How to set up lead routing in a CRM

Before you touch a routing rule, list every channel that creates a lead record. Demo requests, pricing-page forms, content downloads, chat conversations, trade-show scans, partner referrals, inbound calls, outbound-sourced meetings, and third-party intent providers all behave differently and deserve different routing paths. For each source, classify intent on a simple three-tier scale: high (buyer explicitly asked to talk to sales), medium (buyer engaged with mid-funnel content or a free trial), and low (buyer gave contact info in exchange for something, like an ebook, with no stated intent). High-intent leads must route in minutes with a tight SLA. Medium-intent routes to a nurture-plus-sales blended queue. Low-intent leads belong in marketing automation until behavior upgrades their tier. If you skip this classification step, every routing rule you build downstream will have to carry the same emergency weight, and your reps will burn out chasing content downloads with the same urgency as demo requests. Document the inventory in a shared sheet before writing a single rule. Review it with a rep and a marketer in the same room so neither side gets to grade their own homework.

• Build a two-column list: lead source on the left, intent tier (high, medium, low) on the right
• Note the typical daily volume per source so you can predict capacity load before go-live
• Flag any source where volume spikes seasonally; routing capacity has to flex with it
• Mark sources that produce disqualified leads frequently so you can short-circuit them in step three

The four main assignment models are round-robin, weighted round-robin, territory-based, and account-based, and each one solves a different problem. Round-robin distributes leads evenly across a pool of reps, which is simple and fair when every rep is interchangeable. Weighted round-robin lets you tilt the distribution toward reps who are ramped, have capacity, or hold a higher quota. Territory-based routing assigns leads by geography, industry, or company size, which matters the moment you have specialists. Account-based routing checks whether the lead belongs to a named account and sends it to the owning rep regardless of other logic. Most mature teams end up with a layered rule: account-based first, then territory-based, then weighted round-robin inside each territory, with a plain round-robin as the fallback. Simpler teams can start with just territory-plus-round-robin and add weighting after thirty days of real data. The mistake to avoid is picking a model because another company uses it. Pick based on how your reps actually specialize today, and plan to revisit the choice once a quarter.

• Map each lead tier from step one to its primary assignment model
• Decide whether weighting should reflect quota size, ramp status, or current open-pipeline load
• Pick a single fallback rule for anything that does not match a territory or account owner
• Document the decision tree in plain English before translating it to CRM rules; it will save you a day of debugging

Territories exist to give every rep a predictable book of business and to keep two reps from showing up on the same lead. Encode them in the CRM using the same dimensions you used in step one classification: country, state or region, industry, company-size band, and segment tier. Resist the urge to carve territories by rep name as a shortcut. Carve them by firmographics first, then assign reps to the firmographic slice. That way, when a rep leaves, you reassign the territory in one place and every downstream rule inherits the change. Build territories as a lookup table with mutually exclusive and collectively exhaustive rows, meaning no lead ever matches two territories and no lead ever matches zero. Edge cases that break this rule, like a prospect in two offices across two regions or a company whose industry does not fit your taxonomy, should route to a named triage queue rather than crash into the fallback. Review the territory table after the first week of production traffic. You will find at least three edge cases you did not anticipate, and the fix is almost always to add a row to the table rather than invent a new rule.

• Draft the territory table with firmographic columns, not rep names
• Verify every possible combination lands on exactly one territory row
• Create an explicit triage queue for ambiguous leads so edge cases do not fall into fallback silently
• Assign reps to territory rows last, so reassignments stay cheap when turnover happens

Round-robin within a territory keeps distribution fair across reps who share coverage. Weighting lets you reflect reality: a brand-new rep in week three cannot handle the same load as a tenured rep on quota, and a rep already carrying forty open opportunities should not get the same slice as one carrying twelve. Pick one or two weighting signals, not five. Common choices are ramp status (new reps get a lower weight for ninety days), current open-pipeline load (reps above a threshold get skipped until they close or disqualify something), and target quota (reps on larger quotas get a larger share of inbound). The weighting formula should recalculate daily, not in real time, to avoid thrashing. Build an audit view that shows the last fifty assignments with the rep, the weight used, and the reason the system picked them. Reps lose trust in weighted systems faster than any other kind, so transparency is non-negotiable. If a rep asks why they just received a lead, the audit view must answer in one click.

