# Message test — https://b2bsignals.ai/ After reading your page, only 3 of 15 personas could name what kind of product this is, unprompted. - **Page tested:** https://b2bsignals.ai/ - **Audience tested against:** SMB sales and BDRs + founders - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/b2b-signals-outreach-that-names-what-they-just-gx4tm3c > These answers are generated by AI, scored on Wynter's B2B Message > Layers framework using behaviorally-diverse simulated personas. The > methodology is real and the critique is directional. What a simulated > persona cannot have is a live budget, a renewal coming up, or a boss > asking about this quarter. --- ## 01 · The scores Every persona answered all four questions. These are four independent proportions of the same panel, not stages of a funnel. | Layer | Question | Cleared the bar | Strength | Of those who passed | | --- | --- | --- | --- | --- | | 1. Clarity | Do they understand what you do? | 3/15 | 88% | 7 without hesitation, 8 with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 15/15 | 84% | 4 without hesitation, 11 with reservations | | 3. Value | Do they actually want it? | 14/15 | 74% | all with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 9/15 | 56% | all with reservations | **Brand alignment** (a side metric, not one of the four layers) — 15/15, 78% strength (all with reservations). Does the page read like the company you actually are? **Fix first: Clarity.** Earliest failing layer, walking the sequence in order — not simply the lowest score. ### What they thought you sell 5 of the personas who named a category got it wrong: - 1× “Sales intent / signal-based lead generation tool” - 1× “Sales intent data / outbound automation tool” - 1× “Sales intent signal / prospecting tool” - 1× “Sales intent-signal / outbound prospecting tool” - 1× “Sales intent/signal-based prospecting tool” --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Clarity **Define how the fit score is calculated, beside the score.** The page shows precise numbers — 94, 88, 86 — under a column called "Fit" and a card labelled "Fit score / 100", but never says what goes into them. Readers said they could not act on the recommendations without seeing the scoring logic, which is doubly awkward because the page also promises "Not a score, not a percentage." Add one line under the queue table naming the inputs, e.g. "Fit = signal recency + role seniority + your ICP filters (company size, stage, title). No black-box intent… *effort low · impact high · tested against Proof next to the claim* **Stop contradicting yourself on scores versus plain-words signals.** "Signal, in plain words — 'Hiring a Head of Growth, posted this morning.' Not a score, not a percentage" sits directly above tables whose most prominent column is a numeric Fit score. Readers flagged the terminology as muddled. Rewrite that block to say the signal is always quotable plain text and the score only orders the queue, e.g. "The signal is a sentence you can quote. The score only decides what you read first." *effort low · impact high · tested against Plain language* **Rank the eleven signals instead of listing them flat.** "Eleven signals. Every one you can quote." then a flat eleven-row table buries the point that a handful drive the results. The line "The three at the top produce most of the replies" is contradicted by the table, which is sorted by detection speed and where "Company followers" is separately called "The highest-converting source we have." Split the table into "The three that produce most replies" and "Also watched", and make clear one engine reads all of them. *effort medium · impact medium · tested against Conclusion first* ### Differentiation **Put MCP access through Claude and Cursor above the fold.** The clearest thing on this page that rivals cannot claim — querying the queue from Claude or Cursor over MCP — never appears in the copy tested here, and where it does surface it reads as a placeholder rather than something working today. Give it a named section with a real example of what a rep types and what comes back, and say whether it ships today or is in beta. That is the reason to choose this over any other signal tool. *effort medium · impact high · tested against Give a reason to choose you* **Turn the quotable-signal test into a head-to-head comparison.** "The only test we apply" and the "WE DON'T CHASE" list are the sharpest differentiating idea on the page, but they are framed as internal policy rather than a contrast a buyer can use. Rewrite as a two-column comparison — what intent-data vendors send you (page views, topic scores, unnamed accounts) versus what arrives in your queue (a named person, a dated public action, a quotable first line) — so the choice is visible without body copy. *effort medium · impact medium · tested against Give a reason to choose you* ### Value **Attach the reply-rate numbers to the accounts that produced them.