# The Cross-Team Contact Collision Standard: Proceed, Coordinate, or Stand Down

*Score any person who matches across your ATS, CRM, and sequencer on five dimensions and log a Proceed, Coordinate, or Stand Down verdict with a named owner.*

- Canonical URL: https://www.refolk.ai/guides/cross-team-contact-collision-standard
- Pillar: Process, data, and compliance
- Format: Framework
- Published: 2026-10-29
- Last reviewed: 2026-10-29
- Reading time: 23 min
- Keywords: sales and recruiting contacting the same person, recruiting employees of your own customers, can candidate data be used for sales prospecting, customer contract non-solicit of employees, ATS and CRM duplicate contact conflict

## Key takeaways

- A cross-team contact collision should be scored per proposed touch, not per person, on five dimensions scored 0 to 2: relationship at stake, purpose compatibility, contract exposure, process timing, and suppression scope.
- A score of 2 on purpose compatibility, contract exposure, or suppression scope is a Stand Down whatever the other scores say, because those three are legal limits, not business preferences.
- Recruiting a customer's champion causes the departure risk CS teams pay to detect: a vendor conference figure puts 12-month churn after a champion leaves at 51%, and 65% non-renewal after an executive change, with no published methodology.
- Outreach dedupes prospects on email only and the key cannot be changed, so a candidate who applied from a personal address will not collide with a CRM work-email record unless you also match on profile URL or phone.
- Published customer non-solicit templates run from 6 months to 5 years after the contract term and usually carve out general advertising and unsolicited applicants, which a targeted InMail to a named customer employee does not qualify for.
- Refolk's index holds 2,590 revenue operations titles in the US against 477 recruiting or talent operations titles, about 5.4 to 1, so the collision check most likely ends up with RevOps-owned matching infrastructure.

This standard covers the moment before a recruiter or a rep contacts someone who already sits in another team's system. That person might be a champion on an open deal, a user at a paying customer, an active candidate, or a portfolio founder. The guide gives recruiting operations and revenue operations a five-dimension score and a three-way verdict (Proceed, Coordinate, or Stand Down) that names who owns the next touch. It is precise enough that two operators grading the same case should reach the same answer.

## What is a cross-team contact collision?

A cross-team contact collision is one person who has records in more than one team's system, usually the ATS, the CRM, and the sequencer, at the moment one team wants to contact them for a purpose the other team did not collect them for. The collision read decides, for that one proposed touch, whether to send it, route it through the other team's owner, or not send it.

Three terms first:

- **ATS (applicant tracking system):** holds candidates and their hiring stage.
- **CRM:** holds accounts, deals, and the people on them.
- **Sequencer:** the tool that enrolls a person in automated multi-step outreach, such as Outreach.

The collisions that do real damage usually look like one of these:

- A recruiter wants to hire the champion on an open deal.
- A rep enrolls an active candidate in a sales sequence.
- A recruiter sends a targeted InMail to a heavy user at a paying customer.
- Any team cold-approaches a founder the company has invested in.

Existing guidance handles each side on its own:

- **Sales rules of engagement** settle which rep owns a contact.
- **Company-level off-limits lists** settle which employers recruiting avoids.
- **Suppression guides** sync unsubscribes between tools.

None of them scores a single person who sits in both the hiring funnel and the revenue funnel. That gap is why these cases get settled in Slack after someone is already annoyed.

One rule shapes everything below: score the proposed touch, not the person. On the same day, the same champion can be a Proceed for a CS check-in, a Coordinate for a sales expansion pitch, and a Stand Down for a recruiter's InMail.

