In this invented example, the marketing report says 1,200 inbound leads last quarter. The sales report says 310 of them were worked. Nobody can say what happened to the other 890, so the conversation drifts toward lead quality, and both teams leave certain the other side is the problem.
In this invented scenario, suppose neither is. Some of those 890 leads may have been lost in transit rather than rejected: a form that wrote to the wrong object, a record that matched an old closed-lost account and quietly merged into it, an assignment rule with no branch for a new region, a rep on holiday still in the round robin. This is lead leakage, the part of revenue leakage that happens before a human has touched the opportunity, and it is the cheapest leak to close because the buyer had already raised a hand.
The research on inbound response has not improved much in fifteen years. Harvard Business Review's 2011 audit of 2,241 US companies, by Oldroyd, McElheran and Elkington, found that 23 percent never responded to a web lead at all and that firms which did respond took 42 hours on average. In 2024 RevenueHero submitted demo and sales forms on 1,000 B2B SaaS websites and heard nothing back from 635 of them. In March 2026 Workato tested 114 B2B companies and found that only one sent a personalized email within five minutes, while the average personalized email took 11 hours and 54 minutes.
Those numbers matter because the buyer moves on. 6sense’s 2025 B2B Buyer Experience Report found that buyers make first contact roughly 61 percent of the way through their journey, that 79 percent initiate it themselves, and that 77 percent spoke first with the vendor they ultimately chose. The report says that first contact reflects a preference formed earlier, not that replying first caused the win. A missed inbound request means missing a conversation with a buyer who chose to initiate contact.
Lead leakage is a systems problem, not an effort problem. Telling reps to respond faster does nothing for the lead that never reached a rep.
Diagnosis: the five places lead leakage happens
Every inbound lead is captured, becomes a record, gets an owner, receives a first touch and reaches a decision. Leakage happens at each handover, and each has its own signature in the data.
1. The form that never becomes a record
The first leak is invisible to sales by definition. A form submission fails to create a usable CRM record: the form tool writes to a list but not the lead object, a sync job errors on a required field, a spam filter catches a real buyer using a personal address, or a deduplication rule merges the new submission into an old record owned by someone who left. Nothing alerts anyone, because nothing technically broke.
2. The record nobody owns
The record sits in a queue, with a default owner, or with the wrong person. Assignment rules were written for last year's territories, a new country or segment has no branch, or round robin keeps assigning to a rep who is out of office. In LeanData's 2017 State of Lead Management survey of 527 B2B sales and marketing professionals, respondents estimated that roughly one in four leads was routed to the wrong account owner.
3. Owned, but never touched
The lead has the right owner and still receives no first activity. This is the oldest leak in the funnel. In a 2013 Journal of Marketing study of 461 sales reps at four firms, Sabnis, Chatterjee, Grewal and Lilien described the "sales lead black hole," noting that about 70 percent of marketing-generated leads are never pursued by sales, and found that follow-up depends on prequalification quality, managerial tracking and how many leads reps are carrying. A rep with a full pipeline rationally ignores cold-looking inbound.
4. Touched, but too late
The lead gets a first touch, hours or days after the buyer filled in the form. In a separate study of 1.25 million leads, the authors found that firms which tried to contact a lead within an hour were nearly seven times as likely to qualify it as firms that waited even an hour longer, and more than 60 times as likely as firms that waited 24 hours or more. Workato's 2026 test found that companies using lead routing tools responded in about 3 hours and 32 minutes on average, against nearly 13 hours for companies without them. Late touches still count as "worked" in the CRM, so this leak hides in plain sight.
5. Disqualified without anyone deciding
The last leak is a decision nobody made. Enrichment returns no company match, a scoring threshold marks the lead as low fit, or a personal email address triggers an automatic rejection, and the lead is recycled to nurture without a human reading it. RevenueHero's 2024 study observed that 30.86 percent of the companies that never replied appeared to run an enrichment tool, and suggested that enrichment may have been used to screen the requests out. Some rejections are right; nobody can tell which, because no reason was recorded.
Framework: the Lead Leak Trace
The Lead Leak Trace measures inbound as five checkpoints and one clock. Each checkpoint is a yes-or-no question with a timestamp behind it, so every lead passed or stopped at a specific point.
| Checkpoint | Question | Evidence in your records | Leak if it fails |
|---|---|---|---|
| 1. Captured | Did the submission land anywhere? | Form tool or chat log submission ID and timestamp | Lost before the CRM |
| 2. Created | Did it create or update the right record, matched to the right account? | Lead or contact create date, account match, merge history | Sync failure, false merge, spam filter |
| 3. Assigned | Did a working owner receive it? | Owner field history, owner's active status, queue membership | Unowned or wrong-owner lead |
| 4. Touched | Did a human or approved automation reach out, and how fast? | First logged call, email or meeting after create date | Untouched or late lead |
| 5. Resolved | Was the lead qualified, disqualified with a reason, or recycled on purpose? | Status change, disqualification reason, recycle date | Silent disqualification or limbo |
Two numbers summarize the trace. The pass-through rate at each checkpoint shows where leads stop. The time from capture to first touch, measured at the median and at the 90th percentile, shows where they slow down. The 90th percentile matters because the average hides the leads that waited a day.
Illustrative example, with made-up round numbers: a company receives 1,000 inbound submissions a month. Its trace shows 960 records created, 900 assigned to an active owner, 720 touched, 610 touched within an hour and 680 resolved with a reason. The funnel report says "720 worked." The trace shows 40 lost before the CRM, 60 with a queue or inactive owner, 180 never touched and 110 touched late, each with a different fix and owner.
To size it in dollars, use the same structure as the leak ledger: leads lost at each checkpoint, times the conversion rate your fast-touched leads achieve, times average contract value, then a recovery factor for what a realistic fix would win back.