** The headline metrics were discounted because the proof rests on two accounts over 90 days and the caveat sits away from the claim. Move the sourcing next to each number: who, what volume, what period, what baseline they replaced. "90% reached the same day" and "9,090 in queue right now" mean little without knowing whose pipeline they came from. Named results from Payoneer or Sommo beside the figure would carry more than the unnamed logo wall does. *effort medium · impact high · tested against Proof next to the claim* **Answer the "will it work on my data" objection with a pilot offer.** Readers said they would only believe the lift after seeing it on their own volume, and the page leaves that unresolved — the free scan of a single LinkedIn post URL is not the same thing. Add a short block answering it directly: what a pilot on their own ICP looks like, how many days before the first queue arrives, and what a success threshold would be. Keep "Scan it free" as the entry point but say where it leads. *effort medium · impact high · tested against Answer the live objection* ### Relevance **Name the role and team size the queue is built for.** Readers worked out who this was for from workflow mechanics — the fifteen-minute queue, the j/k/a/e keys — rather than from any line they could point at. Add a named-audience line under the headline, e.g. "For founders and two-to-ten-person outbound teams doing their own prospecting" and state whether it suits an SDR pod or a solo seller. It costs one line and stops the reader inferring. *effort low · impact medium · tested against Name the audience* ### Brand alignment (side metric) **Replace the bare logo wall with one named customer result.** "Teams already running on B2B Signals" followed by Payoneer, Sommo, ERA and the rest carries no name, number or quote, and the incomplete product screenshots read as unfinished next to the plain-spoken founder voice everywhere else. Swap one logo for a two-line attributed result — person, role, company, what changed — and finish the screenshots so what is shown matches what the copy promises. *effort medium · impact medium · tested against Proof next to the claim* --- ## 03 · What is working ### The problem and audience land within the first two lines Four respondents said the headline and early proof made the problem and intended audience clear without hunting. One specifically credited the same-day signal messaging versus three-week generic outreach framing, and another said the first-line opener test made the page worth reading rather than skimming. > the headline "Twenty people worth writing to. Every morning." plus the "paste a LinkedIn URL, we show you everyone who engaged" bit tells you in ten seconds this is for outbound sales/SDR types > > — Sales Manager, Technology, 1-10 > It was clear within the first two lines — "Twenty people worth writing to. Every morning" plus "We watch hiring posts, competitor engagement and job changes across your market" told me exactly what this does before I hit the fold. > > — Founder, B2B Services, 201-500 > the "first line you can quote back to them" test is the one bit of specificity that made me stop and actually read rather than skim past as fluff > > — Founder, B2B Services, 11-50 > It's obvious within the first line: "Twenty people worth writing to. Every morning" plus the subhead about hiring posts, competitor engagement and job changes tells me exactly what problem this solves > > — Business Development Representative, B2B Services, 51-200 ### The claimed lift is large enough to justify a workflow change if it survives validation Two respondents said the signal-based scoring with AI-written openers beats cold outreach by a margin that would warrant changing their workflow — explicitly conditional on the results holding beyond the pilot. > If the numbers hold — 55% acceptance vs 24% cold, 30% LinkedIn reply vs 6%, 8% email reply vs 2% — that's a real jump over what I'm doing manually right now > > — Business Development Representative, B2B Services, 1-10 > the acceptance/reply table (55% vs 24% acceptance, 30% vs 6% reply) is the kind of number that, if it held up outside a two-account sample, would actually move our pipeline math > > — Head of Sales Development, SaaS, 11-50 ### Respondents named MCP integration through Claude and Cursor as the differentiator Four respondents pointed to MCP integration via Claude and Cursor as a concrete claim competitors do not make. One qualified it, saying it is presented as an unverified placeholder rather than something demonstrated. > "Your signals live inside Claude and Cursor through MCP... Nobody else in this category has this" — is the one thing that would actually tip a shortlist decision > > — Sales Development Representative, SaaS, 51-200 > The MCP integration — "Your signals live inside Claude and Cursor through MCP... Nobody else in this category has this" — is the one concrete differentiator I'd actually weigh against alternatives > > — Business Development Representative, B2B Services, 1-10 > The MCP integration line — "Nobody else in this