## The five dimensions at a glance

The read scores five dimensions from 0 to 2 each: relationship at stake, purpose compatibility, contract exposure, process timing, and suppression scope. Two are business judgements and three are legal limits, and that split decides how the scores become a verdict.

| Dimension | Question | A high score proves | How it lies |
|---|---|---|---|
| Relationship at stake | What live relationship does another team hold? | A misstep costs revenue or a hire | Stale role fields, person already left |
| Purpose compatibility | Was this data collected for this use? | The reuse lacks a lawful footing | "Fresh sourcing" that began in the other system |
| Contract exposure | Does a customer contract limit hiring them? | The touch may breach a clause | Scope read too broadly or too narrowly |
| Process timing | Is either process at a decision point? | Now is the worst moment | Slipped close dates, stale candidate stages |
| Suppression scope | Has the person said stop to any same-sender brand? | The send is not allowed | An email-only match misses the opt-out |

A 0 means clear. A 1 means a condition the other owner has to be involved in. A 2 means high stakes on the two business dimensions, and a hard stop on the three legal ones.

These rules are what make two operators agree:

1. Score only from evidence you can cite: a record, a contract clause, or a list entry. Memory and Slack threads do not count.
2. A dimension with no evidence either way scores 1, never 0. Unknown is not the same as clear.
3. If you are torn between two scores, take the higher one and write down why.
4. If identity is not resolved across all three systems, do not score yet. Fix the match first.
5. Where the person's location changes which law applies, score under each law and keep the stricter result.

## How do you score relationship at stake and process timing?

Relationship at stake measures what the company loses if the touch goes badly. Process timing measures whether this is the worst week to find out. The people who own the CRM and ATS records score both, using what those records say today.

### Relationship at stake

| Score | Revenue side | Hiring and investing side |
|---|---|---|
| 0 | No open deal, not a user at a paying account | Not an active candidate, no investment tie |
| 1 | Named contact or user, no decisive role | Sourced, applied, or early screen |
| 2 | Champion, signer, or executive sponsor | Past first interview, or a portfolio founder |

A champion is the person inside a buying or customer account who argues for your product internally. A champion scores 2 because of what their departure costs. These are the most-cited published figures:

| Event | Reported figure |
|---|---|
| Champion leaves a customer | 51% churn within 12 months |
| Executive leaves a customer | 65% do not renew |
| CS acts within 48 hours of an executive change | 33% more likely to renew |
| Past champions who do not return unprompted | 91% |

Treat these as directional. The 51% and 65% figures come from a conference talk by the CEO of a customer-intelligence vendor, and the methodology has not been published. The 91% figure is also vendor-reported. The argument does not depend on the exact number, though. Hiring a customer's champion away triggers the same departure risk that CS teams pay for job-change alerts to catch. No published data measures how often deals slip when a champion leaves an open opportunity, so on open deals the score rests on the account owner's judgement, and the log should say so.

> Recruiting a customer's champion is a churn event you cause yourself, on your own calendar.

**How this dimension lies:**

- An optimistic rep has ticked the champion role in the CRM, but nobody else on the account recognises the name. Check the activity history.
- The champion has already left the company. If so, the relationship you would damage belongs to the former employer, not to the person. Check their current employer on their public profile before scoring.
- The ATS still shows an active stage for someone who was rejected weeks ago. Check the date of the last stage change.

### Process timing

| Score | Meaning |
|---|---|
| 0 | Neither process is near a decision |
| 1 | A process is live but not at a decision point |
| 2 | Late-stage deal, renewal window, final round, or offer |

Use the definitions your teams already work to: the CRM stage names that count as late stage, and the renewal window CS already uses. I do not set a number of days, because no source supports one. Timing lies when close dates slip and nobody updates them, or when the person is sitting at final round in the ATS but has already accepted elsewhere. Ask both owners, not just the systems.

#### Relationship at stake against process timing

Horizontal axis runs from Low relationship at stake to High relationship at stake. Vertical axis runs from Quiet process to At a decision point.

| Quadrant | What it means |
| --- | --- |
| Low stake, hot window | Coordinate. Let the live process reach its decision first. |
| High stake, hot window | Stand Down until the window closes, with a dated re-check. |
| Low stake, quiet | Proceed, provided purpose, contract, and suppression all score 0. |
| High stake, quiet | Coordinate. The relationship owner decides when and how the touch happens. |

*Once the legal dimensions are clear, the stakes and the timing together decide whether a touch goes now, goes through the owner, or waits.*

## How do you score purpose, contract, and suppression?

These three dimensions are legal limits, so a 2 on any of them is a hard stop that no business case overrides. Each is settled by a specific document: the compatibility record for purpose, the master services agreement for contract exposure, and the suppression lists for suppression scope.

