use the JTBD switch…”

04 Separate the no-decision loss from the competitive loss before you do anything else.

The single most expensive mistake in win/loss is logging a stall as a competitor win. They are different losses with different remedies, and conflating them sends the work in the wrong direction.

When you survey buyers who passed on you, a large share never bought anything from anyone. 40–60% of B2B buyers say "no decision", your real competitor is the status quo

"When you survey buyers who didn't buy, 40–60% say they made no purchase decision at all."

· April Dunford, Lenny's Podcast, 2026-04-28 · “40–60% of B2B buyers…”

Dixon's machine-learning analysis of 2.5 million recorded sales conversations lands on the same range from a different direction, and adds a second number that should change how you read every stalled deal. 40-60% of B2B deals are lost to "no decision", and 87% of deals show medium-to-high indecision

"40-60% of deals are lost to \"no decision\" and that 87% of deals show medium to high levels of indecision."

· Matt Dixon, The JOLT Effect, 2022 · “40-60% of B2B deals…”

So in interview question 7, listen for whether the buyer made a decision at all. "We decided to wait" and "we decided to handle it in-house" are status-quo losses, not competitive ones. The remedy for a status-quo loss is narrative and problem framing. The remedy for a competitive loss is a comparison and a sharper differentiator. Most teams default to battlecards because battlecards are legible. Run the loss-reason audit before choosing where the time goes. No-decision-as-competitor vs. battle-card-driven competitive workflows Status quo / no-decision is the real competitor

One trap to flag for sales, because the fix is theirs, not yours. When a deal is stalling on indecision, the instinct is to hammer the cost of inaction harder. That works on a buyer who is comfortable with the status quo. It backfires on a buyer who is already afraid of getting the decision wrong. Sell to the buyer's mindset, not to product features

"When faced with indecision, 73% of reps return to hammering the status quo, but this backfires 84% of the time, making buyers more indecisive."

· Matt Dixon, The JOLT Effect, 2022 · “When buyers are indecisive,…”

05 Cluster the lost reasons into six gap types, each routed to a named owner.

Theme extraction is where raw interviews become decisions. Cluster every closed-lost reason into one of six gap types. Each type routes to a different function, which is the whole reason to separate them.

Run extraction monthly, not when someone asks for it. A theme matters only if it updates something. The positioning gap and the competitive gap look similar in the transcript and demand opposite responses: one rewrites the story, the other arms the seller for a comparison they can already win. Status quo / no-decision is the real competitor

06 Pattern-match every won deal against the best-deal cohort.

Sampling skewed to closed-lost only throws away half the signal. Won deals tell you what is working and which stories to reinforce, and they surface the customer evidence your case studies need.

For each won deal, ask what it shares with your fastest, stickiest, highest-expansion customers. The deals that match the best-deal cohort are your narrative-reinforcement signal: the messaging that converted them is the messaging to push harder. The ones that closed for reasons you did not expect are the more interesting find, because they hint at a buyer type or use case the current positioning does not name. Flag any won deal that fits your allowed-claims rules as a case-study candidate, and never pull a public case study from a transcript without written customer approval.

When the market itself breaks, the won-and-lost loop flips into the primary product-research input. Shruti Kapoor ran Wingman through exactly this during the 2022 downturn. Switch from selling mode to listening mode when the market breaks

"We realized, we cannot be out there necessarily selling right now. What we need to be doing is to be much more receptive to what customers are saying, versus trying to sell them on our product. So we changed gears into listening mode to say, ‘Let's use this time to just go back and understand how people are using our product.’"

· Shruti Kapoor on Clari blog, 2022 · “Switch from selling mode…”

07 Name the real alternative, not the one the battlecard names.

The competitive cluster of your loss themes should be checked against one question: are sellers losing to the competitor the battlecard names, or to a different alternative the buyer actually considered? The right input for positioning is what sales is losing to this quarter, not the analyst-defined market map. Stop positioning against ghost competitors customers don't actually consider

When a win/loss review surfaces a competitor the marketing battlecard does not name, that is a finding, not noise. Reps narrate the deal in language the website does not use. The alternative they are winning or losing against rotates faster than the site updates. Frontline customer contact is the PMM substrate

"Your sales team knows months before anyone else when a position is failing."

· April Dunford, LinkedIn (April 2026) · “The sales team detects…”

For competitive-displacement programs, a competitor being in the account is not the same as the account being ready to switch. Readiness shows up only when several signals line up at once: install maturity and renewal proximity, spend and expansion patterns, where the product sits in the org, adjacent technology changes, and segment-level switching behavior. A list of every company using Vendor X is a starting roster, not a target list. True switching readiness requires five aligned signals, not just competitor usage.

08 Turn the competitive theme into a battlecard that updates on event, not on a calendar.

The competitive gap themes feed battlecards. The failure mode is not authoring. It is adoption, and adoption is a distribution problem. Battlecard adoption fails on distribution, not content quality, embed in the workflow or the card dies

"68% of sales opportunities are competitive and nearly two-thirds of CI teams produce battlecards."

· Crayon (synthesized from operator's published work) · “Battlecard adoption fails on…”

A card that lives in a doc nobody opens is the same as no card. Embed it in the CRM at the opportunity level when the competitor is tagged, surface it in the deal-stage channel, and validate the content with top-performing sellers before distribution. Keep it accurate, brief, and consistent in format across competitors. One wrong claim ends seller trust in everything you publish.

The deeper shift is from authoring a static document to curating the inputs. When the win/loss interviews, competitor product updates, and field reports update, the card updates. Battle cards become workflow primitives, not Notion pages

"AI can transform static battle cards into dynamic tools that provide real-time competitive insights directly within seller workflows."

· Gartner, Innovation Insight: Rethinking Battle Cards in the Age of AI (G00832921), 2025-06-19 · “Battle cards become workflow…”

That means the win/loss loop is the feeder. Your job is to keep the inputs current and verified, then let the card pull from them. A battlecard updated on event beats one rewritten on a quarterly calendar.

09 Emit each theme as a signal to the function that owns the fix.

Synthesis is not done when the themes doc is written. It is done when each theme has left the doc and reached an owner. For the quarter, propagate:

• Positioning gaps to the positioning statement and homepage, with the verbatim buyer language attached.
• Competitive gaps to the battlecard inputs, traced to the specific lost deals.
• Capability gaps to product, ranked by how often they appear in loss reasons.
• Price gaps to pricing, with the deal references that show the mismatch.
• Won-deal patterns to the narrative and the case-study pipeline.

Trace every downstream change to the deals that justified it. "We repositioned against the in-house build because it drove four of this quarter's eight enterprise losses" is a propagated theme. "We should improve messaging" is a note nobody will act on.

10 Keep a human verifying every theme before it ships.

AI handles the at-scale parts of this loop well: transcribing interviews, routing replies, clustering reasons, drafting theme summaries. It does not handle the judgment about which loss is a status-quo loss and which is competitive, and it will fabricate a clean theme when you run it across thin or contradictory transcripts without a checkpoint. Verification, not execution, is the irreplaceable human job

Customer-marketing listening programs show the same division of labor. The system scales the routing and the pattern extraction; the human writes the question and decides which signals are worth acting on. The highest-value customer marketing motion is a systematic listening program, not a campaign calendar

Before any theme enters the synthesis doc, check the generated summary against the source transcript. No invented quotes. Every quote in the themes file traces to a recorded interview or call. The verification is the load-bearing step, not the generation.