In short
- The handover is an incentives problem, and most attempts to fix it are definitions projects.
- Marketing is paid on leads this quarter. Sales is paid on pipeline this quarter. An old contact belongs to neither.
- A rejected lead needs a destination, or rejection becomes deletion by another name.
- Routing failures are silent, because an unrouted lead looks identical to a worked one in a dashboard.
- Every stage needs a named owner, including the stage where a lead is turned down.
- Benchmarks from other companies tell you almost nothing here. Use your own closed-won rate.
The definitions are usually fine
Every company that has this problem has already tried to solve it with definitions.
There is a document. It says what a marketing qualified lead is, what a sales qualified lead is, and what has to be true for one to become the other. Both teams signed it. It made no difference.
That is worth sitting with, because it rules out the obvious explanation. The definitions are not the binding constraint. If they were, writing them down would have worked.
The actual cause is how each side is paid
Marketing is measured on leads generated this quarter. Sales is measured on pipeline closed this quarter.
Now consider a contact who enquired eighteen months ago, was passed to sales, was judged not ready, and has been sitting in the database since.
That contact appears in neither number. Working it does not help marketing’s lead count, because the lead was already counted. It does not help sales’ quarterly pipeline, because the odds on any single old contact are low and the effort is the same as working a fresh one.
Nobody is being lazy. Both teams are behaving exactly as their targets instruct. The contact falls through a gap that the measurement system created.
This is why the handover is never fixed by a better definition of MQL. The definition describes the moment of transfer. The failure happens afterwards, in the space neither team is paid to occupy.
Rejection needs a destination
Watch what happens when sales declines a lead.
In most systems, nothing. The lead status changes and the record stops. There is no next step, no owner, and no future date on which anyone will look at it again. Rejection has become deletion, without anyone deciding to delete anything.
The fix is small and structural. A rejected lead re-enters nurture automatically, with the rejection reason attached, and a defined window before it can be routed to sales again.
Two things follow from that. The contact stays alive, and marketing gets the rejection reasons, which is the most useful feedback available for improving what it sends. Most companies collect that data and never read it, because nothing is built to act on it.
Routing failures are invisible by design
The second silent failure is a lead that never reaches anyone.
A form fills in with an unmatched value. A round-robin points at a departed employee. A workflow sits paused after somebody edited it. In every case the lead enters the system, appears in the count, and is never touched.
The reason this survives is that dashboards report on activity, not on the absence of it. A lead sitting unrouted looks exactly like a lead being worked, right up until someone asks why the conversion rate is falling.
Two checks find most of it:
- Count leads created in the last quarter with no owner assigned. Any number above zero is a routing fault.
- Count leads with an owner and no logged activity in fourteen days. This is capacity or motivation rather than routing, and it needs a different conversation.
What a working handover looks like
Written down, it fits on one page, and every row has a name in it.
| Stage | Owner | Maximum time | On breach | |—|—|—|—| | Enquiry received | Marketing ops | 1 hour to route | Alert to a named person, not a shared inbox | | First human response | Sales | 1 working day | Reassign automatically | | Qualification decision | Sales | 3 working days | Escalate to sales lead | | Rejected lead returned | Marketing | Same day, automatic | Rejection reason mandatory | | Nurture re-entry | Marketing | Defined window | Re-route only after the window |
The important column is the last one. A service level agreement with no consequence on breach is a statement of intent, and statements of intent are what everyone has already tried.
Why external benchmarks will not help you here
You will find published MQL to SQL conversion rates by industry, and they vary enormously, because every company defines the two terms differently.
A company with a loose MQL definition will show a low conversion rate and might be performing well. A company with a strict one will show a high rate on far fewer leads. Comparing the two numbers compares the definitions rather than the performance.
The comparison that means something is your own, over time, with the definition unchanged. If you change the definition, note the date and treat what came before as a different series.
Questions, answered
Should marketing own leads after sales rejects them? Yes, and with a rule rather than a habit. Automatic return, mandatory reason, defined window.
How long should sales get before a lead comes back? Short enough that the contact is still warm, and long enough to be fair. Three working days for a decision is a common and workable answer.
What if sales says the leads are bad? Ask for ten examples and go through them together. Either the pattern is real and the targeting changes, or it is a handful of memorable bad leads standing in for the whole set. Both outcomes are useful and neither survives an argument conducted in the abstract.
Do we need a revenue operations function for this? It helps, and the constraint is ownership rather than headcount. Someone has to be measured on what happens between the two teams. Until that is true, adding a function relocates the problem rather than solving it.
Where does an old database fit in this? It is the same failure at scale. Contacts that belong to no current quarter’s number, sitting in a system nobody is paid to work.