Most of the conflicts in this market come from three habits. Vendors print different numbers on the same pricing page. Revenue "facts" are mostly GetLatka estimates or founder claims passed between blogs. Benchmark posts recycle each other's figures. This register collects every ## Contradictions entry from the 15 research notes, plus the places where hub pages disagree with each other. Each row says which number we use and why.
How to read the tables: Claim A is the claim we lean towards unless "Our read" says otherwise. Which page uses which names the hub pages, so the fix list at the end can be worked through. All prices and limits are as of October 2026 unless a row says otherwise.
Company numbers
| Topic | Claim A (source) | Claim B (source) | Which page uses which | Our read |
|---|---|---|---|---|
| Instantly ARR | ~$20M ARR, Dec 2024 (What a Startup) | $38.2M for 2026 (GetLatka estimate) | Instantly shows both; History of cold email software uses $20M; Market size, Build costs and team, Distribution use $38.2M | Not a contradiction in time, but $38.2M has no second source. Always label it "GetLatka estimate". |
| Instantly early scale | $2.4M ARR and 2,900 customers after 9 months (GetLatka interview) | "$8M ARR in less than 2 years" (Discovered Labs podcast) | Instantly, Build costs and team use 2,900 | Different points in time. Both are founder-told. |
| Instantly customer count | "40,000+ customers" (developers page); "50,000+ sales teams" (pricing) | "700,000+ businesses" (benchmark report) | Instantly shows the range; Market size uses 50,000+; The AI shift, AI SDR layer quote 700k+ | 700k counts sign-ups, not payers. Use 40–50k as the vendor's paying-customer claim. |
| Instantly agency revenue before the SaaS | $10–15K MRR (GetLatka interview, above) | $30–50K/month (Instantly llms.txt) | Research notes only | Founder story told two ways. Immaterial. |
| Instantly headcount and founders | ~200 staff (GetLatka 2026 estimate) | "50+" (llms.txt); 15 (StarterStory, stale) | Build costs and team, Distribution use ~200 | 200 is an estimate that includes data, CRM and support teams. Founders: llms.txt lists Raul Kaevand and Nils Schneider as co-CEOs; StarterStory and GetLatka each name one. |
| Smartlead ARR | $14M, 2024 (GetLatka estimate) | "$20M+ ARR" (ZoomInfo review; yespress) | Smartlead shows both; Market size, Build costs and team use $14M only | Both second-hand; the $20M+ comes from competitors' content. Quote "$14–20M+, unverified". |
| Smartlead customers vs revenue | "Used by over 100,000 businesses" (pricing) | $14M revenue (GetLatka) implies ~$140 per business per year | Market size makes the calculation | Only reconcilable if "used by" counts trials and churned accounts. |
| lemlist $40M ARR date | October 2024 (Founderpath) | Announced January 2026 (arr.club) | lemlist uses Jan 2026 | Jan 2026 fits the $28M January 2025 figure. Use it. |
| lemlist current ARR | ~$50M ARR, founder-claimed, dated 28 Apr 2026 by arr.club; ColdIQ says "just crossed" in July (ColdIQ) | $53M revenue 2026, up from $40M in 2025 (GetLatka estimate) | lemlist uses $50M claimed; Market size uses GetLatka $53M; Non-English markets says "$28M ARR (2024)" | Use ~$50M (founder-claimed, 2026). Non-English markets is out of date and misdates the $28M figure. |
| lemlist funding | None; founder says he turned down offers (lemlist) | "$30M in total funding across 1 round", "$150M valuation" (GetLatka) | lemlist says none; Funding and exits reports GetLatka with a flag | No primary source for any round. Treat lemlist as bootstrapped. |
| lemlist founding year | 2017 (beehiiv, above) | 2018 (arr.club; Product Hunt launch Jan 2018) | lemlist "2017–2018"; Distribution "launched 2018" | Built 2017, launched 2018. Both pages are compatible. |
| Salesforge revenue | $3M ARR within 12 months (founder on Indie Hackers, Mar 2025) | $135.5K for 2024 (GetLatka estimate) | Salesforge uses $3M with a flag | GetLatka is ~20× below a 100-person team's plausible revenue. Use $3M, labelled founder-claimed. |
| 11x ARR | ~$10M claimed vs ~$3M real (TechCrunch) | "$14M reported" (Adrata) | 11x uses TechCrunch and flags Adrata | TechCrunch is the only reporting with named sourcing. |
