Confession
Recon is not an enrichment vendor.
We don't sell B2B contact data. Buy fresh emails from Apollo or ZoomInfo and bring them in. Recon scores who's already in your CRM.
The lead
Find your real ICP.Cut the rest.
The closed-loop intelligence layer for B2B GTM teams running on HubSpot. OAuth in 60 seconds. First brief in 48 hours.
Built by an operator for the operator he used to be. EU-hosted, GDPR by default, audit log on every read.
The cost of the wrong ICP
The wrong ICP is a six-figure leak across every line of your GTM budget.
Four things burn every quarter you chase the wrong accounts: paid spend on the wrong segments, AE hours on the wrong deals, tooling for the wrong workflows, and the opportunity cost of the deals you should have been working instead. Forrester pegs wasted marketing spend at 21% on bad data alone. Dixon’s JOLT research shows 40 to 60% of qualified pipeline dies to “no decision” because the account was never the right fit. Recon names which segments to cut, which to double down on, and which campaigns to pause, cited to your closed-won.
- 40-60%
- of qualified pipeline is lost to “no decision” because the account was never a real fit
- 21%
- of marketing budget the average B2B team wastes on bad data and poor targeting
- 67%
- more likely to exceed quota for sales teams that operate from a disciplined ICP
Matthew Dixon, The JOLT Effect (Challenger), 2022
Forrester Research, sustained benchmark
TOPO / Gartner sales analytics research
Run the numbers for your team.
Three inputs, a slider, the answer. No email gate, no demo wall.
Ads, events, sponsorships, content syndication
£
Loaded at £130k / rep / year
Outreach, ZoomInfo, Apollo, ABM tools, etc.
£
30%
Industry research consistently puts the off-ICP share at 25-50% of B2B pipeline. Slide left if you trust your targeting, right if you do not.
Estimated annual waste
£174,000
That is what you are spending today on customers who will never close.
Wasted paid spend (ads + events)£45,000
Wasted sales hours£117,000
Wasted tooling licences£12,000
With Recon
Recon Team is £5,988 / year. Your potential recovered spend is £168,012. That is roughly 29x return on the subscription.
Methodology: loaded cost per UK B2B SaaS sales rep blended at £130k/year (SDR and AE midpoints, fully loaded with employer NIC, benefits, OTE at plan, ramp). Off-ICP wastage applied proportionally across paid marketing, headcount time, and tooling. ROI is computed against the Recon Team plan at £499/month annualized. Adjust inputs to match your reality.
Architecture
The closed loop, in one diagram.
Sources flow in. Recon classifies, derives, recommends, briefs. Every recommendation books a verdict in the outcomes ledger, and the verdict sharpens the next call. The recon_* properties write back into HubSpot, opt-in and scoped, so reps see the call inside the tool they already use.
Sources
- HubSpot CRM
- Apollo
- GA4
- Google Ads
- LinkedIn Ads
- Meta Ads
- n8n / CSV
OAuthwebhooks
Recon, EU-hosted
Classify · derive · recommend · track
Supabase EU (Frankfurt)
Claude Sonnet 4.6 + Haiku 4.5, EU regions where supported
- · 17-class source classifier (Haiku)
- · Recommendation outcomes ledger (the closed loop)
- · Audit log on every Recon AI call
- · AES-256-GCM tokens · RLS per workspace · SOC 2 in progress
opt-inwrites
Outputs
Brief
Ranked calls each morning, every claim cited
Outcomes ledger
30/60/90-day verdicts on every recommendation
Slack alerts
Drift, emerging segments, fix-confirms
Track Record
Audit-grounded ROI, every call dated
recon_* properties
Writeback into HubSpot, opt-in and scoped
Recon MCP
MCPQuery your ICP from Claude, Cursor, Windsurf
Status: status.get-recon.com
SLA: 99.9% uptime target on Team and Scale
Sub-processors: published list
Compared
Status quo vs Recon
Your options today: an annual ICP workshop, a bespoke warehouse, an enterprise ABM platform like 6sense or Demandbase, or an AI copilot from your CRM vendor. The first three go stale within months. The copilots fire suggestions and never check whether they worked. Recon is the one that closes the loop.
| Capability | DIY / 6sense / Demandbase / Champify | Recon |
|---|---|---|
| Time to first useful answer | 8-12 weeks of BI work, or 6-month ABM rollout | First brief within 48 hours of connect |
| Sources unified out of the box | CRM only, or a third-party intent layer bolted on | HubSpot, ad platforms, analytics, Stripe, in one place |
| Tracks whether recommendations worked | Einstein and Breeze ship suggestions and never check | 30/60/90-day verdicts on every call, audit-grounded |
| Writes intelligence back into HubSpot | No. Intelligence stays inside their dashboard | recon_* property writeback drives your existing lists, views, dashboards |
| Analyst with full GTM context | Generic Claude or ChatGPT, no CRM context | Reads every connected system before answering, cites every claim |
| Engineering required | Data engineer plus analyst, or a 6-week ABM onboarding | OAuth, then nothing |
| Refresh cadence | Quarterly review cycle | Continuous, off CRM and ad-platform webhooks |
| Seat-based pricing | Yes, every named competitor | No. Unlimited team members on every plan |
| Annual cost | £40k+ for the BI route, £80-200k for 6sense or Demandbase | From £299/month, no lock-in during early access |
6sense, Demandbase, and Champify are trademarks of their respective owners. We compete on positioning and pricing, not on intent data depth.
Confessions
Three things Recon is not.
We'd rather lose the call than ship a customer who expected something we don't do. The system is small on purpose.
Confession
Recon is not a pipeline forecaster.
We don't predict next quarter's bookings. We tell you which segments close and which drain your AE hours. Forecasting stays in your CRM.
Confession
Recon is not a multi-touch attribution platform.
We don't model 14 touchpoints across 9 channels. We attribute closed-won to the source that put it in motion. Single-touch, sourced, honest.
From the desk of Recon AI
Stop guessing who your best customers are.
Connect HubSpot. First brief lands within 48 hours, every recommendation cited to your real deals. Verdicts get tracked in the ledger. The model sharpens on your data, not someone else's.