• Pick no more than two weighting signals; three or more compound into noise
• Set the weighting refresh cadence to daily; real-time recalculation causes thrashing
• Expose an audit log of the last fifty assignments with the weighting reason for each
• Review weight parameters monthly and after any headcount or quota change

Every inbound lead needs a defined time window inside which the assigned rep must make a first touch, and the CRM has to measure it automatically. Industry research is consistent: the odds of qualifying a web lead drop by about half every five minutes after submission for the first hour, and by an order of magnitude after a day. Set your SLA accordingly. A reasonable baseline for high-intent inbound is five minutes during business hours and the next business morning outside them. Medium-intent leads can tolerate a four-to-eight hour window. Low-intent leads do not get an SLA; they get a nurture cadence. Define what counts as a first touch so the clock does not become theater. Minimum bar: a logged call attempt that reached a person or a sent personalized email, not an automated bounce. The clock starts the moment the lead is created and stops the moment the first touch is logged. Expose the countdown to the rep in their queue view so the pressure is visible, not hidden. Expose the breach rate to the manager in a weekly dashboard. The combination of a visible timer and a visible dashboard converts SLA from a policy document into actual rep behavior inside two weeks.

• Set a specific first-touch window per lead tier in minutes or hours, not vague terms like "fast"
• Define what counts as a first touch so the clock cannot be stopped with hollow activity
• Show a live countdown in the rep queue view and a breach dashboard in the manager view
• Set escalation rules: at seventy percent of the window, notify the rep; at a hundred percent, notify the manager and reassign

The hardest routing failures are the ones nobody designed for. A rep is on vacation. A rep leaves the company. A rep is in a demo and misses the five-minute window. A lead comes in at two in the morning on a Saturday during a product-hunt launch. For each of these, you need an explicit rule, not a hope. Vacation handling: reps toggle an out-of-office flag in the CRM, and the routing engine skips them in round-robin until the flag clears. Departures: an admin-only reassignment workflow transfers ownership of all open leads and opportunities to a successor on day one, with a running audit trail. Missed SLA: at the full window, reassign to the next rep in the territory round-robin and notify the manager. After-hours: route to a weekend-ready pool or hold until the next business morning with a system-generated acknowledgment email so the lead knows a human will respond. Each fallback should be documented in the same decision tree you built in step two. If a lead ever ends up unassigned for more than fifteen minutes, that is a bug in the fallback logic, not a one-off. Fix it in the rules, not with a manual reshuffle.

• Add an out-of-office toggle per rep and wire the routing engine to respect it
• Build an admin workflow that reassigns a departing rep in one action, with audit trail
• Define reassignment at the SLA breach moment, not a "soft" reminder that nobody enforces
• Decide whether after-hours inbound waits or routes to a weekend pool, and set expectations with the lead via acknowledgment email

Never flip routing on in production without running a full dress rehearsal. Build a test harness that generates thirty to fifty synthetic leads covering every tier, every territory, every firmographic combination, and every edge case you argued about in design. Watch where each lead lands. Verify the SLA clock starts correctly. Verify the audit log records the rule that fired. Verify fallbacks trigger when you deliberately break a condition, like marking every rep in a territory out-of-office. Rehearse the departure workflow on a test rep account. Rehearse an SLA breach by holding a lead past its window and confirming the reassignment and manager alert fire. Fix every bug you find before any real lead touches the system. The alternative is explaining to a rep why their first week of inbound went to a competing rep because of a rule you did not test. A thorough test harness takes an afternoon. Debugging live routing in front of frustrated reps takes a week and costs you political capital you do not get back.

• Create a synthetic lead generator that covers every tier, territory, and edge case
• Verify assignment, SLA start, audit log entry, and acknowledgment email for each test lead
• Deliberately trigger every fallback: out-of-office, SLA breach, missing firmographic data
• Keep the test harness checked in so you can replay it after every future routing change

Routing is never finished. Buyer behavior shifts, campaigns change the mix of inbound sources, reps leave, territories expand, and quotas adjust. Build a weekly routing health dashboard with four numbers: average time-to-first-touch by tier, SLA breach rate by tier, assignment-balance variance across reps in the same territory, and the volume of leads landing in the triage queue. Any one of these moving in the wrong direction signals a rule that needs tuning. Rising breach rates usually mean capacity shortage, not rep laziness. Growing variance in assignment balance usually means a weighting signal has drifted. Growing triage volume usually means a territory row is missing from the lookup table. Review the dashboard with sales operations and sales leadership once a week for the first quarter after go-live, then monthly after that. Treat routing the way a product team treats a shipping product: measured, iterated, and never finished. The teams that get routing right are not the ones with the smartest rules on day one. They are the ones who keep the rules honest for a hundred weeks in a row.

• Build a dashboard with time-to-first-touch, breach rate, assignment variance, and triage volume
• Review the dashboard weekly for the first quarter, monthly thereafter
• Change rules only on evidence from the dashboard, not on anecdotes from the loudest rep
• Snapshot the rule configuration every month so you can diff changes against outcomes later