A flat daily send cap across every client workspace looks fair until one campaign starts booking. Then the "fair" cap becomes a throttle on the only motion that is working. Agencies that run many ICPs feel this weekly. Client A is quiet. Client B is hot. The mailbox pool is shared capacity with non-shared reputation.
GTM keeps send windows, limits, and compliance gates inside client-isolated workspaces. The ops question is not whether to cap. It is how to cap so a hot campaign can expand without burning the quiet ones or the domain that carries them.
Cap what actually has a reputation budget
Mailbox caps and domain caps are different knobs. Coldops and other operator guides land in a familiar band for cold outreach: roughly 20 to 40 emails per mailbox per day once an inbox is mature, with lower starts during warm-up. Domain totals matter more than people admit. A handful of mailboxes on one domain still share DNS, authentication, and complaint history.
Google's sender guidelines push the point without giving you a magic daily number. Keep spam rates low in Postmaster Tools, ramp volume gradually, and avoid sudden spikes. Caps exist so you never invent a spike when a junior operator "just needs more volume for Client B."
Write three numbers per client workspace at launch:
- Per-mailbox daily max
- Per-domain daily max across that client's sending domains
- Reserved headroom you will not allocate on a normal day (often about 20%)
The reserve is how you feed a hot campaign without rewriting policy under adrenaline.
Separate baseline allocation from burst allocation
Baseline is what each workspace gets every weekday so quiet clients still warm and still learn. Burst is temporary extra volume drawn from reserve, not stolen from another client's baseline.
A clean pattern:
- Allocate baseline across workspaces from domain capacity minus reserve.
- Score which campaigns earn burst (positive reply rate after OOO/bounce filtering, conversations booked, placement health).
- Grant burst for a short window (one to three days) with an explicit end date.
- Pull burst back when the signal fades or when bounce/complaint rates climb.
Do not permanently reassign another client's baseline because one AE yelled. That teaches every AM to escalate instead of earn.
Protect the hot campaign without mixing identities
Hot does not mean "pour every mailbox into it." Keep each prospect on one consistent sender through the sequence when you can. Inbox rotation across providers helps capacity, but identity drift confuses replies and muddy attribution.
If Client B needs more volume, prefer:
- Adding a warmed mailbox already owned by that client's domain set
- Raising that mailbox toward its mature daily max
- Opening the reserved headroom for that workspace only
Avoid yanking mailboxes from Client A's workspace mid-sequence. Client isolation exists so one client's send pattern cannot rewrite another's reputation story.
When the hot campaign should not get more volume
More send is the wrong cure when:
- Bounce rate is climbing
- Spam rate in Postmaster Tools is rising toward the danger zone Google calls out (keep well under 0.3%)
- Positive replies are flat while raw replies are mostly OOO and "not interested"
- The sequence step getting the extra volume is the known weak step
In those cases, pause growth, fix list quality or copy, then re-earn burst. Starving a bad motion is correct ops. Starving a healthy one because the spreadsheet is rigid is not.
Operator checklist for weekly rebalance
- Review delivered, bounce, positive reply, and booked conversations per workspace
- Confirm no workspace is sitting on unused baseline while another is blocked on reserve
- Expire burst grants that hit their end date
- Document any permanent baseline change with the reason and the domain health snapshot
Agencies that optimize for conversations booked per client workspace treat caps as operating policy, not as a vanity daily total. GTM is built for that motion: find, qualify, write, send, reply, book inside isolated workspaces with send limits you can actually run. If that is the model you want, join the waitlist at https://trygtm.com/waitlist.
Multi-provider rotation without multi-client bleed
Some agencies spread mailboxes across Google and Microsoft tenants to widen capacity. That can help a hot campaign if each mailbox still belongs to the correct client domain set and still respects per-domain caps. It becomes a problem when rotation is used to hide that one client's domains are unhealthy. Caps should make health visible, not launder it.
Sources
- https://coldops.io/blog/how-many-cold-emails-per-day
- https://www.mailpool.ai/blog/email-throttling-strategies-managing-send-limits-across-multiple-providers
- https://coldicp.com/email-throttling-cold-outreach/
- https://support.google.com/mail/answer/81126
- /workspace/trygtm/knowledge/product.md