| 11x Series A month | September 2024 (TechCrunch) | April 2024 (Neuronfeed) | Funding and exits uses Sept; 11x notes the dispute | Use September 2024 (TechCrunch). |
| Artisan Series A lead | Glade Brook (TechCrunch) | YC and HubSpot Ventures (Wikipedia) | Artisan | Wikipedia likely lists earlier backers. Use TechCrunch for the round. |
| Apollo customers | 40,000 paying (About) | 100K+ (newsroom) | Apollo.io | Different definitions (paying vs total). Quote both with labels. |
| Apollo CEO | Tim Zheng (About) | Matt Curl (Aug 2026 release) | Apollo.io flags it | The newer release is more likely current. The handover date is unknown. |
| Apollo growth | ARR path $96M (2023) → $150M (May 2025) (Sacra) | Revenue "more than fivefold" since Aug 2023 (Apollo, quoted by Sacra) | Apollo.io uses ~$150M | Reconcilable only with very steep 2025–26 growth. No primary source for "fivefold". |
| Clari + Salesloft CEO | Andy Byrne at announcement (Salesloft) | Steve Cox at close (BusinessWire) | Salesloft | Sequential, not contradictory. The close-date release wins. |
| Outreach size | 680 staff (GeekWire, Nov 2024) | 1,400 staff (GetLatka, 2025) | Outreach flags both | GeekWire reported actual layoffs. Use ~680. |
| Amplemarket Series A | $12M (vcbacked) | $13M (salestools.io) | Amplemarket uses $12M, notes $13M | Immaterial. |
| Unify total raised | $52M (company release) | ~$59M (Sacra, incl. $6.6M seed) | Unify shows both | Sacra adds the seed. Both are right on their own terms. |
| HeyReach base | US HQ (arr.club) | Team in North Macedonia (founder podcast) | HeyReach | Likely a US entity with a Macedonian team. |
| Expandi customers | 12,500+ (SaaStock) | 3.7K (GetLatka) | Expandi | Probably users vs paying accounts. Neither is audited. |
| Reply.io basics | Founded 2014 (Reply) | 2016, HQ San Jose; revenue falling $14.7M→$8.7M while staff rose 59→79 (GetLatka) | Reply.io uses 2014 and flags GetLatka | The GetLatka series contradicts itself. Do not quote the decline. |
| Saleshandy scale | "25,000+ customers" (own about page) | $1M ARR, 100K customers, $10 ACV (GetLatka) | Saleshandy, Market size flag it | GetLatka cannot be right for a $34–189/mo product. Ignore it. |
| Snov.io origin and size | Ukrainian company, "10,000+ teams" (Snov.io blog) | New York HQ, 150,000 clients, $22.7M (GetLatka) | Snov.io flags it | Use the company's own statement on origin. Revenue is unknown. |
| Other founding details | Klenty 2015, Delaware (IIMB); LGM 2019 (own site); Hunter "7+ million" users (Hunter) | Klenty 2016, Palo Alto (GetLatka); LGM 2020 (GetLatka); Hunter "2M customers" (GetLatka) | Klenty, La Growth Machine, Hunter flag each | GetLatka's company metadata is unreliable. Prefer first-party pages. |
| PlusVibe jurisdiction | "US" (indexator) | Governed by Polish law (privacy policy) | PlusVibe | Delaware-style front, Polish operating law. Relevant if PlusVibe is an EU competitor. |
| Maildoso scale | 10M+ emails/day across 400k+ mailboxes (25/day average) | "Up to 15 emails per day" per mailbox (Maildoso) | Maildoso flags it | Both are Maildoso's own numbers. At least one is rounded marketing. |
| Review counts | Smartlead G2: 419 (G2) or 449 (G2); Trustpilot 140 reviews / 3.0 (Trustpilot) | Smartlead G2 306, Trustpilot 85 / 3.4 (Prospeo); lemlist G2 from 1,345 to 2,268 depending on the page | Customer voice | Counts depend on the fetch date and the G2 page. Quote ranges, not single counts. |
| Instantly Trustpilot | TrustScore hidden for a guideline breach (one fetch, 4 Oct 2026) | Normal score of about 3.8 (second fetch) — both Trustpilot | Instantly flags it | Unresolved. Check manually before citing. |
| Market size | Cold email software $1.8B in 2025 (Dataintelo) | Bottom-up sum of native tools ≈ $0.2B (research note); hub estimate $0.22–0.32B | Market size uses the bottom-up range | Analyst figure is not reproducible. Use bottom-up. |
Pricing and packaging
| Topic | Claim A (source) | Claim B (source) | Which page uses which | Our read |
|---|---|---|---|---|
| Instantly Growth price | $47 monthly, $37.60 annual (pricing, Oct 2026) | "$37/month" in many reviews and the AUTHORING example; $37 in May 2024 (StarterStory) | All current pages use $47; Build costs and team uses the $37.60 annual price correctly | $47 monthly is current. $37 is the old monthly price or the new annual price. |
