Homechevron_rightFeatureschevron_rightICP Matching

Feature 03 · ICP Matching

ICP filters tell you the size. We tell you who'll buy.

Apollo, ZoomInfo, and Sales Navigator return 8,400 firmographic matches. None of them know which of those will respond, take a meeting, or close. We give you 284 contacts scored above 80 on a 100-point conversion model, with the specific signals that drove each score.

See ICP scoring live

0-100

Conversion score on every contact

Signal categories evaluated per account

24-72h

Score refresh cadence

4-5×

Reply rate vs un-scored firmographic outbound

03ICP Matching

ICP filters tell you the size. We tell you who'll buy.

Apollo returns 8,400 firmographic matches. We return 284 scored above 80 on a 100-point conversion model. AI agents research each account for buying signals, funding, hiring, tech stack, leadership changes, and score every contact. Below threshold, skipped. Above, prioritized.

Independent account research per contact, not just firmographics.
0 to 100 conversion score, refreshed every 24-72 hours.
Auto-skip low-fit prospects. Protects your sender reputation.

How the score is built

app.reachiq.ai

88/ 100

apartment Company Analysis

Andrea Knox, Northwind Strong Match

analytics

Criteria evaluated

check_circle

Criteria matched

7 / 8

Evaluation criteria 8

Matched Not matched

check

Industry

Operates in SaaS / Financial Services. Multiple verified sources confirm fit with the target industries criterion.

90% Must have

check

Company Size

Mid-market band at ~150 employees, matching the 50 to 500 headcount range defined in the ICP.

85% Must have

check

Revenue

Estimated ARR in the $30M to $60M range, above the $20M threshold required; medium confidence on exact figure.

80% Nice to have

check

Regions

Headquartered in San Francisco with strong North America and Europe operations. Matches regional criterion fully.

90% Must have

How the score is built

Eight signal categories, weighted by what closes.

Each signal is weighted by how strongly it correlates with actual conversion across millions of B2B emails. A funding announcement is worth +18 points. A vague intent visit is worth +2.

manage_search01

Account discovery

Contact enters your ICP via search, CSV upload, or auto-discovery. AI agents kick off independent research per account.

insights02

Signal scan

LinkedIn activity, news mentions, funding rounds, M&A, leadership changes, tech stack shifts, press, podcast appearances, G2 review activity.

tune03

Weighted scoring

Each signal contributes points based on its measured correlation with conversion. A fresh raise is +18. A Bombora visit is +2. Vanity signals are throttled.

crisis_alert04

Threshold gating

Above 80 is reach out today. 60-80 is nurture pool. Below 60 is skip unless something changes. Re-scored every 24-72 hours.

Real signals, not vanity signals

What we score, and what we throw out.

A lot of "intent data" out there is just visit data, someone from acme.com browsed a Bombora topic page. That's noise. We weight situational signals that correlate with real buying motion.

High-weight signals

What predictive lead scoring rewards.

Each carries real weight because each correlates with actual conversion in the campaign data we trained on.

Just raised funding. 60-90 day window of rapid scaling. +18 points.
Hiring SDRs or AEs. Revenue pipeline is the priority. +12 points.
Leadership change at VP+. New exec evaluates new tooling in their first 90 days. +14 points.
Tech stack change. Removed a tool means they're shopping. +10 points.
Product or feature launch. Sales and marketing ramping in parallel. +8 points.

Throttled signals

What gets down-weighted.

Signals that historically don't predict conversion get small weights or are filtered out entirely.

Generic intent topic visits. Often a vendor or analyst, not a buyer. +2 points max.
Old news mentions. Anything over 90 days is decayed in weight.
Cold LinkedIn likes. Engagement without context. Ignored unless paired.
Static firmographics alone. Size and title fit isn't enough on its own.
Company-level signals without role match. A raise at the wrong department won't lift a contact's score.

Threshold gating

The score doesn't just rank, it gates.

Three score bands, three different behaviors. The ICP score actively decides who gets emailed, who waits in nurture, and who gets skipped.

80-100

Reach out today.

Above-threshold prospects flow straight into the active sequence. Sorted by score so your SDR works the strongest first.

Enrolled in the next outbound sequence automatically
Prioritized in the daily SDR task queue
Slack signal alert fires when a watchlist contact crosses 80

60-80

Hold in nurture.

In-the-fit-window but no urgent buying signal. Held in nurture and re-scored every cycle. If a signal lands, they get promoted.

Lightweight content-led touches, not full sequence
Re-scored every 24-72 hours against fresh signals
Auto-promoted to active when score crosses 80

0-60

Skip unless something changes.

Not enough fit, not enough signal. Sender reputation gets protected, your team's time doesn't get burned on bad-fit replies.

Auto-removed from active sequences
Tracked but not contacted
Re-evaluated if a fresh signal flips the score

Questions

ICP Matching, explained.

Each contact's score is a weighted sum across eight signal categories: firmographic fit, role and decision authority match, recent funding, hiring activity, leadership changes, tech stack shifts, news and content signals, and engagement signals. Weights come from how strongly each signal correlated with actual booked meetings and closed deals in the campaign data we've trained on.

Every 24-72 hours, depending on signal volatility. High-velocity signals like a fresh funding round trigger an immediate rescore. Slower signals like tech stack changes get picked up on the regular cadence. You'll see the last refresh timestamp on every contact.

Yes. Default threshold is 80 for active outreach, 60 for nurture. You can move both. You can also flag specific criteria as must-have vs nice-to-have, which forces the score to gate on those rather than letting other signals compensate.

The base model is trained on cross-customer outbound data, segmented by industry, role, company stage, and region. So even on day one, you inherit the weights learned from millions of similar sends. As your own campaign data comes in, the weights tune to your specific motion.

Most B2B teams burn through their domains by emailing thousands of low-fit prospects. The bounces, unsubscribes, and spam complaints stack up and the domain reputation tanks. ICP-matched outbound to 200 above-threshold prospects produces 4-5× the reply rate of cold outbound to 2,000 firmographic matches, with a fraction of the deliverability damage.

Keep exploring