CRM Signal Routing: From Buyer Event to Rep in Under a Minute

A target account visits your pricing page three times in a week. Two different people from that company download the integration guide. Their VP of Sales started six weeks ago. Every one of those facts is knowable, and in most B2B companies not one of them reaches the rep who owns the account until the account fills in a form, if it ever does.

Routing is the part of the system that closes that gap: the rules that turn a scored signal into a task, on the right desk, with the context attached, fast enough to matter. It is unglamorous, it is where most of the pipeline leak lives, and it is almost entirely a configuration problem once the signals exist in one place.

Why round-robin is the wrong default

Most CRMs route new leads by rotation or by territory. Both answer "who gets this" without asking "does this deserve anyone's time right now". A form fill from a student and a third pricing visit from a target account's economic buyer are routed identically. Reps learn that the queue is mostly noise and stop working it promptly, at which point the signals that were real go cold.

Routing has to be driven by score, not by arrival. And score has to be two numbers, not one.

Fit and intent, separately

Fit is how well the account matches the ICP: industry, size, geography, tech in use, roles present. It changes slowly and comes from enrichment.

Intent is what the account is doing right now: pricing visits, document downloads, replies, usage thresholds, event attendance. It changes daily and decays fast.

A single blended score makes a perfect-fit account with no activity look like a poor-fit account browsing constantly. Two scores make the right action obvious:

	Low intent	High intent
High fit	Nurture, watch triggers, awareness audience	Rep now, with the timeline, within the hour
Low fit	Nothing	Investigate; the ICP may be wrong

The bottom-right cell is where ICP learning comes from. Do not discard it. The Lead Scoring Ruleset has the full model including decay and caps.

Where the signals come from

Routing is only as good as what it can see. The CRM natively sees form fills and email opens, which are the weakest signals available. Everything stronger lives elsewhere:

• Web behaviour, resolved to an account, from a first-party signal ledger.
• Product usage, from the application database.
• Outbound replies, from the sending tool.
• Ad engagement, from the platforms.
• Hires and funding, from a data provider.

The integration work is getting all of these into one place with one schema and one account key, then writing the resulting scores and the triggering events back to the CRM record. The ledger does the first part; a sync does the second. Once the CRM record carries fit_score, intent_score, last_signal and last_signal_at, every routing rule can be a simple workflow.

The routing table

Written down, before it is configured, as rows:

IF   fit >= 70 AND intent >= 60 AND no open opportunity
THEN create task for account owner, priority high, due in 2 business hours,
     attach last 5 signals, attach drafted first touch

IF   fit >= 70 AND intent < 60
THEN add to nurture segment "high-fit-quiet", add to awareness audience,
     set trigger watch on pricing.viewed, careers.role_posted, funding.announced

IF   fit < 40 AND intent >= 60
THEN create task for RevOps: review ICP fit, no outreach

IF   any reply to outbound
THEN stop all sequences for account, exclude from retargeting,
     notify owner, log signal

IF   open opportunity exists
THEN suppress from sequences and paid audiences

Every rule has an owner and a response time. A rule with no response time is a dashboard tile. The response time is what makes it routing.

Suppression is half the system

The rules that stop things are as important as the rules that start them, and they are the ones nobody writes:

• An account in an open opportunity is never enrolled in a sequence and never in a paid audience. Spending to reach someone a rep is already talking to is waste, and getting a cold email while negotiating is worse than waste.
• An account that replied, positively or negatively, has every sequence stopped everywhere within minutes, not at the next daily sync.
• An unsubscribe anywhere is an unsubscribe everywhere. One list honouring it and three ignoring it is a compliance failure, not a data quality issue.
• A customer is excluded from acquisition motions and included in expansion ones, which are different sequences with different owners.

Two motions colliding on one account is almost never a strategic disagreement. It is a missing suppression rule.

What arrives on the rep's desk

Not "Lead assigned". A task that contains: the account, its fit reasons, the signals that fired and when, the people involved with roles, the observable fact worth referencing, and a drafted first touch that references it. The Account Research Brief is the format. Agents assemble this well; the rep's job is to read, adjust and send.

Then the outcome writes back: sent, replied, meeting, disqualified, with a reason code from a closed list. That is what lets the scoring model learn and the routing table improve. A routing system that does not record outcomes is a routing system that will be exactly as good in a year as it is today.

Where it goes wrong

The source field gets overwritten. Most CRMs default to updating "source" on every touch, which destroys channel reporting silently. Set it once, protect it.

Fields do not survive conversion. The click ID and the scores are on the lead and vanish when the lead becomes an opportunity. Check ten opportunities.

Response times are aspirations. Two hours is written down; the median is two days. Measure it, show it on the sales lead's screen, and treat a miss as a process failure with a name on it.

The table is never revisited. Thresholds set in January are wrong by June. Review the routing table against outcomes quarterly, alongside the scoring weights.

The engineering spec configures the CRM routing section of an outbound engine in a few clicks and produces the blueprint we build from; the same logic applies whichever CRM you run.

Frequently asked questions

What is CRM signal routing?

CRM signal routing is the set of rules that turn a scored buying signal, from the web, the product, outbound replies, ads or external data, into a specific action in the CRM: a task for the right rep with context attached, a nurture segment, an audience, or a suppression. It replaces round-robin and territory assignment, which route by arrival rather than by whether the account deserves attention now.

Why should fit and intent be scored separately?

Because a blended score cannot distinguish a perfect-fit account with no activity from a poor-fit account browsing constantly, and those need opposite actions. Two scores feed a routing matrix: high fit and high intent goes to a rep immediately; high fit and low intent goes to nurture with trigger watches; low fit and high intent is investigated because the ICP may be wrong.

How do signals from outside the CRM get in?

Through a first-party signal ledger that receives web, product, outbound, ad and external events under one schema and one account key, computes scores, and syncs the scores and the triggering events back onto the CRM record. Once the record carries fit, intent and last signal, routing rules are ordinary CRM workflows.

What suppression rules does routing need?

Accounts in open opportunities are excluded from sequences and paid audiences; any reply stops every sequence for that account within minutes; an unsubscribe anywhere applies everywhere; customers are excluded from acquisition and included in expansion. Most collisions between motions are a missing rule here, not a strategy problem.