Speed to lead and routing

A demo request from a 900 person logistics company arrived at 10:14 on a Tuesday. The assigned account executive was in a pipeline review. The record sat in a Salesforce queue, picked up a templated email at 14:32, and got no reply. Six weeks later that company signed with a competitor whose form booked a call before the confirmation page had finished rendering.

Nothing in that sequence is a motivation problem. Nobody was lazy. The routing rules assigned a high-fit lead to one human being with no timer, no fallback and no alert, and then the CRM dutifully recorded a four hour response as though that were normal. It is normal. That is the whole problem.

Response time is the cheapest conversion lever in SaaS lead generation because it costs no media spend, no new content and usually no new headcount. It is also the one most teams never measure correctly.

What the speed to lead research actually says, and what has aged badly

The canonical numbers come from two studies, both older than most people citing them realise. Harvard Business Review’s 2011 audit of 2,241 US companies found a median first response of 42 hours, with 23 percent never responding at all. The same research reported that firms contacting a prospect within an hour were roughly seven times more likely to qualify the lead than those waiting one more hour, and about sixty times more likely than those waiting a full day.

The second is the Lead Response Management study led by James Oldroyd, which analysed call logs rather than surveys and found the odds of reaching a lead falling off a cliff after the first hour. Drift’s 2021 test of 433 B2B companies is the most recent large replication: 7 percent replied within five minutes, and more than half took five business days or never replied.

Here is the honest caveat nobody includes. That research predates live chat, scheduling links, product-led signup and AI qualification agents, and it suffers from obvious selection bias. Companies that respond in four minutes tend to have better sales management, better CRM hygiene and better products. The response time is partly a proxy for competence.

So treat the 60x figure as a headline, not a forecast. The version you can defend in a board meeting is the one you calculate yourself: pull twelve months of inbound leads, bucket them by minutes between form submit and first human touch, and compare meeting-held rate per bucket. Every team I have watched run that analysis has found a real, ugly gradient. The size varies. The direction never does.

What a realistic median looks like in B2B SaaS

Most teams that have never instrumented this sit somewhere between one and six hours median, with a tail past 24 hours covering evenings, weekends and anything routed to a person on holiday. Teams running form concierge tooling on their qualified slice report medians under two minutes, because a machine does the routing and the prospect books themselves.

The table below reflects patterns from inbound audits rather than a published dataset, so treat the percentages as typical rather than precise.

Median first human touch	Usual cause	Typical effect on meeting held rate	First fix
Under 5 minutes	Qualify and book on the form itself	Best band, use it as your baseline	Monitor it so it does not decay
5 to 60 minutes	Round robin plus a Slack alert that people watch	Roughly 10 to 25 percent below the best band	Add a pooled fallback queue
1 to 4 hours	Assigned to one named owner, no timer	Often around a third lower	Reassign automatically after 10 minutes
4 to 24 hours	Batch CRM sync plus daily list work	Roughly half	Fix the sync before coaching the rep
Over 24 hours	A shared inbox nobody owns	A fraction of the best band	Assign an owner this week

Two details matter more than the headline. First, the distribution is bimodal in almost every company: a fast cluster during working hours and a dead zone that starts at 17:30 and ends at 09:00. Second, your worst response times land on your best leads surprisingly often, because enterprise-fit leads get routed to a named senior AE who is in back-to-back calls, while the small stuff goes to an SDR pool that actually has capacity.

If you are still running the lead quality argument in your weekly meeting, check the median first. A team debating lead quality while its median response time is over an hour is arguing about the wrong number. Slow follow-up degrades a good lead until it is indistinguishable from a bad one, and then you go and blame the source mix.

The routing architecture, in the order you should build it

Build it in this sequence. Skipping step one is why most routing projects produce duplicate records and angry account executives.

Keep the whole tree under about a dozen branches. I have inherited routing configurations with sixty nodes, three of which had been unreachable since a field was renamed in 2023, and nobody could safely change anything. Complexity in routing is a liability you pay for every time sales reorganises.

Covering evenings, weekends and timezones without hiring

You almost certainly have more coverage than you think. An SDR working 08:00 to 17:00 in London and one working 08:00 to 17:00 in New York already produce fifteen hours of continuous coverage, from 08:00 UK through 22:00 UK. Most teams never wire the routing to use it, because assignment rules are built around territory ownership rather than the clock.

Three patterns work, in descending order of how much I like them.

The first is follow-the-sun pooling for Tier A only. Outside a rep’s business hours, high-fit leads fall through to whichever region is awake, with a note in the Slack alert saying this is a cross-region assist and ownership transfers back at 09:00. Keep this to your top tier so nobody wakes up to forty reassigned records.

The second is a rotating first responder shift, one person per day, who owns anything unaccepted for ten minutes between 07:00 and 19:00. It is a light commitment, roughly what an on-call rota costs an engineering team, and it collapses the ninetieth percentile far more than it changes the median.

Weekends are the third case, and the answer is usually to not staff them. Saturday and Sunday typically carry three to eight percent of weekly inbound at B2B SaaS companies, which does not pay for a shift. Put a real scheduling link on the confirmation screen so the prospect books Monday themselves, write an auto-reply that says plainly when a human will respond rather than pretending to be one, and staff Monday morning heavier because the backlog stacks on top of the normal Monday peak.

