Argument · Updated 2 August 2026 · 13 min
The handoff is the leak: how to fix the gap between B2B marketing and sales
A practical guide to defining, measuring and repairing the point where a qualified buyer moves from B2B marketing to sales.
A buyer asks for a demo. The form creates a record, enrichment adds a company name and an assignment rule sends it to a sales queue. Marketing can see that the lead was delivered, but nobody in sales accepts it because the account already has an owner and the new record looks like a duplicate. An automatic email is sent, so the workflow appears active, yet no person is responsible for deciding what should happen next.
This is an illustrative example, but it shows why a handoff cannot be measured by asking whether a record moved from one system or owner to another. The record moved. Responsibility did not.
A marketing-to-sales handoff is complete only when an eligible buyer or account reaches the right owner with enough context to act, the owner accepts responsibility, the next action happens within an agreed window and the outcome returns to the system. Assignment is one event inside that process. It is not proof that the process worked.
What a marketing-to-sales handoff actually is
The word handoff suggests that marketing finishes its work and passes a complete package to sales. B2B buying rarely works that neatly. A person may research anonymously, download material for a future project, attend an event, return with colleagues, ask a technical question and only later request a conversation. The relevant unit may be one contact, an account or a group of buyers whose activity only becomes meaningful when it is considered together.
The handoff is therefore a change in commercial responsibility. Before the change, marketing is responsible for interpreting and developing the signal. After it, a named sales owner is responsible for accepting, working or returning it with a usable reason. The system between them has to preserve the evidence that justified the change.
The action, account context and agreed criteria that indicate a sales conversation is appropriate.
A named person or actively managed queue accepts the record, its context and the response clock.
The buyer is contacted, qualified, returned, recycled or disqualified with a reason that improves the next decision.
This definition separates delivery from acceptance. An email notification proves that a notification was sent. A changed CRM owner proves that a field changed. Neither tells you whether somebody understood the reason for the handoff, accepted the work or completed the next action.
Not every marketing signal belongs in sales
Many handoff disputes begin before routing. Marketing and sales have not agreed which behaviour means that a buyer wants a conversation, so a webinar registration, repeated website visits and a demo request may all enter the same queue even though they call for different treatment.
Current buyer research gives teams a reason to be careful. In its 2025 B2B Buyer Experience Report, 6sense surveyed nearly 4,000 recent buyers and found that buyers initiated 79% of first engagements with sellers. On average, first contact happened 61% of the way through the reported buying journey. The exact percentage should not become a universal rule, but the broader finding matters: many buyers control when they want direct contact and arrive with substantial research already completed.
That means an interaction is evidence, not an automatic instruction for sales. The system needs to distinguish at least three kinds of signal.
A contact, demo, pricing or consultation request should enter a priority route with a clear owner, relevant context and an agreed response clock.
Repeated activity across relevant people, product use or another defined signal may justify account review and considered outreach, but it is not the same as consent to an immediate sales conversation.
A report download, newsletter subscription or single event registration usually belongs in an appropriate follow-up or nurture path unless other evidence changes its meaning.
The distinction protects both teams. Sales receives fewer records with a clearer reason to act, while marketing does not force early-stage researchers into a sales process simply to satisfy a volume target. If sales consistently returns one class of record as too early, the useful question is not which team is wrong. It is whether the entry rule sends the right signal to the right treatment.
Define the handoff contract before automating it
A reliable handoff is built from standing decisions that marketing, sales and revenue operations can inspect. Together, those decisions form a handoff contract. It does not need to be a legal document or a long service-level agreement. It needs to answer the questions that otherwise get remade for each record.
- Unit of demand
- Are you routing a person, a company, an opportunity or a buying group assembled from several contacts?
- Entry rule
- Which observable evidence makes this record eligible, and which conditions suppress or delay it?
- Required context
- Which source, action, account facts, prior conversations and consent information must travel with it?
- Route and fallback
- Who should receive it, and where does it go when territory, ownership, capacity or data does not match the main rule?
- Acceptance
- Which recorded event proves that the receiving owner has taken responsibility rather than merely received a notification?
- Response clock
- When does the clock start, what counts as a meaningful response and which exceptions pause or change the expectation?
- Disposition
- Which outcomes can the owner record, including accepted, contacted, qualified, returned, recycled, duplicate and disqualified?
- Review trigger
- Which evidence should reopen the rule, such as repeated returns, a new segment, changed territories or a growing exception queue?
This is where apparently simple automation often breaks. A routing rule can only use information that exists in the right place when the rule runs. In a narrow but useful example, Salesforce documented in June 2026 that its standard lead assignment rules cannot evaluate Campaign fields for several lead creation paths because campaign membership is stored on a separate record. The rule may look correct in a diagram while lacking the context it expects at execution time.
