Email deliverability benchmarks for agencies

The agency-specific risk: shared infrastructure, separate reputations

When an agency sends on behalf of multiple clients through shared sending infrastructure - a shared IP range, a shared platform account, or even just shared institutional practices - one client's poor list hygiene or aggressive sending habits can create reputation damage that bleeds into how receiving providers treat mail sent on behalf of other clients from the same infrastructure. This is the single biggest deliverability risk specific to agency work, and it has no equivalent for an in-house marketing team sending only its own mail.

Metrics need to be tracked per client, not blended

An agency-wide average bounce rate or complaint rate across all clients hides exactly the information you need: which specific client's list or practice is causing a problem. A blended 1.5% bounce rate across ten clients could mean every client is at roughly 1.5%, or it could mean nine clients are under 0.5% and one is at 11% - and only the second scenario needs urgent attention, but both produce the same blended number.

A per-client scorecard

Isolating infrastructure between clients

The most direct fix for the shared-infrastructure risk is isolating sending infrastructure per client wherever volume justifies it - separate sending domains, separate mailboxes, ideally separate IPs for higher-volume clients. For smaller clients where dedicated infrastructure isn't practical, at minimum track and alert on each client's metrics individually, so a problem client can be identified and addressed before it accumulates enough volume to meaningfully affect shared reputation.

Onboarding new clients carries the same risk as a new domain

A new client's list - especially one inherited from a previous agency or in-house effort with unknown list hygiene history - should be treated with the same caution as a cold list from an unfamiliar source: verify before the first send, start with a smaller test volume, and watch bounce and complaint metrics closely for that specific client before scaling to full campaign volume.

Setting client expectations around deliverability

Part of managing this risk is setting expectations with clients directly: a client who insists on sending to a purchased list, ignoring agency hygiene recommendations, is introducing risk that may affect other clients sharing infrastructure. Documenting this risk and your hygiene standards as part of onboarding gives you a clear basis to push back when a client's practices threaten shared reputation.

Put list-sourcing standards in the client contract

The clearest way to reduce the shared-infrastructure risk described above is addressing it before onboarding, in the contract or statement of work itself: specify what list sources are acceptable, who is responsible for hygiene before the first send, and what happens if a client insists on sending to a list that fails your standards. This turns a difficult in-the-moment conversation with a client into a reference to an agreement both sides already accepted, rather than a case-by-case negotiation under pressure once a problem has already started affecting other clients.

Include a specific clause about isolation and escalation: if a client's sending starts affecting shared infrastructure, you need the contractual standing to pause that client's sending or move them to isolated infrastructure quickly, without that becoming a larger relationship dispute in the moment it matters most. Revisit the contract language itself periodically, not just at the point of signing - a client's list-sourcing practices can change over the life of an engagement, and a standard that was adequate at onboarding may need updating as the relationship and the client's own program evolve. Assign a specific person to own this periodic contract review, rather than leaving it to whoever happens to notice a client's practices have drifted, since an unowned recurring task is the kind of thing that quietly stops happening after the first year.

What not to do

Do not report deliverability to leadership or to clients using only a blended, agency-wide number - it hides the client-specific problems that actually need fixing. Do not onboard a new client's inherited list onto shared infrastructure at full volume without a hygiene pass and a smaller test send first. Do not let a single difficult client's insistence on poor list practices override your own sending standards when other clients' deliverability depends on the same infrastructure staying clean.

MailPilot tracks reputation per mailbox and per domain, which maps directly onto a per-client view for agencies managing several accounts - a problem client's mailbox shows its own numbers rather than disappearing into an averaged total.