| When Instantly raised the price | 27% rise in August 2026 (IndustryLens) | "We could not fetch an archived page to date the change" | Instantly, History of cold email software date it Aug 2026; Pricing landscape says it could not | Use August 2026 (single source). Update Pricing landscape. |
| Instantly Hypergrowth cap | 125,000 emails (monthly card) | 100,000 (annual card and comparison table), same pricing page | Instantly shows both; Pricing landscape, Sequencing & sending, PlusVibe use 125k | Use 125k for monthly and note the annual card. |
| Instantly placement tests | Ticked on all Outreach plans (pricing) | Separate $47/$97 plans with 2 free tests (plans overview) | Inbox placement testing uses the help version | The help centre is more specific. Treat it as a paid add-on. |
| Instantly AI reply credits | 5 credits per reply "regardless of whether you send, edit, or skip it" (help) | "Rejected drafts in HITL mode incur no credit charges" (blog) | AI reply agent | Unresolved. It matters for anyone pricing a reply desk against Instantly. |
| Instantly affiliate rate | "20%–35%" | "20–40%" — both on instantly.ai/affiliate | Distribution uses 20–40% | Tiered programme. Quote "20–40%, tiered". |
| Smartlead API gating | Pro ($94) and up (pricing.md) | API "add-on" on Pro (one fetch summary); GTM engineers says "$174+" | API, webhooks and MCP, Smartlead, CRM sync say Pro; GTM engineers says $174+ | Use Pro ($94). Fix GTM engineers. |
| Smartlead client workspace price | $29/month (pricing.md) | $39 (add-on row on pricing page) | Smartlead shows both; Business models, White-label and agency portals, Workspaces, roles and SSO use $29 | Use "$29–39". |
| Smartlead Prime send cap | 500,000 (pricing) | 510,000 (llms.txt) | Smartlead, Pricing landscape, Sending limits and throttling use 500k | Immaterial. Use 500k. |
| Smartlead SmartDelivery | "Unlimited tests, no extra fees" (marketing page) | Add-on at $49 / $174 / $599 (pricing) | Business models, Inbox placement testing, Warmup networks use the add-on prices | Pricing page wins. |
| EmailBison price | $599 flat for 500k emails (pricing) | $499 (ModernInbound); tiers from $99 (LeadHaste) | All pages use $599 | $499 was the pre-2026 price. The $99 tiers have no other source. |
| Clay entry plans | Launch $167, Growth $446, 3,000 data credits (clay.com, fetched 4 Oct 2026) | Launch $185, Growth $495 (Sacra); 2,500 credits (Databar); Explorer $149 (SyncGTM) | Clay, AI personalisation use $167 / 3k; Contact data and enrichment uses $185 / 2,500 and $495 | Vendor page wins. $185/$495 may be monthly vs annual. Fix Contact data and enrichment. |
| Unify pricing | $0 / $20 / $60 per seat (unifygtm.com, Oct 2026) | $1,740/mo annual Growth (Landbase, Apr 2026); "~$700/mo" (Sacra) | Unify uses the vendor page | A real 2026 repricing, not an error. The date of the switch is unknown. |
| Apollo plans | Organization $149 monthly, Professional $94 (Docket) | Top tier cut to $119 in March 2026, Enterprise added (Costbench); Professional $99 (Prospeo); live page lists Free/Basic/Professional/Custom (Apollo) | Apollo.io uses "Basic from $49/user/mo annual (third-party)" | Apollo's own prices did not render. Treat every Apollo price as third-party. |
| Amplemarket credits | 13,500 / 35,000 / 40,000 email credits per user per year (pricing) | 15,000 / 70,000 / 100,000 (ZoomInfo blog) | Amplemarket uses the vendor numbers | Vendor page wins. ZoomInfo is a competitor. |
| Woodpecker entry price | ~$24/mo annual (500 prospects) | $29 (LeadHaste); $35 (Capterra) | Woodpecker uses ~$24 annual; Pricing landscape, Business models use $35 | All derive from $7 per 100 prospects with different billing periods. Label the period. |
| Mailshake plans | $49 / $99 with 2 / 5 accounts (Mailshake) | $29 Starter, 10 accounts on $99 (Prospeo) | Mailshake uses the official page | Official page wins. |
| HubSpot Sales Hub Starter | $20/seat/mo | $7/seat/mo — both on HubSpot pricing | HubSpot Sequences flags $7 as likely promotional | Use $20 as list price. |
| La Growth Machine | Pro and Ultimate both €120/identity (fetch) | Pro "$110/month billed annually" on the same page | La Growth Machine flags it | Parsing error likely. Check the live page before quoting Ultimate. |