The general lesson applies beyond one platform. For every routing condition, check where the data originates, when it becomes available, which object holds it and what happens when it is missing. A fallback queue only works if somebody owns the queue and the system raises an exception before the record becomes stale.
A lead is not handed off when it is assigned
The cleanest operating model treats the handoff as a sequence of visible states. Each state needs an entry event, an owner and a valid next state. This prevents a single status such as qualified from carrying several incompatible meanings.
- 01Captured
The original action and source are recorded without yet claiming that sales should act.
- 02Eligible
The record meets the agreed entry rule and carries the required context.
- 03Routed
A rule assigns a named owner or managed queue and starts the relevant clock.
- 04Accepted
The receiving owner acknowledges responsibility or returns the record through a defined exception path.
- 05Worked
A meaningful action occurs and its timestamp is recorded separately from an automated acknowledgement.
- 06Resolved
The commercial outcome and reason are recorded so marketing, sales and operations can learn from it.
The valid path is not always forward. A record can be returned because it is a duplicate, recycled because the timing is wrong, or moved back to marketing because the person asked for information rather than a sales conversation. These are legitimate outcomes when the reason is recorded. A record that simply disappears from a queue is not.
Modern CRM platforms can expose much of this without a separate analytics project. HubSpot’s current lifecycle stage documentation, for example, distinguishes lifecycle stage from Lead Status and provides calculated properties for the dates records enter and exit stages, plus time spent in a stage. The tool does not decide what your stages should mean, but the underlying events make waiting time and skipped states measurable.
Who should own the handoff?
Marketing and sales both participate in the handoff, but joint participation cannot mean that ownership is vague. One operating owner should maintain the state model, routing logic, instrumentation and exception review. In many companies this is revenue operations. In a smaller team it may be marketing operations or a commercial leader with enough access to see both sides.
That operating owner should not unilaterally decide what sales must accept or what marketing must produce. The leaders carrying the commercial consequences approve the contract, while the teams doing the work contribute evidence and use the agreed states.
Maintains the definitions, fields, routing rules, timestamps, fallback paths and exception review. Smaller companies may place this responsibility with a commercial leader.
Approve which demand deserves sales capacity, what response is expected and which return reasons are valid. A senior commercial leader resolves trade-offs that affect targets or territory.
Capture the evidence, keep queue coverage realistic, accept the work and record what happened. Their recurring exceptions show where the contract needs to change.
Suppose marketing wants to route every attendee from a high-value event because attendance shows substantial interest, while sales says the list contains researchers, partners and existing customers. The operating owner should not settle the argument by choosing a lower score. The teams need to define which event behaviours, account traits and existing relationships make a record eligible, then test whether accepted records progress differently from returned ones. The contract changes when the evidence changes, not when one team needs a better dashboard.
Six common handoff failures and what they reveal
Handoff problems are easier to diagnose when the visible symptom is connected to the missing control. The following patterns are not proof of one cause, but they give the audit somewhere specific to begin.
| What you can observe | What to inspect |
|---|---|
| Assigned but never workedThe CRM shows an owner, yet no meaningful action follows. | There may be no acceptance state, no response clock or no escalation when an owner is unavailable. |
| Returned as poor qualitySales sends back a large share of one campaign or source. | The entry rule may confuse research activity with buying intent, or required account context may be missing. |
| Correct person, wrong ownerA new contact routes away from the seller already working the account. | Identity resolution, account matching or existing ownership may happen after the assignment rule. |
| Records collect in a queueThe main rules run, but exceptions wait without a decision. | The fallback has a destination but no active owner, capacity rule or escalation deadline. |
| Timing is the only return reasonMarketing receives no explanation beyond not now. | The disposition model may be too vague to distinguish early research, missing budget, wrong person or a future review date. |
| Lead volume rises while acceptance fallsMore records cross the boundary, but sales accepts a smaller share. | Scoring, capacity and the definition of eligible demand may not have changed together. |
Several symptoms can share one cause. Missing account matching can create duplicates, wrong owners and false response-time breaches. This is why the audit should trace records through actual events rather than inferring the problem from a single dashboard.
What AI should automate and what it should not decide
AI does not inherently make the handoff better or worse. It changes how quickly the system can collect context, recommend an action and execute follow-up. The result depends on whether the definitions, evidence and authority around that action are visible.
AI can be useful for assembling account history, identifying likely duplicates, extracting the reason for a request, summarising relevant prior conversations, recommending a route, drafting a response and grouping free-text return notes into themes. These tasks remove searching and copying from the path between a signal and an informed action.
Some controls should remain deterministic because consistency matters more than interpretation. Consent and suppression rules, customer exclusions, territory authority, permissions, named fallback owners and escalation deadlines should not change because a model found a different pattern in the wording.
Human judgment remains important when the system encounters a new segment, an ambiguous strategic account, conflicting ownership, a commercially sensitive qualification decision or evidence that the contract itself should change. Human judgment does not mean every record must wait for manual review. It means the system has a named person with authority to inspect exceptions, override an output and revise the rule.
