
Lead Management
Distribute leads fairly: rules, capacity and confirmed ownership
Route leads transparently to exactly one eligible person using a routing card, owner protection, eligibility, capacity, acceptance and a bounded fallback.
Short answer
Distribute a lead only after the business case is unambiguous and ready for routing. Then build a set of genuinely eligible people: check the existing relationship, responsibility, language, region, permission, working state and available capacity before selecting anyone. Round robin or weighting should decide only among those eligible people. Routing succeeds only when exactly one person accepts ownership and CRM contains a dated next action. If nobody accepts, use a limited and traceable fallback path—never endless reassignment.
Key takeaways
- Use round robin or weighting only inside a pool that passes relationship, must-have, permission and operating-state checks.
- Keep the contact, business case, routing decision and task separate so a new enquiry does not overwrite a relationship owner or another case.
- Test incumbent owners, destination objects, synchronisation and visibility permissions with concrete cases before the pilot.
- Technical assignment is not acceptance; evidence one accepted owner, a next action and a bounded fallback.
Why equal rotation alone is not fair lead distribution
Consider a fictional Monday morning. Three new enquiries arrive in sequence. The first belongs to an existing customer managed by Nora. The second requires an English-speaking product specialist. The third is a standard German quote request. A simple rotation distributes the contacts in turn to Nora, Mehmet and Lea. The dashboard looks balanced. Operationally, it is wrong: Nora’s existing customer loses the known contact, Mehmet receives an enquiry outside his product scope, and Lea is not working that day.
Round robin was not the problem. It was applied too early. The rotation selected a person before the team had defined who was eligible and able to work each case.
Fair distribution does not mean that every person receives the same number of contacts. It means the selection rule is understandable before the result and the actual workload is auditable afterwards.
Qualitative Reddit discussions surface the same conflicts: existing accounts should reach the current owner, extra conditions such as territory or capacity make routing hard to understand, urgent leads reach the wrong rep, or cases sit without escalation. These are useful problem signals, not evidence of frequency or effectiveness. A robust design must therefore be tested against the company’s own cases.
Keep the contact, business case, route and task separate
In everyday language, a “lead” often means a person, enquiry, opportunity and work item at once. Safe distribution treats them as four separate objects.
Object | What it represents | Its own state | Common failure |
|---|---|---|---|
contact | known person or organisation | identity, relationship, contact owner | every new enquiry overwrites the relationship owner |
business case | a specific need at a point in time | case ID, need, status, source | several needs from one person are combined into one case |
route | the decision about the eligible pool and selected person | rule version, candidates, reason and time | only the new owner remains visible; the decision path disappears |
task | the next action someone must actually perform | assignee, due time, status and outcome | assignment is treated as completed outreach |
The contact and opportunity owner may be the same person, but need not be. One person can retain the customer relationship while a specialist temporarily owns an opportunity or task. Every automation must therefore state which object it changes and whether owner synchronisation affects another object.
The routing card makes every decision auditable
The routing card is not another form for sales to maintain. It is the small shared state that the workflow, CRM and team read for one distribution decision.
Field | What it contains | Why it is needed |
|---|---|---|
routing input | unique event and case ID, timestamp | prevent the same start from being distributed twice |
business need | confirmed need and desired outcome | distinguish interest in a contact from a routable case |
existing relationship | confirmed account, contact or opportunity owner | protect continuity deliberately or hand it over with a reason |
must-have criteria | product, language, region, role and required permission | remove ineligible people before selection |
operating state | active, working, absent or blocked | avoid treating every configured user as available |
capacity state | open, limited, full or unknown with observation time | expose stale or invented capacity |
eligible pool | user IDs that meet every must-have criterion | apply round robin only to the right set |
selection rule | relationship, highest free capacity, rotation or weighting | explain and reproduce the result later |
rule version | explicit version and effective date | avoid mixing old and new decisions |
assignment | selected person, time and attempt | separate a technical change from human work |
acceptance | accepted, declined, pending or expired | prove one active responsibility |
next action | activity, due time and destination state | keep the lead actionable after routing |
fallback | next permitted pool or named human owner | never fall into an anonymous residual state |
If a must-have field is missing, the state is `review`, not `default owner`. If capacity is unknown, a simple rotation may still be acceptable—but only as an explicit pilot assumption, not as a claim that selection is capacity-aware.