category has this" letting you query signals from inside Claude or Cursor — is the one thing that would actually differentiate it on a shortlist, if it's real and not "capture pending" vapourware; right now it's an unverified screenshot placeholder > > — Senior Sales Development Representative, Technology, 201-500 --- ## 04 · What the personas said ### The fit score is quoted as a precise number but never defined Three respondents said the value proposition depends on understanding how the fit score is calculated, and that the page presents precise numbers without defining them. They said they could not accept the recommendations without that scoring logic being transparent. > the only soft spot is "fit score" being thrown around before you know how it's calculated > > — Head of Sales Development, SaaS, 11-50 > That's worth a meeting, but only a technical one where I can ask exactly how "fit" is scored, what happens when the scraped signal is stale or wrong, and whether their "2 accounts, first 90 days" sample means anything at our volume > > — Founder, B2B Services, 11-50 > the only fuzzy word is "fit score" (94, 88, 91 etc.) — it's presented as precise but never defined, so I can't tell if it's a real model output or a made-up number dressed up to look rigorous > > — Business Development Representative, B2B Services, 51-200 ### Terminology is muddled by renaming established terms and listing signals without ranking Two respondents said the copy conflates established terms — signal, queue, fit score — with the product's own feature names, and that listing multiple signal types at once without ranking obscures the fact that one engine drives them all. > the mix of signal types (hiring, competitor, influencer, G2, RSVPs) listed all at once without ranking makes you do a bit of work to see it's one engine, not five bolted-on features > > — Head of Sales Development, SaaS, 11-50 > "signal," "queue," "opener," "fit score," "ICP qualified" all get thrown around like established category terms when really they're just their names for scraping + templated first lines > > — Sales Manager, Technology, 1-10 ### The two-account, 90-day sample is treated as the page's central credibility problem Six respondents dismissed or discounted the headline reply-rate and acceptance metrics because the proof rests on two accounts over the first 90 days. Several said they would need a live pilot on their own data and volume before believing the numbers, and one noted the sample-size caveat undercut the metrics even though it read as honest. > which is a sample size I wouldn't hang a purchase decision on, and I'd want to see it across more accounts and a longer window > > — Sales Development Representative, SaaS, 51-200 > But it's "2 accounts, first 90 days" — that's a pilot, not proof, and I'd want to see it against my own list before I believe it transfers. > > — Business Development Representative, B2B Services, 1-10 > the only proof offered is "two accounts, first 90 days" against platform averages, which is thin for a claim that big > > — Senior Sales Development Representative, Technology, 201-500 > That's worth a meeting, but only a technical one where I can ask exactly how "fit" is scored, what happens when the scraped signal is stale or wrong, and whether their "2 accounts, first 90 days" sample means anything at our volume > > — Founder, B2B Services, 11-50 > "2 accounts, first 90 days" — that's a sample size a rounding error could produce, and until I see it across more accounts or at our own volume it's not proof, it's a promising anecdote dressed as a metric > > — Founder, B2B Services, 11-50 > I'd want to see it live on our own market before I'd trust the "2 accounts, first 90 days" numbers as anything more than a best-case anecdote > > — Sales Manager, Technology, 51-200 > 2 accounts, first 90 days > > — Senior Sales Development Representative, Technology, 11-50 ### Incomplete screenshots and an unnamed logo wall undercut the founder credibility Two respondents said the visual proof worked against the page: incomplete screenshots undercut the authenticity the founder tone had built, and the logo wall carried no named case studies or customer quotes to back the claims. > the parts that don't are the ones still marked "capture pending," which undercuts the otherwise credible, practitioner voice > > — Sales Development Representative, SaaS, 51-200 > the logo wall (Payoneer, Somm, ERA, Clarity Global...) with zero named case studies or quotes behind it — I'd want to hear from one of them directly, e.g. "Payoneer's SDR team saw X" > > — Senior Sales Development Representative, Technology, 201-500 ### The target reader is inferred from workflow mechanics, never stated Four respondents said they worked out who the page was for from the mechanics and workflow detail rather than any explicit persona statement, and one asked directly for role naming and team size context. Several treated this as an omission rather than a failure to understand. > I'd want a line naming the role directly — "built for BDRs and SDRs running outbound" — plus a mention of team size or quota context, so I'm not inferring it from workflow details > > — Business Development Representative, B2B Services, 51-200 > The reader is inferred rather than stated outright — it's someone running or managing outbound (SDR, founder, sales lead) at a B2B company > > — Founder, B2B Services, 11-50 --- ## 05 · The hardest read An adversarial pass over the findings. Every claim below was checked against the panel's own answers; unsupported ones were dropped. - **The page's core numbers are its largest liability, and the honesty of the caveat does not rescue them.