How to measure true first touch instead of CRM task completion

You need three timestamps and most CRMs give you the wrong ones by default.

Timestamp one is the form submission, captured client side and written into a hidden field at submit. Timestamp two is record creation in the CRM. Timestamp three is the first genuine human contact: a dial with a connect or voicemail, a one-to-one email send, or a chat reply typed by a person.

The gap between one and two is where reported numbers quietly cheat. A Marketo or Pardot batch sync running every five to fifteen minutes means a team measuring from CRM creation is already understating response time by a quarter of an hour before anybody has done anything. Webhook the form directly into the CRM for Tier A and keep the batch sync for everything else.

Never measure task completion. A rep can mark “Call 1” complete without touching a phone, and in any team under quota pressure, some will. Measure against the activity record with a real dial duration or a sent message ID.

Report median and ninetieth percentile, weekly, split by tier, source and hour of submission. The mean is worthless here because a single Friday evening lead answered on Monday adds 60 hours to the pile. Put the two numbers on one chart, and when they diverge you are looking at a coverage gap rather than a habit problem.

An escalation ladder for high fit leads

Speed applied evenly is waste. Calling a student researching a dissertation within ninety seconds teaches your reps that the alerts are noise, and once that happens the alerts stop working for the leads that matter. Tiering is what makes urgency sustainable.

The escalation ladder above works only if somebody owns the misses. Publish response time by rep weekly, unedited, in the same channel as the alerts. Nobody enjoys this for the first fortnight and then it stops needing enforcement.

Codify the tiers and the timers in writing. If you do not already have one, the sales and marketing SLA template covers the response commitments both sides sign up to, and the lead scoring template is where the tier definitions themselves should live so that routing and scoring do not drift apart.

Which routing tool solves which problem

Buy by problem, not by category. Most teams overbuy here because a vendor demo makes complex routing look inevitable.

Category	Representative tools	Problem it solves	Budget signal	Pick it if
Native CRM assignment	Salesforce assignment rules, HubSpot workflows	Territory and simple round robin	Included in your existing licence	Under roughly 300 inbound leads a month and one clean segment
Lead to account matching and routing	LeanData, Default	Dedupe, account matching, deep trees, SLA timers, audit trail	Low five figures to well into six, depending on seats	Salesforce shop with ABM, overlapping territories or partner routing
Form concierge and instant booking	Chili Piper, RevenueHero, Calendly routing forms	Qualify and book the meeting on the form itself	Per seat, typically tens of dollars a seat a month	Your demo form already hands off to a scheduling step
Enrichment at submit	Clearbit, Clay, Apollo	Fills firmographics so routing rules can branch on something true	Usage based, varies enormously	You route on company size, industry or country

My default recommendation for a Series A company doing 200 to 600 inbound leads a month: native assignment rules plus a form concierge product, and nothing else. That combination delivers most of the available gain for a few hundred dollars a month. Move to LeanData or Default when you have territories that overlap, a partner channel, or an ABM motion where account ownership genuinely conflicts with round robin. There is a fuller breakdown in the guide to lead capture and routing tools for SaaS.

What this costs, and where it fails

The direct cost is modest. Tooling for a mid-sized team runs somewhere between 500 and 4,000 dollars a month depending on the stack. The hidden cost is RevOps maintenance, realistically half a day to a day a month once things are stable, plus a painful fortnight when sales reorganises territories.

Three failure modes show up repeatedly.

Speed without qualification is the most common. Teams turn on instant alerts for every form fill, reps chase unqualified traffic for three weeks, trust in the alert dies, and response times revert while the dashboard still shows the rule as active. Tiering prevents this. So does keeping Tier D genuinely out of the calculation instead of quietly inflating your numbers with it.

The second is the auto-reply that counts as a response. An automated acknowledgement sent in twelve seconds is useful, and it is not a first touch. If your reporting counts it as one, your speed to lead metric is decorative.

The third is diminishing returns. The move from four hours to five minutes is a genuine revenue event. The move from five minutes to ninety seconds is mostly a dashboard event, and the engineering effort required for that last stretch is better spent on conversion rate or on the cost per lead side of the equation. Know where your gains stop.

One more honest note. Fixing response time raises meeting-held rate, which raises pipeline, which raises the number of opportunities your AEs must work. If capacity is already the constraint, the gain shows up as longer sales cycles rather than more closed revenue. Model it with the lead value calculator before you promise a number, and sanity check the downstream assumptions against SaaS lead conversion benchmarks.

The audit to run this week

Do the audit before you buy anything. In most companies the first three findings are a broken sync, a missing fallback timer and a tier of leads assigned to somebody who left in June. Those are free to fix, and they usually account for the bulk of the gap.

Once the numbers are clean, set a target that matches your ACV. High-velocity self-serve businesses should push for instant booking on the form. Enterprise teams with a 90 day cycle can live with fifteen minutes and should spend the saved effort on multi-threading. Then size what the improvement is worth using the lead goal calculator, and check the economics against your current cost per lead so that you are comparing a routing fix against the alternative of simply buying more leads. Nine times out of ten the routing fix wins, and it wins this quarter rather than next year.