| lemlist Email plan | "Unlimited users" and unlimited senders (pricing) | Footer says "per user" (Valley); "1 sender" (older review) | lemlist flags it | Senders went unlimited in 2026. The per-user question is open. |
| Expandi Agency | Custom, 10+ seats (Expandi) | ~$149/seat (LeadHaste) | Expandi | Negotiated. Do not quote $149 as list. |
| Salesforge Agent Frank | $499/mo billed quarterly (pricing) | "$599/month annually" (Agent Frank page) | Salesforge | Salesforge's own pages disagree; 11x's review flags it too. |
| Infra list prices | Primeforge $4.50/mailbox annual (Primeforge); MailReach $19.50 (MailReach) | Primeforge "from $2.88" (Inboxkit); MailReach $20–25 (MailReach blog, oneaway); lemwarm $29/$49 vs $25 | Mailforge, MailReach, Warmup networks use vendor prices | Vendor pages win. Competitor tables lag or show volume discounts. |
| Google Workspace list price | Business Starter ~$7/user/mo (EmailBison blog) | ~$5 (Maildoso help content); Google's own page did not render | Google vs Microsoft vs SMTP, Inbox & domain infrastructure use ~$7; Volume math: what 10k and 100k emails a day cost declines to state a price | Unverified either way. This number drives the "how are $3 inboxes possible" question in Open questions. |
| Google CASA cost | $540–720 (TAC Security) and ~$700 (Nango) | $3,000–6,000 for AL1 (Leviathan); "$15k–75k legacy" (Unipile, unverified) | Build costs and team budgets $540–720 in its table and $0.5k–6k in its advice; OAuth verification and the end of basic auth shows the spread | Price depends on lab and assurance level. Budget $6k until two quotes are in. |
Provider policy and deliverability rules
| Topic | Claim A (source) | Claim B (source) | Which page uses which | Our read |
|---|---|---|---|---|
| Outlook.com May 2025 enforcement | Non-compliant high-volume mail goes to Junk first and "may be rejected" later (Microsoft policy page) | Immediate rejection with 550 5.7.515 (URIports; Resend) | Microsoft rules: Outlook.com and Microsoft 365 limits says Junk first, possible 550; Provider crackdowns and History of cold email software say "rejects" | Observed behaviour is rejection; the policy text says Junk first. Pages should say "policy: Junk then reject; observed: 550 from day one". |
| Microsoft per-mailbox 2,000/day external limit (MERRL) | Cancelled January 2026 (Office365itpros) | "Dormant rather than dead"; withdrawal announcement not fetched | Microsoft rules: Outlook.com and Microsoft 365 limits, Provider crackdowns, OAuth verification and the end of basic auth say cancelled; Provider rules: Google, Microsoft and the ESPs says dormant | Office365itpros reports the cancellation. Align Provider rules: Google, Microsoft and the ESPs. |
| Microsoft 2026 cold-email rules | Microsoft Learn limits page and Office365itpros mention no registration scheme | Bulk-sender registration, 14-day domain age, 3M accounts blocked in Q1 2026 (Litemail) | Provider crackdowns flags Litemail as unverified | Litemail cites no primary source. Do not build on it. |
| Gmail Postmaster reputation | Google's guidelines still say check "domain and IP address reputation with Postmaster Tools" (Google) | Those dashboards were retired on 30 Sept 2025 (Twilio) | Reputation and blocklists, How filtering works treat the guideline as a leftover | Retired. Do not promise customers Google reputation data. |
| Gmail scope classification | gmail.readonly and gmail.modify are restricted (Google) | Listed as sensitive (Unipile) | OAuth verification and the end of basic auth uses restricted | Google is authoritative. Restricted means CASA. |
| CASA naming | AL1 / AL2 assurance levels (Google; ADA) | "Tier 2 / Tier 3" (Nango, Unipile) | OAuth verification and the end of basic auth | Same programme, old and new names. Use AL1/AL2. |
| App passwords on Workspace | App passwords remain the stated exception after less-secure apps were switched off (Google) | "Google does not allow plain password authentication for Workspace accounts" (Maildoso help) | OAuth verification and the end of basic auth, Sending architecture use Google's version | Probably a tenant-policy setting, not a platform rule. |