### Purpose compatibility

Purpose limitation is the data protection principle that personal data collected for one purpose can be reused for another only if the new purpose is compatible with the first. The ICO puts it plainly: you must only reuse personal information for a new purpose if it is compatible with the original purpose, and you must have a lawful basis for the new purpose. If the original lawful basis is not enough, you must find a new one. The ICO lists five factors to weigh:

1. Any link between the original purpose and the new one.
2. The context of collection, especially your relationship with the person and what they would reasonably expect.
3. The nature of the personal data.
4. The possible consequences of the new processing for the person.
5. Whether appropriate safeguards are in place.

Reuse is likely to be incompatible if it is very different from the original purpose, would be unexpected to the person, or would have an unjustified impact on them. The ICO notes that this assessment resembles a legitimate interests assessment, so a DPO who already runs those can run this one.

So, can candidate data be used for sales prospecting? Not by default. No ruling or ICO example covers moving data from recruiting to sales or from sales to recruiting. My reading is that someone who applied for a job would not reasonably expect to end up in a sales sequence. Unless a compatibility record says otherwise, the answer is no.

| Score | Condition |
|---|---|
| 0 | Requesting team uses data it collected for this purpose |
| 1 | Reuses the other team's record, compatibility recorded yes, lawful basis named |
| 2 | Reuses with no compatibility record, or a recorded no |

This dimension lies when someone says "I found them myself". If the touch was prompted by seeing the person in the other team's system, treat it as reuse, whatever route was used to look them up afterwards.

> **Note:** Re-check the legal footing before relying on this
>
> The ICO has amended its purpose limitation guidance to reflect the Data (Use and Access) Act, and this guide does not cover EDPB guidance for EU data subjects. Re-read the current ICO page before adopting the rubric, and have your DPO check EU cases against EDPB material.

### Contract exposure

A customer contract non-solicit of employees is a clause in which one party agrees not to recruit the other party's staff for a set period. Nobody has published how often these clauses appear in B2B SaaS or services contracts. Template terms are documented, though, and they vary widely:

| Published template | Duration | Carve-outs |
|---|---|---|
| Balanced clause-library template | Term + 12 months | General ads, unsolicited applicants, terminated staff, people 6 months past separation |
| Law Insider personnel sample | Term + 12 months | General recruiting, job fairs |
| Law Insider provider sample | Term + 6 months | None listed, covers staff involved in the deal |
| Law Insider mutual sample | Term + 5 years | None listed |

Some clauses carry a real penalty. One published sample sets it at six times the employee's gross monthly pay. Read three things in the clause:

- **Duration:** when it ends relative to the contract term.
- **Scope words:** "material contact", "involved in the services", or all employees.
- **Carve-outs:** general advertising, unsolicited applications, former staff.

The clause also has an outer limit set by antitrust law. The DOJ and FTC Antitrust Guidelines for Business Activities Affecting Workers say that any agreement restricting a company's ability to hire another company's workers, such as an agreement not to cold call them or one requiring the current employer's permission before hiring, may be an unlawful no-poach agreement. A non-solicit in a customer contract is usually part of, and ancillary to, the main commercial relationship. I have not reproduced the guidelines' ancillary-restraint test here, so read it in the guidelines directly. Two Commissioners dissented from the guidelines, so also confirm they are still in force. In practice, never enforce a clause more broadly than it is written, and send any clause that has no general-advertising carve-out to legal for review rather than stretching it yourself. One published clause analysis says California Business and Professions Code Section 16600 voids employee non-solicits, so check the governing law with counsel.

| Score | Condition |
|---|---|
| 0 | Not a hiring touch, no clause, clause expired, carve-out applies, or counsel confirms void |
| 1 | Clause exists but its scope has not been read against this person |
| 2 | Clause in force, person in scope, approach targeted |

### Suppression scope

Suppression scope is the set of opt-outs and objections that bind this particular send, across every brand that could count as the same sender.