Use a fixed rule hierarchy to resolve conflicts
Evaluate rules from top to bottom. A later stage must not silently reverse an earlier protection decision.
The Microsoft overview of unified routing separates classification of incoming work from later assignment in a different system and considers factors including skills, presence and capacity. This is not a description of ONE LOOP functionality. As an external system reference, it supports the design rule of building an eligible candidate set before selecting one person.
Stage | Decision question | Possible outcome | Stop or evidence |
|---|---|---|---|
1. protection | Is the case unique, current and eligible for distribution? | continue, review or suppress | duplicate, closed case, missing identity or ineligible work stops |
2. relationship | Is there a confirmed relationship or opportunity owner? | prefer the eligible incumbent or hand over deliberately | multiple matches, absence or ineligibility needs a visible reason |
3. eligibility | Who meets the product, language, territory, role and permission requirements? | build the eligible pool | no candidate triggers fallback, not the removal of every constraint |
4. operability | Who is working in the relevant period and can accept more work? | reduce the pool by working state and capacity | unknown or stale state remains visible |
5. selection | Which eligible person receives the case? | relationship, capacity, round robin or weighting | store the reason, candidate set and rule version |
6. fallback | What happens without an eligible or accepting person? | triage queue or named owner | limited attempts, no endless ping-pong |
A “high-value lead” must not bypass eligibility. A high score may change the order of eligible work, but it cannot make an unauthorised or unsuitable user eligible.

The robust workflow from routable lead to accepted task
From routable case to confirmed ownership
Secure the case
Confirm a unique case and entry ID, the need and the intended outcome.
Build eligibility
Check relationship, product, language, region, role, permission and working state.
Select one person
Apply capacity, rotation or documented weighting only inside the eligible pool.
Evidence acceptance
Verify acceptance time, responsible person, next action and due date in the destination system.
No eligible user, unknown capacity, owner conflict or expired acceptance deadline?
Send the case and routing card to named triage. Do not silently relax criteria, overwrite an owner or create parallel offers.
Recommended flow: check eligibility and operating state before selection; complete assignment only through acceptance and a next action.
1. Secure the routable business case
Do not start routing from “contact created” alone. Confirm that a specific open case exists with a unique ID, a stated need and a desired next outcome. A repeated technical input may read the existing route, but must not create a second opportunity or another active assignment.
If the contact or account match is ambiguous, this stage ends in review. Distribution does not solve an identity problem by choosing an arbitrary owner.
2. Build the eligible candidate set
Check the existing relationship, must-have criteria, permission and operating state in sequence. The result is not “everyone in sales”; it is a concrete list of eligible user IDs with an observation time. A user who exists in the team but lacks permission, is inactive or does not cover the product does not belong in the selection.
If nobody remains, send the case to visible triage. Do not silently relax constraints merely so the automation can write a name.
3. Select for capacity and explainable fairness
Within the eligible set, the business can use a confirmed relationship owner, highest known free capacity, round robin or documented weighting. One primary method and one explicit tie-break rule are enough for the first pilot. Several competing rotations across different workflows soon become impossible to explain.
Write the selected person, candidate set, rule version and time atomically to the routing card. A parallel run should recognise the active decision and reuse it rather than draw again.
4. Confirm acceptance and the next action
After technical assignment, exactly one person receives a visible work offer. They accept or decline with a limited reason. Only acceptance sets `accepted at` and starts the operational response clock. A concrete next action with a due time is created at the same point.
If the predefined acceptance deadline expires on the relevant business calendar, offer only the still-open case through the agreed fallback. A late acceptance by the first owner must not compete with the new acceptance; the current routing version determines which action is valid.
When round robin works—and when it does not
Round robin is simple, transparent and often useful for comparable work. It is only a selection method, however, not a full routing strategy.