| Workspace AUP | Workspace AUP URL (workspace.google.com) | Now renders as the Google Cloud AUP | Provider rules: Google, Microsoft and the ESPs | Unclear whether a Workspace-specific AUP still exists. Low stakes. |
| Purchased lists | Allowed if they "follow the local privacy regulations", framed as a deliverability risk (lemlist Sending Policy) | Banned outright (Mailgun AUP; Twilio email terms) | Provider rules: Google, Microsoft and the ESPs, Platform liability: what the sequencer itself risks | Not a factual conflict: sequencers and ESPs have different rules. That gap is why cold email runs on mailboxes. |
| Safe daily sends per mailbox | Instantly recommends 30 campaign emails/day (Sending limits and throttling); Mailforge 30, max 100 (Mailforge); ScaledMail 25 Google / 10 Microsoft (ScaledMail); Maildoso 15 | Instantly's warmup post: ramp to "80–100 emails per day" (Instantly) | Sending limits and throttling ~30; Volume math: what 10k and 100k emails a day cost, How filtering works, Microsoft rules: Outlook.com and Microsoft 365 limits 10–30; Google vs Microsoft vs SMTP 15–30; Sending architecture 20–50; In-house SDR teams Apollo 50–100 | 10–30 is the 2026 norm; 80–100 is 2022-era advice. Align Sending architecture. |
| Safe Microsoft volume per mailbox | 20–100/day (EmailBison) | 10/day (ScaledMail); 30–50 (getle.ad) | Google vs Microsoft vs SMTP, Volume math: what 10k and 100k emails a day cost use ScaledMail's 10 | Vendor-set caps. The conservative figure is the safe planning number. |
| Apollo warmup | Discontinued in 2024 (MailReach; Docket) | Offered via Warmbox.ai (Apollo KB; deliverability page) | Apollo.io says reinstated through a third party; Amplemarket says Apollo sells warmup | Apollo's own October 2026 docs win. |
Benchmarks and performance claims
| Topic | Claim A (source) | Claim B (source) | Which page uses which | Our read |
|---|---|---|---|---|
| Average cold reply rate | 3.43% for 2025 (Instantly); 3.7% for H1 2026 (Saleshandy) | 0.45% for 2025 (Belkins) | Reply benchmarks, Volume math: what 10k and 100k emails a day cost show the range; 11x, Instantly, AI SDR layer, Replies, booking and handoff cite 3.43% alone; Campaign analytics, Copy and personalisation layer cite a Backlinko ~8.5%; Signals layer says 1–3% | Denominators differ (delivered vs total sent; auto-replies in or out). Every page quoting a single number should say whose and point to Reply benchmarks. |
| Why Belkins is low | Belkins excludes auto-replies (lemlist) | Belkins switched the denominator from openers to total sends (Belkins) | Reply benchmarks | Use Belkins' own explanation. |
| Saleshandy positive vs total | Total reply rate 3.7% | Positive reply rate "3% to 5%" — same report | Reply benchmarks | A positive rate above the total rate means different bases. Unusable as a pair. |
| "Platform-wide decline" | Woodpecker: 5.1% (2024) → 3.43% (2026) (Woodpecker) | 3.43% is Instantly's figure, not Woodpecker's (Instantly) | Reply benchmarks | The "decline" compares two vendors' datasets. Not a trend. |
| AI SDR reply performance | 9.7% reply rate (11x homepage); "up to 50% higher" than manual (SuperAGI via Autobound) | 0.5–1.5% after week six vs 4–6% human (leadgen-economy) | AI SDR layer, 11x, The AI shift show both sides | Every number is vendor or competitor content. No neutral test exists. |
| AI SDR churn "50–70%" | Attributed to UserGems (Autobound) | The UserGems post does not contain it; Prospeo attributes it to Reddit | AI SDR layer, The AI shift call it unsourced | Do not quote it. The 11x "70–80% early churn" (TechCrunch, former staff) is the only sourced figure. |
| ESP matching uplift | Matched 87–92% vs cross-provider 74–81% inbox (EmailBison); 94–96% vs 85–88% (Prospeo); "5–15% improvement" (Octave) | "Same rate" for authenticated Google→Microsoft mail (getle.ad) | ESP matching flags all as unaudited | No published method anywhere. Treat as plausible, unmeasured. |
| Warmup network size | Instantly "1,000,000+" / "4.2M+ accounts"; Smartlead founder "5–7M mailboxes" | Smartlead "35,000+" (Marketbetter) | Warmup networks, Built-in warmup say not comparable | Accounts vs domains vs "real inboxes". Do not rank networks by size. |
