Lead response and routing

What this problem is

Lead response and routing is the work of noticing a new enquiry, deciding who should own it, and sending a first useful reply. It is not marketing. It is operations: a repeating handoff between channels, people, and the CRM.

Is this your problem?

• Enquiries sit in a shared inbox until someone happens to look
• The same lead is contacted by two people, or by nobody
• First replies are delayed after hours, at lunch, or on busy days
• The CRM record is created after the conversation, if at all
• Managers cannot see response time without exporting a spreadsheet

Current workflow

A typical 10–50 person service firm receives enquiries through website forms, email, phone, and sometimes LinkedIn or a portal. A coordinator or salesperson opens the inbox, reads the message, looks the company up, decides who should take it, types a reply, and later copies the details into the CRM. Out-of-hours messages wait until the next working morning.

Where it fails

• No single queue: the same lead arrives in two channels and is treated as two leads
• Assignment depends on who is online, not on rules the team agreed
• Context (source, listing, previous jobs) is not attached to the first reply
• Exceptions — existing customers, junk, urgent tenders — use the same path as new leads
• Measurement happens, if at all, from memory rather than timestamps

How to measure the baseline

Pick a two-week window. For every inbound enquiry, record the timestamp it arrived, the timestamp of the first useful reply, who owned it, and whether it was a new opportunity, an existing customer, or junk. Do not use “we usually reply the same day” as a baseline. Use lead_response_minutes, cases_per_week, and conversion_rate_pct on the same cohort. Evidence should be labelled observed (system timestamps), reported (staff estimate), or unknown.

Metrics that matter

• Lead response time — Minutes from inbound enquiry to first useful human or system response. (minutes)
• Cases per week — Volume of discrete items (leads, invoices, tickets) in a typical week. (cases/week)
• Minutes per case — Average handling time from pickup to completion of one case. (minutes)
• Conversion rate — Share of cases that reach the intended commercial or operational outcome. (%)
• Exception rate — Share of cases routed to human exception handling. (%)

Root causes

• Ownership is informal: “someone in sales will see it”
• The CRM is a reporting tool, not the system of work
• There is no after-hours rule, so volume waits in the inbox
• Qualification criteria live in people’s heads
• Channel tools are not connected, so routing cannot be deterministic

What should be automated

• Capture every inbound channel into one queue with a source stamp
• Create or match a CRM record before a human replies
• Apply routing rules the team can explain (region, service line, round-robin)
• Send a confirmation that the enquiry was received, without pretending it was qualified
• Flag obvious junk and existing-customer messages for a different path

What should remain human-controlled

• Judgement on unusual or high-value opportunities
• The first substantive commercial conversation
• Exceptions: complaints, legal, existing accounts, anything that looks sensitive
• Changing routing rules when the team or offer changes

Example intervention architecture

Illustrative only — not a client result. Intake from form and shared inbox writes a structured record. Matching looks up the CRM on email and company name. Rules assign an owner or a rota. A short acknowledgement can go out automatically. Anything below a confidence threshold, or marked sensitive, waits in a human queue. The CRM remains the system of record; the inbox is not.

Failure modes

• Auto-replies that sound like a salesperson and damage trust
• Wrong owner assigned because the rules are coarser than the business
• Duplicate records when matching is too weak
• Silent failure: the automation stops and nobody notices for a week
• Leads marked qualified by a model with no review

When not to automate

• Volume is a handful of enquiries a week and one named owner already responds fast
• The offer, territories, or team change every month — rules will rot
• There is no CRM, or the CRM data is too poor to match against
• Enquiries are mostly relationship-based and arrive through personal inboxes you cannot govern
• Expected time saved is smaller than the cost of connecting the tools and training the team

Related problems

• Lead follow-up

Sources

• The Short Life of Online Sales Leads (Harvard Business Review, 2011)
• ICO guidance for organisations (UK data protection)