Lead routing doesn't usually fail because the rule is wrong. It fails because of what happens in the thirty seconds before the rule ever runs, a missing enrichment field, a rep who left the company three weeks ago, a duplicate record splitting activity across two owners. The rule fires exactly as written and still sends the lead to the wrong place, because the rule was never the weak point. Everything feeding into it was.
This is the part most teams skip when they set up routing. They design the logic, round robin here, territory split there, and never build the layer around it that catches what the logic alone can't handle.
The Routing Methods and What Each One Actually Assumes
Every routing method makes a bet about what matters most for a given lead. Picking the wrong one for your motion causes problems no amount of clean data will fix.
Round robin. Simplest option, cycles leads evenly across a rep pool. It assumes every rep and every lead are roughly interchangeable, which is true for an SDR team working a single inbound queue and false almost everywhere else.
Weighted or capacity-based. Routes based on current pipeline load, quota attainment or seniority, not strict rotation. Fixes round robin's biggest flaw, a rep who just closed three deals shouldn't get the exact same next-lead priority as one sitting idle, at the cost of more logic to maintain.
Territory-based. Routes on geography, industry vertical or account size. Necessary the moment reps specialize, useless the moment your ICP data is incomplete, since a territory rule can't route what it can't classify.
Account-based. Routes the lead to whoever already owns the account, not whoever's next in a queue. Required for any team running ABM. It's also the method most likely to break silently, because it depends entirely on account ownership data staying current.
Where It Actually Breaks
Almost never at the rule itself. Three places, consistently:
- The fallback queue nobody watches. Every routing setup has a default bucket for leads that don't match any rule cleanly. That bucket is supposed to be a safety net. In practice it quietly becomes a graveyard, because nobody owns monitoring it and a lead that lands there can sit for days before anyone notices.
- Fields the rule depends on, missing at the moment routing runs. A territory rule needs company size. An account-based rule needs current ownership. If that field is blank when the lead hits the rule, the rule doesn't error. It just makes the best guess it can with incomplete information, and that guess is usually wrong.
- Ownership that's stale by the time the lead arrives. A rep changes territory, gets promoted or leaves the company. The routing rule still points at them until someone manually updates it, so the lead routes correctly according to the rule and incorrectly according to reality.
Our piece on CRM data hygiene goes deeper on why this data drifts in the first place. The fix here is narrower: build routing so it degrades safely when the data underneath it isn't perfect, instead of assuming it always will be.
Building SLA and Escalation Into the Routing Layer, Not Around It
Routing a lead to the right person is half the job. Making sure that person actually acts on it is the other half. It's the half most setups treat as a separate problem instead of part of the same system.
The research on this is old but hasn't been contradicted by anything newer. A 2011 Harvard Business Review audit of 2,241 companies found an average first-contact time of 42 hours, with 23% of companies never responding to inbound leads at all. A 2007 MIT study run with InsideSales.com across more than 15,000 leads found that contacting a lead within 5 minutes instead of 30 made a rep roughly 100 times more likely to actually reach that person.
That gap is what an SLA timer is for. The routing rule should start a clock the moment a lead lands with an owner. If it isn't claimed or actioned inside a set window, typically 5 to 15 minutes for inbound, it should escalate automatically to a manager, a backup rep or a round-robin re-route, not sit waiting for someone to notice.
Testing Routing Rules Before They Touch Real Leads
Most routing changes go straight to production, because testing them feels like overkill for "just a rule." It isn't. A single misconfigured territory rule can misroute every enterprise lead for a week before a rep thinks to check.
Two things catch this before it costs you pipeline. Run new rules in shadow mode first, log where a lead would have routed without actually routing it there, and compare that against what a human would have done for the same batch of leads. And treat fallback-queue volume as a live health metric, not an afterthought, a sudden spike means a rule just broke, usually within minutes of whatever change caused it.
FAQ
What's the simplest routing method to start with?
Round robin, if your team is genuinely interchangeable on skill and territory. The moment reps specialize by industry, region or account size, round robin starts actively costing you fit. It's worth moving to territory or account-based routing at that point, even though it's more setup work.
How fast should an SLA escalation actually trigger?
For inbound leads, 5 to 15 minutes before escalation is the range the response-time research supports. Outbound-sourced leads and lower-intent signals can tolerate longer windows, an hour or more, since the urgency isn't the same as someone who just filled out a form on your site.
Do we need a dedicated routing tool, or can native CRM rules handle it?
Native CRM routing (Salesforce assignment rules, HubSpot's workflow tool) handles round robin and simple territory logic fine. Account-based routing with real-time ownership lookups, weighted capacity logic or cross-object rules usually outgrows native tools and needs either a dedicated routing app or a custom-built layer on top.
If your routing setup is quietly losing leads and you can't tell where, get in touch and we'll trace exactly where it's breaking.