In the US, CAN-SPAM covers email whose primary purpose is commercial advertisement or promotion, and it makes no exception for business-to-business email. Opt-outs must be honoured within 10 business days. Separate lines of business count as separate senders only if they are presented to recipients as separate. Where several senders share a message that is not structured to name one of them, each must honour opt-outs. The FTC has not said whether a recruiting email counts as commercial.

In the UK, PECR does not require you to honour a corporate subscriber's email opt-out. UK GDPR still applies, though, and an individual's right to object to direct marketing is absolute. Whether a recruiting approach counts as direct marketing under PECR has not been established.

That leads to an unverified inference that needs legal review: opt-outs probably run one way. A sales opt-out would bind every commercial email sent under the brand, but a recruiting opt-out may not legally bind sales.

#### Which opt-outs bind a send, widest first

1. **Individual objection** - The person's own request to stop. Under UK GDPR, the objection to direct marketing is absolute.
2. **Same-sender commercial opt-out** - A CAN-SPAM opt-out held by any brand presented as the same sender, honoured within 10 business days.
3. **Team-only preference** - A recruiting opt-out. Its legal reach over sales email is not established.
4. **Corporate subscriber opt-out** - A company-level request. PECR does not require it to be honoured for email.

*An individual's own objection binds every team, while a corporate or team-only opt-out has a narrower or unsettled reach.*

| Score | Condition |
|---|---|
| 0 | No opt-out or objection on any matched record |
| 1 | An opt-out exists whose reach to this touch is unsettled |
| 2 | A personal request to stop, or a same-sender opt-out on a commercial send |

My policy is to score a person's own request to stop as 2 for every touch, whichever system recorded it. The law may not always require that, but the person does not think in terms of your teams.

## How do five scores become a verdict and a named owner?

Apply the verdict rules in order, and the first row that matches wins. Hard stops come first, then the combination of stakes and timing, then any remaining caution, and every verdict names the person who owns the next touch.

| Condition | Verdict | Owner of next touch |
|---|---|---|
| Purpose, contract, or suppression scores 2 | Stand Down | Existing owner, and the requester logs a re-check trigger |
| Relationship and timing both score 2 | Stand Down until the window closes | Existing owner |
| Any other score above 0 | Coordinate | Existing owner sets the sequence and the date |
| All five score 0 | Proceed | Requesting team, which notifies the other owner |

> **Rule:** Hard stops are never averaged
>
> A 2 on purpose, contract, or suppression ends the read at Stand Down, however low the other four scores are. Adding scores together hides exactly the legal limits the read exists to catch.

#### From scores to verdict

1. **Resolve** - One person ID is linked to every ATS, CRM, and sequencer record
2. **Hard stops** - Any 2 on purpose, contract, or suppression means Stand Down
3. **Stakes and window** - Relationship and timing both at 2 means Stand Down until the window closes
4. **Caution** - Any remaining score above 0 means Coordinate
5. **Owner** - Log the verdict with the named owner of the next touch

*The read checks legal hard stops before business stakes, so a commercial case never outweighs a legal limit.*

The "existing owner" depends on the relationship:

- **Open deal or paying account:** the account owner, or the CSM after the sale.
- **Active candidate:** the recruiter of record.
- **Portfolio founder:** the deal partner.

Under Coordinate, the existing owner decides whether the other team makes contact, when, and through whom. The requester supplies a draft and waits. Under Stand Down, the requester logs a dated re-check or a trigger event, such as the deal closing, the renewal signing, the clause expiring, or the candidacy ending. A Stand Down caused by an individual objection may never clear, and the log should say so.

