# Round 5 Synthesis - Segmentation and Firm WTP: Every Yes Came With an Escape Hatch (8 interviews + 4 carry-forward closes, simulated) *Study lead synthesis, 2026-08-07. Simulated personas: every number below is a directional prior, not validation. Round 5 of max 5. CONVERGED - see final call.* ## Headline finding The firm asks worked exactly as designed and the answer is now unambiguous: five rounds of real pain, real dollar losses, and persona-authored product specs converted into zero signed commitments, zero cards, and one forwarded factory list. Every yes arrived with a commitment-deferral mechanism the persona named precisely and unapologetically: Dana says yes at her own $500 number but wants the engagement letter before a card ("send me the actual engagement letter first"); Kelly says yes at $25/month but "I'm not giving you a card number for something that doesn't exist yet"; Derek agrees $2,500 is the right price and counters "ask me again at month nine"; Priya's yes becomes a tracker note that fires when a finalist exists, which is not this quarter; Tom's "easiest yes in five conversations" collapses to a $50-75 trial the moment the cold-start truth is spoken aloud. The second finding is the one that decides whether anyone should build this: the personas ran the segmentation arithmetic themselves, out loud, and every single account bounds at beer money. Dana $1,500/year. Priya $550-1,100. Tom $400-600 ("it doesn't sound like a company, it sounds like a hobby"). Kelly $300. Jenny $600. Marcus $4-5k in a theoretical trust-established year against $0 of his actual 2026 volume. Derek one purchase, maybe, someday. Aisha zero times six thousand. The only unconditional-feeling yes in the round, Jenny's $600/year, came attached to her own proof that the buildable product would not have prevented the $4,050 loss that made her want it. The study's five-round verdict: the pain is real, the products are correctly specified because the buyers specified them, and the market, as simulated, prices all of it as substitution or rounding-error insurance with a circular trust gate in front of the first sale. ## Scorecard | # | Persona | Wedge tested | Fit | Pain | Firm-ask result | New $ this round | 12-mo account value (their own math) | |---|---|---|---|---|---|---|---| | 01 | Dana, DTC hardware sourcing lead | W1 network | strong | 4/5 | Yes at $500/candidate, conditional on escrow + engagement letter; no card today; cold-start line held | none (Herrera passed QC; subcontractor permit risk surfaced by luck, not product) | ~$1,500 (3 non-China events) | | 02 | Marcus, OEM procurement | W1 network | medium | 3/5 | No on Riverstone at any price; pilot only on a low-stakes case that does not exist; Dave sign-off required pre-spend | ~$1,050 Meridian floor trip; $7,400 YTD verification | $4-5k theoretical, $0 realized against real 2026 volume | | 03 | Rita, DTC control (retired R3) | none | weak | 1/5 | No interview; carry-forward final | none | $0 (free tools suffice, final) | | 04 | Tom, 63, CNC shop owner | W5 shopped-print check | medium | 4/5 | Yes at $200 at full confidence; shrinks to $50-75 trial against cold-start; live candidate (Ferris) on the bench | ~$225 sunk into Ferris, unresolved; $650 Vandermeer anchor | $400-600 | | 05 | Kelly, 34, fab shop successor | W5 both halves | medium | 4/5 | Yes at $25/mo for visibility, no card, no date; commission structure killed on her own arithmetic | $32,000 Cimarron lost on freight (first real reason in 3 tries) | $300, flexing $180-600 | | 06 | Hank, foundry sales (retired R4) | none | no | 3/5 | No interview; carry-forward final | none ($10,830 cumulative stands) | $0 (ownerless pain, final) | | 07 | Priya, OEM sourcing engineer | W1 desk pass | medium | 4/5 | Conditional yes at $550, finalist-only; no candidate at finalist stage; declined to buy this quarter; Renee's free checklist ate the easy half | none ($1,250 R4 figure stands) | $550-1,100, some years $0 | | 08 | Sal, rep control (retired R3) | none | no | 2/5 | No interview; carry-forward final | none | $0 (hostile incumbent, final) | | 09 | Jenny, trading co owner | W4-adjacent monitoring | medium | 4/5 | Yes at $600/yr flat with a named next action (18-factory list this week); proved on her own math it would not have caught Ningbo | $405 mandatory VeriFact audit (Fujian) | $600 | | 10 | Derek, search fund CEO | W4 month-11 report | weak (down from medium) | 3/5 | Agrees $2,500 is right, declines to sign; "check back at month nine"; refuses peer-channel referral | none; let ZoomInfo + all 3 subscriptions lapse | one-shot $2,500, contingent, forward frequency zero | | 11 | Aisha, PE roll-up associate | W4 embedded module | medium | 3/5 | No: will not forward the one-pager; names exactly what she does to cold one-pagers (3 seconds, archive) | none; Vantage signed with the $2.1M tie never verified | $0 through her (Beacon Ridge channel only) | | 12 | Greg, EMS control (retired R4) | none | no | 1/5 | Close-out check only: zero events trailing 12 months, zero forecast | none | $0 (process-gated, final) | Roster shape check: 1 strong, 6 medium, 2 weak, 3 no. Derek downgraded on commitment behavior. Four retirements held as designed with no rescue attempts. The study closes without ever skewing positive: the final round subtracted a fit rather than adding one. ## Cross-cutting patterns **1. Every yes shipped with its own escape hatch, and the personas named the hatch precisely.