Situation | Round robin works when … | Add another rule when … |
|---|---|---|
small team with the same offer | everyone in the pool may handle the same cases and is available | leave, part-time work or different permissions are absent from pool state |
several languages or products | an eligible language/product pool is built first | the rotation runs across all users before eligibility |
existing customer relationship | the incumbent should deliberately receive no preference | continuity, commission or account responsibility must be protected |
different capacity | weights are current, justified and auditable | a permanent 2:1 weight merely reflects an old staffing model |
several queues or workflows | one shared workload view exists | each queue rotates “fairly” while loading the same person repeatedly |
very different case complexity | the cases are genuinely comparable | a complex case and a short standard enquiry count as equal work |
The Microsoft documentation for assignment methods illustrates an important boundary in a different system: round-robin order can be maintained per queue, so people who belong to several queues may still receive back-to-back or concurrent work. Skills, presence, capacity and working hours also need suitable configuration. This is not a description of ONE LOOP functionality; it demonstrates why “even per queue” does not automatically mean “fair overall”.
Protect existing owners without freezing old responsibility
Do not overwrite an existing owner automatically. First determine whether the relationship is confirmed, the current case still belongs to that owner’s scope and the person remains able to work it. There are then three explicit outcomes:
Outcome | When it fits | What is recorded |
|---|---|---|
retain incumbent | confirmed relationship, suitable expertise and current capacity | protection reason and current confirmation |
hand over the opportunity | relationship remains, but a specialist owns this case | contact, opportunity and task owners separately |
route to triage/fallback | owner is absent, unauthorised, unsuitable or relationship is ambiguous | reason, new responsible party and return rule |
An option such as “only unassigned contacts/opportunities” is an important guard, but not a complete policy. Depending on configuration, it prevents an overwrite; it does not show whether an old owner is still responsible or reachable. Conversely, deliberate reassignment must not unexpectedly change contact and opportunity owners together. Test that coupling on the real destination object during the pilot.
Capacity needs a timestamp and a definition
“Available” can mean calendar time, presence, open tasks, active opportunities or manually declared capacity. Those values are not interchangeable. Choose one definition for the first pilot and store its source and observation time.
A robust minimum:
Eligibility is binary: the person may and can handle this kind of case.
Capacity is a current operational value with a source and observation time.
Unknown capacity does not automatically mean free.
Selection compares only values with the same definition.
A manual override stores its reason, author and time.
When work differs materially, the team may later use workload units instead of contact counts. Those weights begin as an operating hypothesis. Validate them against actual handling time and open inventory rather than presenting them as a scientifically optimal allocation.

Assignment, notification and acceptance are three events
An owner field proves only technical assignment. An internal message proves only that a notification was created. Confirmed acceptance is the evidence that a person took responsibility.
State | Evidence | Permitted next action |
|---|---|---|
`assigned` | owner, routing version and assignment time stored | send notification and start acceptance deadline |
`offered` | exactly one active work offer visible | allow acceptance or a reasoned decline |
`accepted` | person and acceptance time stored unambiguously | activate next action and operating response clock |
`expired` | business calendar and deadline timestamp recorded | close current attempt and trigger one fallback step |
`fallback` | named queue or person owns the case | human triage; no anonymous residual list |
`handed over` | old and new owner plus handover reason stored | only the new owner may perform the open task |
The right deadline depends on opening hours, team size, channel and promised response time. This guide therefore gives no universal number of minutes. Define it before the pilot and measure it from assignment; customer response time is measured separately from the original enquiry.
What the Salesforce time split shows—and what it does not
For its State of Sales, 7th Edition, Salesforce surveyed 4,050 sales professionals across 22 countries. The anonymous survey ran from August to September 2025 and all respondents were third-party panelists. The report allocates an average workweek as 40% selling and 60% not selling.
How an average sales workweek is allocated
Share of reported workweek · Surveyed August–September 2025 · report 2026
- Selling40 %
- Not selling60 %
View data table
| Category | Share of reported workweek |
|---|---|
| Selling | 40 % |
| Not selling | 60 % |
Global vendor survey; not time tracking and not evidence of lead-routing performance.
Source: Salesforce · State of Sales, 7th Edition · Surveyed August–September 2025 · report 2026. 4,050 sales professionals in 22 countries; anonymous survey of third-party panelists.
No Germany-only or SME-only subset reported. ‘Not selling’ covers several activities and is not synonymous with unnecessary administration. No routing, conversion or ONE LOOP effect. Scale 0–100%.