### Who should own the read itself?

The read needs one process owner, and in most companies that will be revenue operations. Refolk's index holds far more RevOps titles than recruiting ops titles, and the sequencer and CRM where matching happens usually sit with RevOps.

| Segment | Titles in Refolk's index | Most common title in a sample of 25 |
|---|---|---|
| RevOps, US | 2,590 | Director of Revenue Operations, 72% |
| RevOps, UK | 300 | Revenue Operations Manager, 52% |
| Recruiting ops, US | 477 | Recruiting Ops Manager, 40% |

**5.4 to 1 - US revenue operations titles for every recruiting or talent operations title**

From Refolk's index, 2,590 against 477. The US RevOps population is also 8.6 times the UK's.

This conclusion rests on headcount alone, so treat it as a likely pattern rather than a rule. Recruiting ops still owns the ATS side of the inputs, and the read fails without them. If you are designing the check and want to compare notes with people who run the same stack, [Refolk](/) can find them from a plain-English description.

I ran this search: `Directors of Revenue Operations at US SaaS companies who use Outreach and Greenhouse` - [see the full result list](https://www.refolk.ai/s/dhsssczvyr).

*Returns RevOps leaders who run the same sequencer and ATS pairing, the peers most likely to have built a cross-system collision check already.*

## The collision read, step by step

No published source documents this procedure end to end. I assembled the order below from the ICO's guidance, sequencer documentation, and contract practice, and each step has a defined finish point.

#### Running one collision read

1. **Resolve identity across all three systems** - Match the person across the ATS, CRM, and sequencer on work email, personal email, profile URL, and phone. Done means one person ID linked to every record, with every email-less prospect checked for a duplicate.
2. **Record the relationship at stake** - The account owner or CSM records the person's role on any open deal or paying account, and the recruiter records any candidate stage. Done means both are written down and scored 0 to 2.
3. **Test data origin and purpose** - The record owner and the DPO record the data's origin, its original lawful basis, and the ICO's five compatibility factors for the new use. Done means compatibility is recorded as yes or no, with the new lawful basis.
4. **Check contract exposure** - For a hiring touch, legal or deal desk searches the MSA with the person's employer for a non-solicit and reads its duration, scope, and carve-outs. Done means the clause is cited, or recorded as none.
5. **Compare process timing** - The two owners compare deal stage, renewal date, and candidate stage. Done means the decision window is identified and scored.
6. **Check suppression across every same-sender brand** - Ops checks the opt-outs and objections held by every brand that could count as the same sender, on every matched record. Done means each list is checked and a re-check is set for right before the send.
7. **Log the verdict and the owner** - Apply the verdict rules in order and log Proceed, Coordinate, or Stand Down with the named owner of the next touch. Done means five cited scores, a verdict, an owner, and a re-check date or trigger.

The sources disagree on when suppression should be checked. The ICO describes checking your own do-not-contact list before sending direct marketing, while sequencer setup guides put deduplication at sync time. Do both. Dedupe when records sync, then run the suppression check again immediately before the send, because an opt-out can arrive between the read and the send.

**Collision read log entry**

```
COLLISION READ: <person ID>
Requested by: <name, team>
Proposed touch: <recruiting / sales / CS / investor> for <purpose>
Records matched: ATS <id> | CRM <id> | Sequencer <id, plus any duplicates>
Match keys used: work email / personal email / profile URL / phone

R  Relationship at stake: <0/1/2>  Source: <deal or account role, candidate stage>
P  Purpose compatibility: <0/1/2>  Source: <compatibility record, lawful basis>
C  Contract exposure:     <0/1/2>  Source: <MSA clause reference, or "none">
T  Process timing:        <0/1/2>  Source: <deal stage, renewal date, candidate stage>
S  Suppression scope:     <0/1/2>  Source: <lists checked, by brand>

Rule matched: <hard stop / stakes and window / caution / all clear>
Verdict: <Proceed / Coordinate / Stand Down>
Owner of next touch: <name, team>
Re-check: <date or trigger event>
Pre-send suppression re-run: <date, result>
```

*One entry per proposed touch. Leave no score blank. Unknown scores 1, with "no evidence" as the source.*

## How does the collision read go wrong?

Most bad verdicts come from one of two errors. A missed identity match produces a false Proceed. A clause or opt-out read more broadly than it is written produces a false Stand Down. The first annoys a customer or a candidate, and the second throws away a legitimate hire or deal, so check for both.