** Not one evasion was vague. Dana: letter before card. Kelly: sign-up link, not a card, "when there's a real product I can log into." Derek: month nine, because signing now "doesn't buy me anything signing in month nine doesn't also buy me." Priya: a tracker note with a named trigger (finalist stage). Tom: a discounted trial until the network can return a real answer. These are honest buyers describing the actual distance between agreement and purchase, and the distance never closed. > "I'm not putting a card down today, though, send me the actual engagement > letter first. I got burned once by a company whose real terms were nothing > like their pitch." - Dana **2. The cold-start gate is universal, circular, and each persona quoted their own price for someone else going first.** Told plainly that no inspector has a catch on record and no network has data yet, nobody walked, but everybody repriced or deferred: Dana wants a named inspector with ten years at Bureau Veritas or SGS as a substitute for a catch record; Marcus wants one documented catch anywhere, any industry; Tom cuts the price 60-75% until "it's a real question, not a guess"; Priya calls the launch product "a nicer checklist question, and I already have a free one of those"; Derek wants E&O, a named investigator, and a law firm cosign; Aisha wants an incumbent's letterhead; Kelly wants a Texas-Louisiana footprint before she contributes data. Every substitute is something a cold-start vendor either does not have or can only get by making the first sale it cannot make. > "Zero catch record on a $1,100 ask against my highest-volume open > candidate is a hard no, not a soft one." - Marcus **3. The personas ran the vendor's kill math themselves, and the arithmetic never favored the vendor.** Round 5's most striking behavior: buyers doing honest unit economics out loud, unprompted, against their own earlier enthusiasm. Kelly multiplied her closed revenue by her own round-1 commission offer and withdrew it ($10-15k/year against Chris at $1,800). Aisha multiplied ceiling revenue by Ray's historical approval rate and got zero. Marcus computed a $4-5k year and immediately labeled it "imaginary" because zero of his real 2026 candidates would have used the product. Tom priced his own account at $400-600 and asked how anyone makes a living. > "Zero and eight to ten was the deal I offered you in round one. I've done > the math now. I'm taking it back." - Kelly > "Four hundred, maybe six hundred bucks a year... it doesn't sound like a > company, it sounds like a hobby somebody's running on the side." - Tom **4. The organization patches the acute half of the pain for free, faster than a vendor can sell, leaving only the hard half with no track record.** Renee added one mandatory line to the audit checklist inside six weeks and it is already standard on the manifold search, at zero cost. Kelly started asking buyers directly whether a drawing has been shopped, and "half of them just tell me." Dana opened a policy conversation about subcontractor disclosure in her supplier agreements. Each free fix eats exactly the product's cleanest selling point (the "nobody asked" failure) and strands the paid product on the harder claim (catching a liar) that it has never demonstrated. This is the same pattern that killed W3 via Hank's index card, now reproduced in the study's strongest segments. > "The checklist catches 'nobody asked.' It doesn't catch 'they said no and > it was actually yes.' That second one is still open, and that's the one > I'd pay for." - Priya **5. The round's two "successful" outcomes both failed on the dimension the buyer actually cares about, which validates the spec and indicts the market's luck-based status quo simultaneously.** Dana's blind $5,500 wire passed QC at 2.1% defect, and three weeks later a Slack acquaintance mentioned by chance that Herrera's undisclosed plating subcontractor had its wastewater permit suspended, exactly the sub-tier risk Dana flagged in R4. The outcome was fine; the information system that produced it was luck. Jenny walked the Ningbo timeline through the honest public-signal product and proved it could not have caught her $4,050 loss, then bought it anyway as $600 of insurance against a different, unpriced risk. Both events sharpen the same conclusion: the sub-tier map and the private-behavior early warning are the real products, and neither is what anyone can currently build or prove. > "I still don't know who does their plating, and it turns out that > mattered, I just got lucky that it mattered to a permit and not to my > parts." - Dana > "Don't sell it to me as the thing that would've caught Ningbo. Sell it to > me as the thing that tells me if a vendor's business itself changes > underneath me. Different product, still worth six hundred bucks." - Jenny **6. Champion-gated and channel-gated accounts confirmed terminal: the gate does not open for a stranger, ever, in this roster.