When to pause a client campaign

Pausing feels like failure when a client is watching weekly dashboards. Leaving a broken campaign live is the real failure. Reputation damage compounds. More volume into a bounce spike or a LinkedIn restriction makes the next month harder, not easier.

Agencies need explicit pause rules per client workspace before the first send. Decide the kill switches while everyone is calm.

Pause is an operating decision, not a panic

SuperSend's deliverability framing is right: pause when placement, bounces, provider behavior, or domain health show that more volume will make the problem worse. You are protecting the rest of the program while you fix the weak layer.

The same logic applies to LinkedIn. Soft warnings (CAPTCHA, forced re-auth, sudden acceptance collapse) are invitations to stop automation, not to push harder.

Signal 1: Bounce spikes

Hard bounces are a list-quality emergency. Industry guidance commonly treats roughly 1–2% hard-bounce rates on active cold campaigns as a stop-and-investigate threshold, with some operators pausing even earlier. Percentages are noisy below a few hundred sends, so also watch absolute spikes from a new list source.

When you pause:

• Stop the affected campaign or sender, not "the whole agency forever"
• Suppress hard-bounced addresses
• Re-verify the remaining segment before any resume
• Do not restart at the old daily cap the next morning

Soft bounces need patience and categorization. Quotas and greylisting are not the same as "user unknown." Still, repeated soft failures on the same address belong on a suppression path.

Signal 2: Restriction and session risk (especially LinkedIn)

Pause LinkedIn automation immediately on:

• Account restriction notices
• CAPTCHA or forced logout loops
• Acceptance rates collapsing versus the client's baseline
• Pending invite piles that stop clearing

Outzeach and PhantomBuster-style safety guides agree on the response shape: stop, complete verification, wait, then resume at a fraction of prior volume only after stability returns. Email can continue if it is healthy and if cross-channel stop rules still honor LinkedIn replies.

Never "share" LinkedIn identities across clients to keep volume up. Isolation exists so one restriction does not become a portfolio incident. Credentials should be forwarded to the connection vendor, not stored as a reusable password stash.

Signal 3: Reply quality collapse

Raw reply rate can lie. Watch the mix.

Pause or throttle when:

• Positives fall while unsubscribes and hostility rise
• Referral and "not now" patterns vanish and only noise remains
• Time-to-human on positives blows past your SLA because volume outran staffing
• Sequences fail to stop on reply (that is a product bug and a brand risk)

If infrastructure is healthy but reply quality is bad, the pause may be on copy, ICP, or entry rules rather than on DNS. Diagnose before you burn another week of credits.

Signal 4: Compliance misses

Any of these is an immediate full stop for the client campaign:

• Sends without a working unsubscribe path
• Suppression not applied (re-contact after opt-out)
• Compliance gate bypassed or misconfigured
• Inaccurate From identity or deceptive subjects
• Provenance gaps on fields you claimed were verified

GTM's compliance gate is designed as a database constraint: no send without a verdict and unsubscribe URL. If your process ever invents a workaround around a gate, pause and remove the workaround. Agencies remain responsible for messages sent on a client's behalf.

Signal 5: Domain and placement deterioration

Pause when:

• Inbox placement tests show spam-folder collapse across providers
• Authentication fails (SPF/DKIM/DMARC breakage)
• Tracking domains get listed or break
• Complaint rates climb
• A major blacklist hits the sending domain or IP

Partial pauses are valid. One bad list source can stop while another segment stays live. One mailbox can stop while siblings continue under lower caps. Scope the pause to the failing layer so you do not teach the client that "pause" means "we disappeared."

What to do during the pause

1. Name the trigger in the workspace log with timestamps and metrics.
2. Fix the layer (list, auth, identity, copy, staffing), not the vanity chart.
3. Tell the client what stopped, why, and what resume criteria look like.
4. Resume with a smaller first segment and written ramp rules.
5. Keep suppression and sequences honest while paused so opt-outs still process.

A simple pause card for every client

Write the numbers the client agrees to before launch. Arguing about thresholds mid-incident is how campaigns stay live too long.

Sources

• Safe multi-touch outreach flow (PhantomBuster)
• Outreach volume control (Outzeach)
• When to pause a cold email campaign (SuperSend)
• Common outbound automation mistakes (Instantly)

Want client workspaces with send windows, compliance gates, and pause-friendly controls built for agencies? Join the GTM waitlist.