HubSpot CRM sequences are one of the most reply-sensitive email channels a modern sales team runs. They look and feel like personal 1:1 email — because technically they are. Each step in a sequence sends from the individual rep's connected inbox (Gmail or Microsoft 365), carrying the rep's SMTP reputation, domain alignment, and spam history with every message.
That design is a feature, not a bug: personal mail from a real sender is the whole point of a sales sequence. But it also means deliverability is no longer a single tenant-wide problem — it is the sum of however many reps you have sequences enrolled under. One rep's misconfigured alias, one rep's previous cold-email burn, one rep whose MFA reset broke the connected inbox — and the team's reply rate quietly halves. Placement testing per rep, before you scale the sequence, is the only way to catch it.
Sequences vs marketing email — two different delivery stacks
HubSpot sends email through two entirely separate pipelines. Treating them as one is the first mistake most teams make.
- Marketing email. Sent from HubSpot's own IP pool through their dedicated marketing MTA. Your marketing sending domain publishes CNAMEs that alias HubSpot's signing keys. Deliverability here is HubSpot's operational problem as much as yours.
- Sequences (and 1:1 email from the CRM). Sent through the individual rep's connected inbox — Gmail API or Microsoft Graph. The message leaves Google's or Microsoft's outbound infrastructure under that user's identity. HubSpot is not the sender; the rep is.
Because sequences inherit the rep's inbox reputation, every placement question is a per-user question. Rep A sending from a well-warmed Gmail has near-perfect placement. Rep B, joining last week, sending under an alias that was never properly SPF-aligned, lands in Promotions or worse. Your sequence performance report looks like a content problem — it is actually an infrastructure problem concentrated in a handful of users.
Open rates within one sequence vary wildly between reps sending identical templates. That is almost always connected-inbox reputation, not copy. A placement test per rep will confirm it in five minutes.
What HubSpot sequences actually require
For a rep's connected inbox to land cleanly in a prospect's primary tab, three things need to be in order:
- Authentication. SPF must include the rep's provider (
_spf.google.comorspf.protection.outlook.com). DKIM must sign under the rep's actual sending domain, notgmail.comoronmicrosoft.com. DMARC must be aligned or passing. - Inbox warm-up. A brand-new rep mailbox with no prior history, sent into a cold sequence on day one, will get throttled by Google and Microsoft. Send manual 1:1 email for one to two weeks before enrolling sequences.
- Reply hygiene. Sequences measure replies. Replies come from recipients who saw the message in their inbox. If the rep's
Reply-Toroutes to a shared alias that nobody monitors, or to a domain that receivers distrust, replies degrade even when placement is fine.
Per-rep placement testing in practice
The test is deliberately simple: every rep sends an identical one-off message from their connected inbox into a seed set that mirrors where their real prospects sit. You record inbox versus Promotions versus Junk per provider per rep. The spread tells you which reps are fit to enrol in a sequence and which need remediation first.
What the seed set should cover
- Gmail personal (
@gmail.com) - Gmail Workspace (corporate Google tenants)
- Outlook.com / Hotmail (consumer Microsoft)
- Microsoft 365 (corporate Microsoft tenants)
- Yahoo / AOL (for consumer reach)
- Regional providers your audience uses
What the message should look like
Use the real first step of the sequence you are about to enrol, including merge fields rendered against a representative contact. A "hello, this is a test" placeholder has almost no predictive value — Gmail's Promotions heuristic is sensitive to exactly the language, link density, and image ratio of the real template.
A native integration for leading CRMs is in private beta — placement tests inside the CRM, alerts on drops.
Reading the results
Patterns that repeat across HubSpot sequence audits:
- One rep Promotions, everyone else Primary. That rep's connected inbox has alignment or warm-up issues specific to them. Pull them out of the sequence until their seed test comes back green.
- Entire team in Promotions on Gmail only. Content classifier is firing on your template, not infrastructure. Reduce link count, drop tracking pixels on the first send, shorten the body.
- Entire team in Junk on Microsoft 365. Usually a corporate DMARC policy that is stricter than your alignment. Check the
d=on the DKIM signature from the rep's connected inbox — if HubSpot is signing underhs-sequences.ioinstead of your domain, alignment fails against a strict DMARC policy. - Random drops over time. Reputation decay from high-volume, low-reply sending. Sequences are not a firehose — they are a conversation tool. If reply rate is under two percent, you are burning reputation faster than you are building it.
A weekly routine that works
Monday Every rep with an active sequence sends one seed test.
Monday Any rep below 80% inbox on Gmail is paused from sequences.
Tuesday Paused reps run through connected-inbox reauth +
HubSpot delivery-settings audit.
Wednesday Re-test. Reinstate reps who recover to >= 80% inbox.
Friday Sequence performance review uses the placement data
as a control variable, not just opens/replies.The point of the weekly cadence is to keep the seed-test data fresher than any single sequence cohort. A sequence enrolled on Monday that runs for fourteen days will finish under the same placement conditions it started — because you verified, not because you hoped.
FAQ
Does HubSpot throttle sequences if my placement is poor?
Can we use a shared mailbox for sequences to simplify this?
How many seed mailboxes do we actually need per rep?
What changes break sequence placement overnight?
p=none to p=reject when HubSpot's signing wasn't aligned to begin with. A new sending alias added without SPF.