** Marcus converted a supposedly solo $1,100 buy into a two-person sale by stating he would walk into Dave's office before picking up a phone. Priya will show Renee a sample before spending her own no-approval budget, and Renee's bar tightened this round ("their opinion with a letterhead"). Aisha archives cold one-pagers in three seconds and forwarded exactly one vendor in two years, warm-vouched. Derek refuses to vouch to his own cohort. Four rounds of "the buyer exists" resolved in round 5 into "the buyer exists behind a door only incumbents and warm referrals open." > "I get four or five vendor one-pagers a week I've never heard of. I give > them three seconds, I archive them, I do not forward them to Ray... and > I'd do the exact same thing to yours." - Aisha ## Vetoes and dealbreakers (added or hardened this round) - No card before a written engagement letter, and no card for a concept: document-before-money is now explicit for the two most willing buyers (Dana, Kelly). A landing page is not an artifact. - Full price for a no-data answer is dead on arrival. Cold-start checks must be discounted to throwaway money ($50-75) or free until the network can return a real answer (Tom). Charging full price and eating the buyer's hesitation produces an irritated non-returning customer, per Tom directly. - The highest-stakes open candidate is never the pilot. Marcus will not risk Riverstone at any price; the pilot slot is reserved for a low-consequence case that may not exist this year. Sell to the pipeline, not the crisis. - Pre-spend gatekeeper sign-off is mandatory for unproven vendors even under the buyer's own budget line (Marcus with Dave, Priya with Renee). The approval threshold only prices trusted incumbents. - Cold outreach has no channel: one-pagers die unread (Aisha), peer groups are closed to unvouched vendors (Derek), and the only forwarding on record in two years was warm-vouched. - Deposition-adjacent products require E&O insurance, a named credentialed individual, and a recognizable firm name; "our platform" and "our agent network" are disqualifying phrases (Derek, Aisha, consistent with R4). - Any contact with the watched party is terminal, restated as a closing condition on an actual yes: "don't let anyone on your side ever call the sales team pretending they need something. I will find out." (Jenny) - Commission on closed work dies whenever the buyer multiplies it against a cheap trusted human (Kelly, final). Standing vetoes all hold: vendor self-reported anything, percentage pricing to the Xometry-burned, subscriptions to the ThomasNet-burned, agents speaking in the user's name. ## WTP read (with required skepticism) Simulated WTP is a directional prior, and round 5's numbers are the study's best instrument (dated, structured, yes-or-no asks at the persona's own volunteered price) applied to its worst subjects (simulated agents that cannot actually feel a card leave a wallet). Read the shape, heavily discount the levels, and note that even simulated buyers, freed of real loss-aversion, declined to commit. Real humans should be assumed strictly worse on every number below. - Zero purchases, zero signatures, zero cards across five rounds and eight firm asks. Artifacts secured: one tracker note (Priya), one promised factory-list forward (Jenny), one request for an engagement letter (Dana), one request for a future sign-up link (Kelly). That is the complete commitment inventory of the study. - Roster-wide revenue ceiling if every stated yes converted at the stated number: roughly $5,000-6,500/year across eight active personas, before any COGS (and W1's COGS is a paid human inspector per event). No account exceeds $1,800/year on its owner's own math. - Every price is a substitution price or an insurance rounding error. Dana pegs to Kevin, Marcus to a Dave floor trip, Kelly to Bluebeam and Chris, Tom to the Vandermeer write-off, Jenny to 6% of her VeriFact spend, Derek to insurance on the holdback. Nothing in five rounds was ever priced as premium over the incumbent equivalent. - The conditional structure is more informative than the dollars: pay-after and escrow are table stakes (behaviorally proven in R4), cold-start discounts are expected, and the first full-price sale in every segment requires a credibility asset (catch record, BV/SGS pedigree, E&O plus law firm, incumbent letterhead) the vendor does not start with. ## Wedge status after round 5 (final) - **W1 Verification network + sub-tier desk pass: ALIVE, BUT SMALL AND GATED. The study's only survivor with a real buyer set.** Demand is genuine and persona-specified: Dana yes at $500 (letter first, cold-start line intact), Priya yes at $550 (finalist-only, Renee gate, no candidate today), Marcus yes in a future year where trust already exists. But round 5 bounded it: $500-1,800 per account per year, 1-4 events each, COGS that include a paid human per event, a circular cold-start trust gate, and free organizational fixes actively eroding the easy half of the value. The one named unlock that does not require a first customer: hire inspectors with independently checkable pedigrees (ex-Bureau Veritas, SGS, Intertek), which Dana volunteered as a full substitute for a catch record. As a business it needs hundreds of Dana-shaped accounts; as a service wedge for a founder with a channel, it is real. Verdict: alive, wounded, niche. - **W2 Agent-amplified rep desk: DEAD.