> **Watch out:** Email-only matching shows no collision at all
>
> Outreach dedupes prospects on email, the key cannot be customised, and candidates often apply from personal addresses. A read that matches on email alone will report a clean Proceed for the exact people this standard exists to catch.

1. **Email-only match.** A candidate applied with a personal Gmail address, and the CRM holds their work email. No collision appears, which is a false Proceed. Match on profile URL and phone as well. A people search that returns each person's public profile across GitHub, LinkedIn, and the open web, as Refolk does, gives you a key that does not depend on which email address they used.
2. **Lead converted without an email.** In Outreach, a CRM Lead or Contact synced without an email address can create a duplicate prospect, and the suppression attached to the first record is then missed. Audit prospects that have no email. Remapping a duplicate does not remove it from sequences, so check enrollment by hand after any merge.
3. **Mismatched identity keys across tools.** Salesforce Account Engagement moved its unique identifier from email address to CRM ID, while the sequencer still keys on email. No published native integration checks an ATS against a sequencer before enrollment. Expect to build the cross-check yourself.
4. **"General advertising" stretched.** A recruiter treats a targeted InMail to a customer's staff as covered by the general-advertising carve-out. It is not, because it was aimed at specific people. Check whether the outreach was aimed at named individuals.
5. **Clause read too broadly.** Treating every customer employee as off-limits when the clause covers only "material contact" staff produces false Stand Downs, and over-enforcement drifts toward the no-poach territory the antitrust guidelines describe. Read the scope words, and check whether California law governs.
6. **Single-sender assumption.** A recruiting brand marketed under the company name may count as the same sender as sales under CAN-SPAM. Check how the From line and the brand appear to the recipient, not how your org chart separates the teams.
7. **Corporate opt-out confused with personal objection.** PECR does not oblige you to honour a company-level email opt-out, but an individual's UK GDPR objection to direct marketing is absolute. Check whose objection it is before scoring suppression.
8. **Vendor statistics taken as measurement.** The 51% churn figure has no published methodology. Use it to explain why a champion scores 2, never to forecast a specific account.
9. **Laundered reuse.** A recruiter spots a name in the CRM, looks the person up again on a public profile, and records the data as freshly sourced. If the other system prompted the touch, score purpose as reuse.
10. **Scoring the person instead of the touch.** A single Stand Down stamped on a person blocks a harmless CS check-in along with the recruiter's pitch. Each proposed touch gets its own read.

## Keeping the collision read current

A verdict is only as current as the records it was scored from. Re-run the read whenever a trigger fires, and re-check the legal inputs on a schedule, because three of the five dimensions depend on guidance that can change.

Re-run a read when any of these changes:

- The deal stage moves, the deal closes, or the renewal signs.
- The candidacy closes or moves to offer.
- The person changes employer, which resets both relationship and contract exposure.
- A non-solicit clause reaches its expiry.
- A new opt-out or objection lands on any matched record.

Re-check the legal footing whenever the underlying guidance moves:

- The current ICO purpose limitation guidance, which has been amended for the Data (Use and Access) Act.
- Whether the DOJ and FTC worker guidelines are still in force.
- Any FTC statement on whether recruiting email counts as commercial.

When the verdict is Coordinate, send the existing owner a note that makes the decision easy and leaves the timing with them.