| lemlist new-mailbox cap | 30–40/day (lemlist help) | "40 (new inboxes)" (Prospeo table) | Sending limits and throttling uses lemlist's own | Immaterial. |
Law
| Topic | Claim A (source) | Claim B (source) | Which page uses which | Our read |
|---|---|---|---|---|
| German B2B email consent | Express prior consent required for email; "presumed consent" applies to phone only (IHK Köln; AG Düsseldorf 2026) | B2B email allowed with "express or presumed consent" (the research brief and the AUTHORING example) | Germany uses Claim A; Founders and small B2B teams and Non-English markets say "prior or presumed consent" | Claim A is correct. Fix Founders and small B2B teams and Non-English markets. This is the most decision-relevant legal error in the hub. |
| GDPR layer for German cold email | DSK view: a UWG breach is also a GDPR breach (DLA Piper DE) | Inteligo Media (C-654/23, Nov 2025): ePrivacy is lex specialis over GDPR Art. 6 | Germany, GDPR and ePrivacy describe both | Open doctrinal question. It changes fine exposure, not whether sending is lawful. |
| ePrivacy Regulation withdrawal | Announced 11 Feb 2025 (Hunton); DLA calls it "officially withdrawn" that day (DLA SE) | OJ publication 6 Oct 2025 (nicfab) | GDPR and ePrivacy | Announced Feb, formally published Oct. Either way the 2002 Directive still governs. |
| Digital Omnibus date | Proposed 19 Nov 2025 (Inside Privacy) | "February 2026" (OnTap) | GDPR and ePrivacy | Use 19 Nov 2025. |
| Sweden B2B | Marketing Act rules apply "to natural persons as well as to legal entities" (DLA SE) | Only applicable to B2C (ICLG) | EU/EEA country matrix, Cold email law: the jurisdiction matrix treat Sweden as opt-out to legal persons | DLA likely refers to the identification rule. Opt-out for companies stands. |
| KASPR fine | €240,000, CNIL decision of 5 Dec 2024 (CNIL) | "Widely quoted… CNIL page could not be fetched"; research note dates the action Jan 2025 via LeadGenius | France, Data sourcing law, GDPR and ePrivacy cite the CNIL decision; Contact data and enrichment flags it as unverified | The law pages cite the CNIL decision itself. Remove the flag on Contact data and enrichment. |
| Japan individual fine | ¥500,000 (e-Gov fetch summary, Act) | ¥1,000,000 for some offences (researcher's recollection) | Australia, New Zealand and APAC avoids the figure | Correctly omitted. |
| Verkada CAN-SPAM count | 3 violations (WilmerHale) | Four ways (FTC) | United States: CAN-SPAM and state email laws | Immaterial. Use the FTC. |
Product and tech facts
| Topic | Claim A (source) | Claim B (source) | Which page uses which | Our read |
|---|---|---|---|---|
| Instantly MCP scope | 31 tools (Campaigns 6, Leads 11, Email 5, Analytics 3, Accounts 6) (help) | Developer docs describe the "full V2 API" (Reply.io review) | AI campaign builder, API, webhooks and MCP, Built-in lead database, Campaign analytics all use 31 | Pages agree. The help centre count is the one to cite. |
| Smartlead MCP tool count | "116+ tools", official, as of April 2026 (Smartlead) | 113 tools in LeadMagic's community smartlead-mcp-server (GitHub, research note) | Smartlead, API, webhooks and MCP, AI campaign builder all use 116+ | Two different servers. Pages are consistent; do not mix the two counts. |
| Instantly MCP launch date | Help article dated 25 Aug 2026 | Reviewed by Reply.io on 21 Aug 2026 | Not dated on hub pages | The help date is a revision date. Launch date unknown. |
| Instantly subsequences | Hyper Growth and Light Speed only (help) | "Advanced Sequences" ticked on Growth (pricing) | Conditional sequences and subsequences uses the help version | Possibly two different features. |
| Instantly ESP-matching fallback | "Proceeds with Gmail without waiting" | "May not reach your full daily limit" while waiting — same article | ESP matching | Behaviour unclear. Test it before claiming parity. |
| lemlist advanced conditions | Multichannel Expert or Enterprise only (help) | Listed across plans (pricing) | Conditional sequences and subsequences | Plan names changed. Pricing page is newer. |
| lemlist database size | 650M+ leads (pricing) | 450M+ (MCP docs) | Contact data and enrichment, Sequencing & sending use 650M; Built-in lead database shows both | Marketing numbers. Pages quoting 650M should add "claimed". |