** Closed R2-R3, control retired, record final. No change. - **W3 Castings/forgings desk: DEAD.** Closed R3-R4, ownerless-pain finding confirmed through Hank's final carry-forward. Its residues (sub-tier scope, shopped-print signal) live inside W1 and W5. No change. - **W4 Succession demand book: DEAD as a standalone wedge, final.** Round 5 executed the last two tests and both failed: Derek, the best-case buyer with live $150k exposure, agreed to the price and declined the signature, the lock-in, and the peer referral, and his forward frequency is zero; Aisha confirmed the standalone product is unbuyable through her at any price and the only funded path is Beacon Ridge's letterhead. What remains is a partnership or BD thesis (sell the module to incumbent DD firms), not a wedge a cold-start founder can open. Jenny's $600/yr public-signal monitor survives as a W4-adjacent micro-product: honest, buildable, confirmed with a next action, and explicitly not a fix for the loss that motivated it. - **W5 Commission rep in software: DEAD in its defining half, TINY in the other.** The commission structure, the wedge's entire reason to exist, was killed this round by the only persona who ever endorsed it, using her own closed-revenue arithmetic. The surviving visibility half is real demand at Bluebeam-class pricing ($25/mo Kelly, $200/check Tom at full confidence) with two unsolved structural problems: the cross-shop data cold start (Kelly will contribute only after a regional footprint exists; Tom will not pay full price for no-data answers) and per-account economics of $300-600/year. It is a feature-sized product in search of a network nobody has a reason to seed. Verdict: not a wedge; at most a future feature of something with distribution. ## What we still do not know These are now questions for real humans; the simulation has no further resolution to offer on any of them. 1. Whether any conditional yes survives contact with a real card. Dana's engagement-letter yes, Kelly's sign-up-link yes, and Priya's finalist-trigger yes are the three testable commitments; each specifies its own artifact, and building that artifact (a one-page letter, a sign-up page, a sample desk pass for Renee) is cheap. Real-human versions of these three asks are the obvious first field tests. 2. Whether an ex-BV/SGS/Intertek named inspector actually clears the cold-start bar with real buyers, and what such people cost to recruit. This is the only cold-start unlock any persona volunteered that does not require a prior customer. 3. Whether the sub-tier desk pass is buildable to Priya's bar (registry verified, near-zero false negatives on "no outside processing") and to Renee-class gatekeepers' evidentiary standards, which tightened, not loosened, under pressure. 4. Whether real shops will contribute quote flow to a shopped-print network before it has coverage. Kelly's reciprocity conditions and Tom's trial pricing sketch the mechanism; simulation cannot resolve whether it bootstraps. 5. Whether frequency times price supports any business shape. On the personas' own math the total roster is worth $5-6.5k/year. A real business needs either much higher event frequency (Jenny is the only compounding account in five rounds) or a segment this study never reached. 6. Whether the gatekeeper walls (Dave, Renee, Ray, IC) behave in reality as uniformly as they did in simulation, where they went 12-for-12 against strangers. If real gatekeepers are even occasionally permeable, W1's economics change materially. 7. Whether simulated consistency is itself an artifact. Five rounds of personas behaving with perfect economic self-knowledge (running their own kill math, holding lines across rounds) may overstate real buyers' discipline in both directions: real humans commit more impulsively and churn more quickly. Only field interviews resolve the direction of the error. ## Convergence call CONVERGED. This is the final designed round, but the call stands on the merits: another simulated round would change no decision. The wedge ranking did not move between R4 and R5 (it only resolved downward: W4 standalone dead, W5 commission dead, W1 confirmed lead-but-niche); WTP is now bounded above by firm asks and below by five rounds of zero commitments; every control behaved as designed through retirement; and all seven remaining unknowns are explicitly unresolvable in simulation (real cards, real gatekeepers, real network bootstraps). The study's terminal output for CROSS-ROUND.md: one wounded-but-alive wedge (W1 verification network with the sub-tier desk pass as lead artifact), two micro-products with named buyers and named conditions (Jenny's $600/yr public-signal monitor, the Bluebeam-class quote-desk flag), three dead wedges, and a field-test plan that is already written in the personas' own conditions: send Dana the engagement letter, show Renee the sample, and find out whether a real Kelly clicks a real sign-up link.