Implementation: six steps to close lead leakage
Reconcile capture against creation. Export the last 90 days of submissions from every inbound source (forms, chat, demo requests, event and list imports) and match them to CRM records by email and timestamp. The unmatched rows are checkpoint 1 and 2 leaks. Check merges into old records and personal-email submissions first.
Audit every assignment branch. List each routing rule and the leads it caught last quarter. Rules that caught nothing, leads that fell to a default owner or queue, and assignments to inactive users are your checkpoint 3 leaks. Add an out-of-office check to round robin, and give every rule an explicit fallback owner who is a person, not a queue.
Start the clock at submission, not at assignment. Many teams measure response time from when the lead was assigned, which hides every minute lost in creation and routing. Store the submission timestamp on the record and measure first touch against it. Our speed-to-lead benchmarks by ARR band suggest what "fast" should mean for your size.
Make untouched leads loud. A lead with no first activity after a set window (a suggested starting point is 15 minutes during working hours, not a benchmark) should alert the owner, then a manager, then reassign. Managerial tracking is one of the levers the Journal of Marketing study found changes follow-up behavior.
Require a reason for every disqualification. Automated rejection is fine when the reason is recorded and reviewable. Add a required disqualification reason, route "no enrichment match" and personal-email leads to a human review queue instead of auto-recycling them, and sample a batch every month to check the rules are rejecting the right people.
Run the trace weekly and assign each checkpoint an owner. Marketing operations usually owns checkpoints 1 and 2, sales operations owns 3, sales leadership owns 4, and checkpoint 5 is shared. Publish the five pass-through rates and the p90 clock every week. Our guide to evidence-gated lifecycle stages covers how to stop status fields from hiding stalled leads.
Workflow: the first hour of an inbound lead
Each stage has one job, one owner and a time budget, and exceptions are caught before the buyer notices them.
Every inbound source writes to one intake point with a submission ID and timestamp. Nothing enters through a side door the trace cannot see.
The record is created or updated and matched to its account. Merges into old records are logged and flagged, never applied silently.
The lead is checked against the ICP your team approved and assigned to an active owner who fits the account. Uncertain cases go to a named person with the reason attached, instead of to a default queue.
The owner reaches out, ideally with a first reply already drafted from what the buyer submitted. If nothing is logged inside the window, the alert and reassignment chain begins. Workato's 2026 test found that 42 percent of the companies that called did so within the hour, so a sub-hour phone touch is attainable, not aspirational.
Every lead ends in one of three states: qualified into an opportunity, disqualified with a reason, or recycled on purpose with a date.
Board narrative: how to report lead leakage
Lead leakage fits on one slide, because it is about buyers who asked and were not answered. Three framings work.
"We traced every inbound lead from submission to resolution. We now know where each lost lead stopped, not just how many converted."
"Leads we answered within the hour converted at a higher rate than leads we answered late. Applying that gap to the leads we lost or delayed gives a recoverable pipeline figure that finance has reviewed line by line."
"Each checkpoint has an owner and a weekly number. We are closing the largest first and will report pass-through rates and p90 response time every quarter."
Cross-domain: where lead leakage connects to the rest of the engine
Lead leakage sets the volume every later stage works with. Closing it is the job of Speed-to-Lead, which watches every inbound source, qualifies and routes each lead, and drafts the first reply. It starts read-only, logs every action with its reasoning, and runs against around 20 of your own past leads before touching a live one. It ships at 85 percent or it does not ship.
The next leak downstream is the handoff. A lead answered in four minutes can still die between SDR and AE, which is what the Handoff Orchestrator addresses, and our analysis of why deals die between marketing, SDR and AE applies the same clock-based approach to that stage. Upstream, checkpoints 2 and 3 depend on matching a lead to its account, which is an identity problem before it is a routing one; our piece on automating territory and account assignment covers the rule engine behind clean ownership.
Seen as a whole, inbound response is one of the nine systems a revenue engine needs, and usually the first to fix. The full set is on the systems page.
Frequently asked questions
What is lead leakage?
Lead leakage is the loss of inbound leads between the moment a buyer submits a form and the moment a qualified person responds. It covers submissions that never become CRM records, records with no working owner, leads that are never contacted or contacted too late, and leads disqualified automatically without a recorded reason. It is the earliest form of revenue leakage in the funnel.
What causes leads to fall through the cracks?
Leads usually fall through the cracks because of system gaps, not rep laziness. Common causes are form-to-CRM sync errors, deduplication rules that merge new leads into old records, assignment rules with no fallback owner, round robin assigning to reps who are away, reps prioritizing their existing pipeline, and enrichment or scoring rules that reject leads silently.
How do you measure lead leakage?
Measure lead leakage by tracing each inbound lead through five checkpoints: captured, created, assigned, touched and resolved. Reconcile form submissions against CRM records, check owner history and active status, and compare first-activity timestamps with the original submission time. The pass-through rate at each checkpoint shows where leads stop, and the 90th percentile response time shows where they slow down.
What is a good lead response time for B2B?
A good B2B lead response time is under an hour, and the best teams answer within minutes. A separate 1.25-million-lead study reported in Harvard Business Review in 2011 found that firms responding within an hour were nearly seven times as likely to qualify a lead as those responding an hour later. Measure from the original submission, not from assignment, and track the 90th percentile as well as the median.
How do you stop lead leakage without adding headcount?
Start by fixing the system, not the staffing. Reconcile submissions against records, give every routing rule a named fallback owner, remove inactive reps from round robin automatically, alert on leads untouched after a set window, and require a reason for every disqualification. Automating qualification, routing and a drafted first reply closes most of the gap before more reps are needed.
