Volume is how much mail is sent over a stated period. Cadence is the recurring pattern of sending. A burst is a concentrated increase within a shorter interval. Two senders can have the same daily volume but very different traffic patterns. Investigate these dimensions separately when transport or placement changes, and avoid treating a small seed test as proof of behaviour at campaign scale.
Measure the actual outbound traffic
Build a timeline from the sending system's records, including campaign sends, application notifications, follow-ups, and background automation. Count the relevant attempts and outcomes with clear definitions. A dashboard limited to the campaign you paused may omit another system still sending through the same domain or infrastructure.
Choose intervals that reveal the question. A daily total can hide a short concentrated send, while minute-level noise may obscure a recurring weekly newsletter. Record recipients or attempts consistently and keep retries identifiable. A retry queue can create traffic after the initial campaign window, so do not assume the visible scheduler describes all outbound activity.
Connect the pattern with route and audience changes
Overlay the timeline with sending IPs, accounts, streams, template versions, and audience segments. A volume increase can coincide with an ESP migration, new tracking domain, or older contact segment. Those changes create competing explanations. Establish which objects changed before attributing the problem entirely to the size of the send.
Inspect receiving-provider responses for rate-related deferrals or other errors. Preserve exact codes and times rather than paraphrasing every temporary response as throttling. A temporary failure can have several documented causes. Folder observations supply another view: accepted mail can be in Spam even when the transport queue looks healthy.
Use responsible pacing rather than a universal schedule
Follow the sending service's and receiver's current guidance for your route. Google advises consistent sending and avoiding sudden increases, but that does not establish one safe ramp for every domain, IP, and audience. Sending history, infrastructure, recipient expectation, and response signals affect what is appropriate.
Do not invent a guaranteed number of days or messages that creates reputation. A new dedicated IP and an established shared pool have different operational conditions. If expansion is justified, use controlled authorized stages, inspect transport and feedback evidence, and preserve a way to pause. Sending more unwanted mail is not a remedy for weak engagement or complaints.
Interpret a seed test within its scale
A representative seed send can check route, content, authentication, and observed folders for its copies. It does not simulate all traffic pressure or every subscriber relationship in a large campaign. Record the surrounding sending activity when testing during a burst or pause, because the broader context can differ between otherwise similar runs.
For a before-and-after comparison, keep route, message, panel, and observation stage as stable as possible. Document any changes in background volume and audience. If placement improves after reducing activity, report the observed association without claiming the reduction was the only cause. Further evidence may be needed to distinguish infrastructure, audience, and timing effects.
Frequently asked questions
Is daily volume enough to diagnose a burst?
No. A daily total can conceal concentrated traffic. Inspect shorter intervals and queue behaviour appropriate to the incident while retaining the broader trend.
Does pausing one campaign stop all sending?
Not necessarily. Check automation, application mail, scheduled follow-ups, retries, and other accounts sharing the relevant identities or infrastructure.
Can a good test justify an immediate large increase?
It describes a small observed sample. Expansion needs its own operational judgement, current provider guidance, audience evidence, and monitoring rather than a guarantee inferred from one green run.
Sources and scope
Source check: November 5, 2026. Source checking is separate from publication. Observations apply to the tested copies and conditions.