Enrich context, summarise history, detect duplicates, classify requests, suggest a route and monitor the response clock.
Apply consent, permissions, territories, exclusions, required fields, named fallbacks and escalation rules consistently.
Handle novel exceptions, accept commercial trade-offs and decide whether new evidence warrants a change to the operating rule.
This boundary also protects information quality. The NIST Generative AI Profile warns about automation bias, where people defer too readily to automated output, and describes high-integrity information as information that distinguishes fact, opinion and inference while remaining linked to its sources. A useful handoff summary should therefore show where each material fact came from and which parts are model interpretation.
Use AI to shorten the distance between evidence and action. Do not let it hide the evidence or the authority behind the action.
How to audit the handoff in an afternoon
Start with a bounded cohort rather than the entire database. Choose a recent period with enough volume to show the normal paths, then separate explicit buyer requests from research engagement and account-level signals. Comparing unlike signals will make both speed and conversion measures difficult to interpret.
- 01Reconstruct the event path
Export the original action, source, creation time, eligibility time, owner history, acceptance event, first meaningful response, current state, outcome and reason. Note which events do not exist or cannot be trusted.
- 02Measure coverage before conversion
Calculate the share of eligible records with a valid owner, the share accepted, the share worked and the share with a complete disposition. A conversion rate cannot explain records that never entered a workable process.
- 03Measure waiting, including the tail
Look at the median time from eligibility to acceptance and from acceptance to meaningful action, then inspect the slowest portion separately. A healthy median can hide a group of records that waited for days.
- 04Read the exception paths
Group returns, duplicates, no-owner records, stale queues and overrides by source, segment and route. Trace examples from each group through the systems that touched them.
- 05Compare the written definitions
Ask marketing and sales to state, separately, which evidence makes a record eligible, what sales commits to do and which return reasons are valid. Compare those answers with the rules running in the systems.
- 06Follow the learning back
Check whether sales outcomes change scoring, campaign treatment, audience choices or follow-up. If the information stops at a dashboard, the return path is not improving the next handoff.
A 2026 Qualified survey of 514 B2B marketing and revenue leaders reported that only 7.2% of respondents said their company responded to demo requests within five minutes, and that roughly one in ten inbound leads was never contacted. Qualified sells software for inbound engagement and the study relies on self-reported data, so those figures should be treated as a reason to inspect your own records rather than as a universal target. The useful benchmark is the one created from your buyers, routes and capacity.
How to redesign the handoff
Do not begin with a new scoring model or another platform. Begin with one type of demand that matters commercially and currently produces disagreement, such as demo requests from existing target accounts or event responses from a priority segment.
First, write the handoff contract for that path and have the accountable marketing and sales leaders approve it. Then create the missing states and reason codes, including acceptance, return and fallback. Test the routing logic against historical records that represent normal cases and known exceptions before it controls live work.
When the path goes live, monitor exceptions more closely than averages. An empty fallback queue may mean the rules work, or it may mean records are failing before they reach the queue. A rising return rate may indicate weaker demand, but it may also show that sales has started recording outcomes more honestly. Read a small sample of the underlying records before changing the rule.
Finally, set a review trigger. Reopen the contract when a new segment becomes material, territory or product structure changes, a source produces repeated returns, accepted records stop progressing or a new automated decision changes the evidence available to the owner. The purpose of the review is not to protect the original workflow. It is to keep the workflow aligned with how buyers and the business actually operate.
A reliable handoff makes four facts visible from one record: why it crossed, who accepted it, what happened next and what the system learned. If the CRM can show only that the owner field changed, the handoff is not finished.
Questions B2B teams ask
- What is a marketing and sales handoff?
- A marketing and sales handoff is the controlled transfer of a person or account from a marketing-owned stage into a sales-owned action. A usable handoff defines what qualifies, which context travels with the record, who accepts it, how quickly they act and what happens when the record is rejected or cannot be routed.
- Who owns the handoff between marketing and sales?
- Marketing and sales should jointly approve the commercial definition. Marketing owns the evidence it supplies, sales owns acceptance and follow-up behaviour, and revenue operations usually implements the shared stages, routing, timing and reporting in the systems. One named operator should be responsible for monitoring the interface end to end.
- Which metrics show whether the handoff is working?
- Measure the share of eligible records routed successfully, time to acceptance, time to meaningful action, rejection reasons, records with no owner, recycled records and progression after acceptance. Review the same cohort across the join instead of comparing separate marketing volume and sales activity totals.
- Can AI fix a weak marketing and sales handoff?
- AI can help classify, enrich, summarise and prioritise records, but it cannot resolve an undefined qualification rule or unclear ownership. Fix the contract and failure routes first, then automate the stable parts while keeping consequential customer decisions reviewable.