| Smartlead Salesforce | Native integration (Smartlead API page) | "No native Salesforce integration" (ZoomInfo review) | Smartlead, CRM sync | Vendor docs win over a competitor's review. |
| Smartlead rate-limit plan names | "Standard / Pro" (rate-limit doc) | Plans are Base / Pro / Unlimited (pricing) | API, webhooks and MCP | Stale docs. |
| EmailBison webhook retries | "Up to 5 retry attempts over 24 hours" (GTM Directory) | No retry count in official docs | API, webhooks and MCP | Unverified. |
| HeyReach | Enrichment credits "per sender" (pricing); executives' profiles "restricted" (HeyReach); G2 4.7 | One-time credits, founder "permanently" banned, G2 4.6 (Valley; Valley) | HeyReach | Valley is a competitor. Prefer HeyReach's own pages, flag the ban scope. |
| Clay Sequencer status | GA 1 Sept 2026 (Clay blog) | A "beta" capped at 4 emails (Deepline, undated) | Clay, The AI shift use GA; GTM engineers says launch date unknown | Deepline is likely older. Single-source GA date. |
| Instantly localisation | French-language URLs indexed (instantly.ai/fr/partenariat-commercial) | English-only selector on the fetched page (instantly.ai/fr) | Non-English markets says English-only | Partial localisation at most. Check before positioning on language. |
| GitHub stars | Warmbly 22 / 34 on topic pages (cold-email, email-warmup) | 267 on the repo page; OpenOutreach 2.4k vs 3.0k | Open source in cold email flags it | Cache dates differ. Treat as approximate. |
| Leadbird ↔ EmailBison | Assumed relationship (research brief) | Leadbird's documented tool is Smartlead (Salesforge review) | EmailBison corrects the brief | No public link exists. |
Pages that disagree with each other: fix list
| # | Issue | Pages to fix | Align to |
|---|---|---|---|
| 1 | German B2B email needs "prior or presumed consent" | Founders and small B2B teams, Non-English markets | Germany: express prior consent; presumed consent is phone only |
| 2 | Smartlead API floor at "$174+" | GTM engineers | API, webhooks and MCP, Smartlead, CRM sync: Pro ($94) and up |
| 3 | Clay Launch $185 / 2,500 credits, Growth $495 | Contact data and enrichment | Clay, AI personalisation: $167 / 3,000 credits (clay.com, Oct 2026) |
| 4 | Microsoft "rejects rather than junks" | Provider crackdowns, History of cold email software | Microsoft rules: Outlook.com and Microsoft 365 limits: policy says Junk first; observed 550 5.7.515 |
| 5 | MERRL "dormant rather than dead" | Provider rules: Google, Microsoft and the ESPs | Microsoft rules: Outlook.com and Microsoft 365 limits, Provider crackdowns, OAuth verification and the end of basic auth: cancelled Jan 2026 |
| 6 | KASPR fine flagged unverified | Contact data and enrichment | France, Data sourcing law: CNIL decision, 5 Dec 2024 |
| 7 | lemlist "$28M ARR (2024)" | Non-English markets | lemlist: $28M in Jan 2025; ~$50M claimed 2026 |
| 8 | lemlist "$53M revenue 2026, $40M in 2025" | Market size | Label as GetLatka estimate next to lemlist's founder-claimed ~$50M |
| 9 | Instantly price change "could not be dated" | Pricing landscape | Instantly, History of cold email software: Aug 2026 per IndustryLens |
| 10 | Per-mailbox norm of 20–50/day | Sending architecture | Sending limits and throttling, Volume math: what 10k and 100k emails a day cost: 10–30/day |
| 11 | Single-number reply benchmarks (3.43%, ~8.5%, 1–3%) | 11x, Instantly, AI SDR layer, Replies, booking and handoff, Campaign analytics, Copy and personalisation layer, Signals layer | Add the 0.45–3.7% range and link Reply benchmarks |
| 12 | lemlist "650M" database without the 450M MCP figure | Contact data and enrichment, Sequencing & sending | Built-in lead database: "650M+ claimed; 450M+ in its own MCP docs" |
| 13 | Smartlead workspace "$29" only | Business models, White-label and agency portals, Workspaces, roles and SSO | Smartlead: $29–39 depending on the page |
| 14 | lemlist funding "None" vs GetLatka "$30M" | none (both pages flag it) | Keep "no primary source" wording on Funding and exits |
Which contradictions matter for the decision
Most rows above are noise: founding years, review counts, a few dollars on a competitor's add-on. Six change what an EU entrant should build or how it should price.