**Coordinate handoff to the existing owner**

```
Subject: Collision read for <person name> at <company>: Coordinate

<Owner name>, <person name> is <their role on your deal, account, or candidacy>.
<Requesting team> wants to <proposed touch> for <purpose>.

Scores: R <n> | P <n> | C <n> | T <n> | S <n>. Verdict: Coordinate.
You own the next touch. Please reply with one of:
1. Go ahead after <date or event>.
2. I will make the introduction myself.
3. Not this cycle. Re-check on <date>.

Nobody on <requesting team> will contact <person name> until you reply.
```

*Send it to the owner named in the log, and copy the requester. Do not contact the person until the owner replies.*

#### Before you call a collision read done

- [ ] The person has one ID linked to every ATS, CRM, and sequencer record, matched on more than email.
- [ ] Every sequencer prospect without an email address has been checked for a duplicate of this person.
- [ ] Each of the five dimensions has a score and a cited source.
- [ ] Every dimension without evidence was scored 1, not 0.
- [ ] For a hiring touch, the MSA with the person's employer was searched and the clause cited or recorded as none.
- [ ] Opt-outs and objections were checked under every brand presented to recipients as the same sender.
- [ ] The verdict came from the ordered rules, with no hard stop averaged away.
- [ ] The log names one owner of the next touch and a re-check date or trigger.
- [ ] The suppression check is set to run again immediately before the send.

## Frequently asked questions

### Can candidate data be used for sales prospecting?

Not by default. Under UK GDPR as the ICO explains it, you may reuse personal data for a new purpose only if that purpose is compatible with the original one, and you need a lawful basis for the new use. Reuse that is very different from the original purpose, would surprise the person, or would affect them unjustifiably is likely to be incompatible. No published ruling covers moving data from recruiting to sales, so record a compatibility assessment and a lawful basis before any candidate goes into a sales sequence.

### Is it OK to recruit employees of your own customers?

Often yes, but check the master services agreement first. Many customer contracts include an employee non-solicit. Published templates run from six months to five years after the term and usually carve out general advertising and unsolicited applicants. A targeted message to a named customer employee is not general advertising. Weigh the relationship as well: hiring a customer's champion triggers the departure risk that vendor-reported figures put at 51% churn within 12 months. Treat that number as directional only.

### Does a sales unsubscribe stop recruiters from emailing the same person?

It may, and the safe practice is to treat it as binding. CAN-SPAM requires each sender of commercial email to honour opt-outs, and brands presented to recipients as one sender count as one. The FTC has not said whether a pure recruiting email is commercial, so the legal reach of the opt-out is unsettled. In the UK, an individual's objection to direct marketing is absolute. Treat any personal request to stop as a hard stop for every team.

### How do I catch ATS and CRM duplicate contact conflicts before outreach?

Match on more than email. Outreach dedupes prospects on email only, and that key cannot be changed. Candidates often apply from personal addresses, so the same person can sit in both the ATS and the CRM with no visible collision. Match on work email, personal email, profile URL, and phone, and link everything to one person ID. Run the check before enrollment, because no native integration that checks an ATS against a sequencer has been published.

### Who owns the next touch when sales and recruiting want the same person?

The team that already holds the live relationship owns it, unless the read comes back Proceed. Under Coordinate, the account owner, CSM, recruiter of record, or deal partner decides whether the other team makes contact, when, and through whom. Under Stand Down, the existing owner keeps the relationship and the requester logs a re-check date or trigger. Under Proceed, the requester sends and tells the other owner.

### Can a customer non-solicit clause create antitrust risk?

Yes. The DOJ and FTC worker antitrust guidelines say an agreement not to cold call another company's workers, or one that requires the employer's permission before hiring, may be an unlawful no-poach agreement. Non-solicits in customer contracts are usually ancillary to the main deal. Still, read the guidelines' ancillary-restraint test directly, confirm the guidelines are still in force, and never enforce a clause more broadly than it is written.

---

*From the Refolk guide library. I revise these guides rather than replacing them, so the current version is always at https://www.refolk.ai/guides/cross-team-contact-collision-standard*