The figures provide context for limited selling capacity in this global vendor survey. They are not a German SME sample or observed time tracking. “Not selling” is not synonymous with unnecessary administration; the report covers several activities. The 40/60 split does not establish that lead routing saves a particular amount of time or that round robin improves conversion or revenue.
For the pilot, the figure is context only. The defensible evidence comes from the company’s own routing data: how many routable cases reached an eligible person, were actually accepted and received a next action?
Resume safely without owner ping-pong or double acceptance
A failed or delayed run reads the actual state first:
Has the same routing input ID already been processed?
Which business case is linked and still open?
Which rule version created the current route?
Is there exactly one active offer to accept?
Was it accepted, declined or has its deadline expired?
Are the owner, permission and next action actually present in the destination?
Which one step is missing?
If the owner is set but notification failed, recreate only the notification. If the task exists but acceptance remains pending, do not create a second task. If acceptance and escalation arrive nearly simultaneously, the first valid decision stored for the current routing version wins; the other action reads that state and stops. A manual handover freezes automatic routing until a person explicitly releases it.
Ten test cases before the pilot
The following cases are fictional tests, not measured ONE LOOP results.
Test case | Expected evidence |
|---|---|
the same routing input arrives twice | one business case and one active routing decision |
a confirmed customer asks about an existing quote | eligible incumbent is protected or handover reason is visible |
company name matches two accounts | review case; no arbitrary owner selection |
lead requires English and product A | pool contains only users meeting both requirements |
selected user is disabled or lacks permission | no assignment to that user; visible fallback |
user belongs to two pools and receives concurrent work | shared capacity is checked or the limitation is logged as pilot risk |
contact has an owner but opportunity is unassigned | only the intended object changes; sync behaviour is evidenced |
owner declines with a reason | one limited new attempt and old work offer closed |
acceptance and expiry occur almost simultaneously | exactly one valid person and one active task |
a person manually hands over an accepted case | automation does not overwrite the new owner |
A test passes only when the routing card, destination object, permission, acceptance state and next action show the same expected result. A green workflow run or sent notification is not sufficient.
Metrics with fixed numerators and denominators
Universal target rates would create false precision. Start with absolute cases and explicit denominators.
Metric | Numerator | Denominator |
|---|---|---|
routable cases | unique open cases with every required input | all reviewed new business cases |
assigned to an eligible owner | cases whose owner met every must-have criterion at decision time | all routable cases with at least one eligible person |
no eligible person | routable cases with an empty candidate set | all routable cases |
acceptance rate | unambiguously accepted assignments | all valid assignment offers |
time to acceptance | time from valid assignment to valid acceptance | reported per accepted assignment; use a distribution, not only a mean |
reassignment rate | cases with at least one valid new routing attempt | all routable cases |
prohibited owner overwrite | confirmed incumbent owners changed against the protection rule | all routes with an existing owner |
fallback inventory | open cases in triage/fallback with named responsibility | all open routable cases |
workload distribution | accepted comparable cases or documented workload units per eligible person | compare only the same candidate set and period |
Report the candidate set as well. A rep may receive fewer leads because they were eligible or available less often. Without that denominator, a correct eligibility rule can look like unfair distribution.
How ONE LOOP starts the lead-distribution pilot
Choose one source, one clear business-case type and a small team. Document the routing card and rule hierarchy first. Then enable appropriate owner protection, build an eligible pool and use exactly one selection method within it. Add a visible work offer, an acceptance deadline defined on the business calendar and a named fallback.
ONE LOOP can assign contacts and opportunities by rule to one or several users, distribute them equally or by weight, and then trigger notifications and tasks. Before the pilot, verify the exact destination entity, incumbent owners, synchronisation settings, selected user lists, permissions, repeated triggers and behaviour when the destination object is missing. Capacity, absence and human acceptance are treated as automated facts only when the concrete setup actually supplies and stores those states for testing.
If response timing is the first problem, read Build a speed-to-lead system. For channel intake, use Move social DMs into CRM or Route WhatsApp leads safely. ONE LOOP can then review a limited lead-distribution pilot with you.
FAQ