** *(high)* Six of 15 respondents — the largest theme on the page — discounted or dismissed the headline reply-rate and acceptance metrics because the proof rests on two accounts over 90 days, and several demanded a live pilot on their own data before believing anything. One explicitly said the sample-size caveat undercut the metrics even while reading as honest, so the disclosure buys goodwill and costs belief at the same time. - **Every positive result on the page is conditional on proof the page does not supply, so nothing here can be banked.** *(high)* The value upside is explicitly conditional — the two respondents who said the lift would justify a workflow change tied it to results holding beyond the pilot. The differentiator is qualified the same way, with one of three respondents calling MCP integration an unverified placeholder rather than something demonstrated. Meanwhile six respondents rejected the proof outright. The page's wins are promissory notes drawn on an account six people said is empty. - **Quantified claims are actively counterproductive because the page publishes numbers it will not explain.** *(high)* Three respondents said the fit score is quoted as a precise number with no definition, and stated they could not accept the recommendations without transparent scoring logic. Precision without derivation reads as a claim to be checked rather than a benefit to be enjoyed, and the same suspicion falls on the reply-rate figures six respondents already discounted. - **The visual proof cancels out the founder tone the page works to establish.** *(high)* Two respondents said incomplete screenshots undercut the authenticity the founder voice had built, and that the logo wall carries no named case studies or customer quotes. Combined with the two-account sample flagged by six respondents, the page shows logos it cannot substantiate and screenshots it cannot complete — the assets meant to prove the claims are the ones inviting doubt. - **The page never names its reader, forcing every prospect to self-qualify from mechanics.** *(medium)* Four respondents reverse-engineered the target reader from workflow detail rather than any explicit statement, and one asked directly for role naming and team size. Clarity of problem does not substitute for clarity of audience: readers should not have to audit the workflow to learn whether the product is for them. - **Renaming established categories hides the product's actual architecture from buyers.** *(medium)* Two respondents said the copy conflates signal, queue and fit score with proprietary feature names, and that listing signal types without ranking obscures the fact that a single engine drives all of them. Paired with the three respondents who could not get a definition of the fit score, the vocabulary problem is not stylistic — it is concealing how the product works. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | Founder | B2B Services | 11-50 | | 2 | Sales Development Representative | SaaS | 51-200 | | 3 | Senior Sales Development Representative | Technology | 201-500 | | 4 | Business Development Representative | B2B Services | 1-10 | | 5 | Head of Sales Development | SaaS | 11-50 | | 6 | Sales Manager | Technology | 51-200 | | 7 | Founder | B2B Services | 201-500 | | 8 | Sales Development Representative | SaaS | 1-10 | | 9 | Senior Sales Development Representative | Technology | 11-50 | | 10 | Business Development Representative | B2B Services | 51-200 | | 11 | Head of Sales Development | SaaS | 201-500 | | 12 | Sales Manager | Technology | 1-10 | | 13 | Founder | B2B Services | 11-50 | | 14 | Sales Development Representative | SaaS | 51-200 | | 15 | Senior Sales Development Representative | Technology | 201-500 | --- ## 07 · Before you act on this The methodology is real, and the critique is directional. What a simulated persona cannot have is a live budget, a renewal coming up, or a boss asking about this quarter. **Validate anything you're betting on with real ICPs who are actually in-market.** Being wrong is more expensive than you think. Finding out is cheaper than you'd guess. Wynter runs message testing with verified B2B professionals — trusted by HubSpot, RingCentral, Shopify, Cognism, Paddle, Veeam, Rippling and Miro. This report is kept for 60 days from 2026-08-25, then deleted along with the personas and their answers.