German consent for B2B email. Two hub pages say German B2B email works with "presumed consent". It does not; one email to a business without express consent is actionable (Germany). This decides whether a German-founded product can sell to German senders at all, or must be built for export plus a consent-first workflow (EU-native compliant outbound, Governed volume for sales teams).
Instantly's real scale and price direction. Instantly is somewhere between $20M (reported, 2024) and $38M (estimate, 2026), with 40–50k claimed customers on a $47 entry plan that just rose 27%. The upper figure is uncorroborated. The direction matters more than the level: the leader is raising prices and bundling, not cutting (Pricing strategy, Cloning the leaders).
Reply benchmarks. With published averages from 0.45% to 3.7% for the same year, no customer promise, kill criterion or ROI model can rest on "the industry average". This is both a risk (your own targets in Kill criteria) and an opening for methodology-explicit, in-product benchmarks (Trust as positioning).
Microsoft enforcement detail. Whether non-compliant Outlook.com mail is junked or rejected, and whether per-mailbox limits are coming back, sets how much a Microsoft-heavy infrastructure strategy can be trusted (Infrastructure strategy: build or partner, Microsoft-first outbound). The evidence favours "rejected from day one" and "MERRL cancelled", but Microsoft has promised "adaptive" limits with no date.
CASA and Gmail scope classification. The spread from $540 to $6,000 a year (and an unverified $15k–75k) is small money. The point is that restricted scopes and annual assessment are certain. That makes reply-reading via Gmail API a compliance project, which shapes Sending architecture and OAuth verification and the end of basic auth.
Smartlead's revenue and pricing gates. Whether Smartlead is a $14M or $20M+ business, and whether its API sits at $94 or $174, decide how much room exists for an agency-first, API-first challenger (Agency operating system, Sending engine as an API). The fix to GTM engineers matters because that page argues the API gate is "the easiest feature to beat".
Resolved on 2026-10-05
| Page | Old claim | New claim |
|---|---|---|
| Founders and small B2B teams, Non-English markets, Cold email law: the jurisdiction matrix | German B2B email needs "prior or presumed consent" | Express prior consent; presumed consent covers phone only (Germany) |
| GTM engineers | Smartlead API gated at "$174+" | Pro ($94) and up (pricing.md, Oct 2026) |
| GTM engineers | Clay Sequencer launch date unknown | GA 1 Sept 2026 per Clay blog (single source) |
| Contact data and enrichment | Clay Launch $185 / 2,500 credits, Growth $495; flag on credits | Launch $167 ($150 annual) / 3,000 credits, Growth $446 ($401 annual) per clay.com; flag removed |
| Contact data and enrichment | KASPR fine flagged unverified; action dated Jan 2025 | €240,000, CNIL decision of 5 Dec 2024, linked |
| Contact data and enrichment, Sequencing & sending | lemlist "650M" database | 650M+ claimed; 450M+ in its MCP docs |
| Provider crackdowns, History of cold email software | Outlook "rejects rather than junks" | Policy: Junk then possible reject; observed: 550 5.7.515 from day one |
| Provider rules: Google, Microsoft and the ESPs | MERRL "dormant rather than dead" | Cancelled 6 Jan 2026 (Office365itpros) |
| Non-English markets | lemlist "$28M ARR (2024)" | $28M in Jan 2025; ~$50M founder-claimed 2026 |
| Market size | lemlist "$53M revenue" unlabelled | Labelled GetLatka estimate beside founder-claimed ~$50M |
| Pricing landscape | Instantly price change "could not be dated" | August 2026 per IndustryLens; $37 is the pre-rise monthly price |
| Sending architecture | 20–50 sends/day per inbox | 10–30/day |
| Business models, White-label and agency portals, Workspaces, roles and SSO | Smartlead workspace $29 | $29–39 |
| 11x, Instantly, AI SDR layer, Replies, booking and handoff, Campaign analytics, Copy and personalisation layer, Signals layer, In-house SDR teams | Single reply-rate average (3.43%, ~8.5%, 1–3%, 5–6%) | Same figure plus the 0.45–3.7% range and denominator caveat, linked to Reply benchmarks |
Not changed: no current page states Instantly Growth at $37/mo unqualified ($37 appears only as the 2024 price or the $37.60 annual price). The "presumed consent" and "$37/mo" examples in AUTHORING.md are outside content/ and were left as is.