Batched versus spread sending: why timing changes placement

Two senders can push identical volume to identical lists and get different results, because receivers evaluate rate and pattern as well as total. Burst sending from a sender with thin reputation looks like a compromised mailbox — which is precisely the behaviour filters are built to catch. Spreading the same volume with irregular intervals looks like people writing to people.

What the receiver reacts to

• Instantaneous rate. Per-minute submission limits exist at Google and Microsoft independently of daily caps, and they bite first.
• Connection concurrency. Twenty parallel connections to one provider earns 421 responses no matter how clean your list is.
• Uniformity. Messages at exactly the same interval, with near-identical size and structure, read as machine-generated. Human mail is irregular.
• Day-over-day change. A sender whose daily volume multiplies overnight is the single most suspicious shape there is.
• Per-destination pacing. Your total volume matters less to Gmail than the volume you push at Gmail specifically.

The practical schedule

1. Pick a daily number well under your provider's cap and under what your reputation currently supports. The effective ceiling for a new mailbox is far below its nominal quota.
2. Spread it across working hours in the recipient's rough timezone, not yours.
3. Add jitter. Randomised gaps — not a fixed interval — between sends. Identical spacing is itself a fingerprint.
4. Cap concurrency per receiving provider. A small number of connections, reused, beats many in parallel.
5. Pace per destination domain. Track how much you are sending to Gmail, Microsoft and Yahoo separately, and throttle each independently.
6. Use more mailboxes rather than a harder one. Ten mailboxes at 30 a day each sits comfortably inside every limit; one at 300 does not.
7. Hold a floor on quiet days. Reputation decays with silence, and a stop-start pattern forces repeated re-ramps.

Respect the deferrals, they are instructions

The receivers tell you when the shape is wrong, in plain text:

421 4.4.2 Message submission rate exceeded
451 4.7.500 Server busy, please try again later
451 4.7.650 The mail server [203.0.113.10] has been temporarily rate limited
421 4.7.0 [TSS04] ... temporarily deferred due to unexpected volume or user complaints

All 4xx, all temporary. Back off, lower the rate, keep the recipients — and add jitter to the retries too, or every deferred message will re-attempt at the same instant and recreate the burst.

Deferrals that thin out while volume holds steady mean your reputation is improving. Deferrals that rise while volume is flat mean complaints, not rate, and the fix is list quality rather than scheduling.

Where this interacts with warm-up

A warm-up ramp is a sending schedule with a slope. The same rules apply, and the slope is the extra variable:

• Grow by a modest proportion every few days rather than doubling.
• Keep the intra-day spread from the first day, not just at high volume.
• Never resume at full rate after a pause; treat a quiet fortnight as a fresh ramp.
• Generate replies, not just opens. Replies and folder rescues are the signals receivers weight; opens are weak and, on Gmail, partly synthetic because image requests are proxied, cached and sometimes prefetched.

Verify by placement, not by acceptance

Your sending platform's delivery rate cannot tell you whether the schedule helped, because accepted mail includes spam-foldered mail. Seed-test on a fixed cadence with several mailboxes per provider, and read Gmail's tabs separately — Promotions carries the `INBOX` label, so any label-based check counts a Promotions delivery as an inbox hit and overstates primary-tab placement.

Cross-reference with what Google publishes: Postmaster Tools v2 reports spam rate, authentication, delivery errors and compliance status, plus a Deliverability analysis report added in June 2026. There is no reputation grade to compare against — the domain and IP reputation dashboards retired with v1, senders having been redirected from 30 September 2025 and v1 removed by the end of 2025. Keep spam rate under the 0.30% threshold and aim below 0.10%.

What not to do

• Do not queue the day's volume at 09:00. It is the fastest way to a per-minute limit.
• Do not use a fixed interval between sends. Uniformity is a signal; jitter is cheap.
• Do not raise concurrency to clear a backlog. Deferrals become refusals.
• Do not retry everything simultaneously after a deferral.
• Do not plan a campaign at your published daily cap. The reputation-based ceiling is lower and invisible.
• Do not assume more volume fixes a complaint rate. It is a ratio, and the curve is itself a signal.

How MailPilot helps

MailPilot's warm-up sends are spread through the day on a compounding ramp rather than batched, which is the pattern that avoids per-minute limits and reputation throttling. It tracks each mailbox's reputation daily, measures inbox placement per provider with a seed-inbox network that reports Gmail's tabs separately, and checks SPF, DKIM and DMARC on the sending domain. Every plan includes unlimited mailboxes, which is the practical way to spread volume rather than pushing one sender to its ceiling.