Managed deliverability, operated at the infrastructure level.

Why does inbox placement need an owner?

A surprising number of senders treat deliverability as a project that finished the day authentication passed. The providers do not see it that way, and the reasons placement decays without active management are worth stating plainly.

1. Reputation is a moving average. Providers judge recent behaviour, not your history. A few heavy sends to a stale segment can pull a months-old reputation down inside a single week.
2. The rules keep moving. Google and Yahoo set their bar on 1 February 2024; Microsoft enforced its own from 5 May 2025; Gmail moved to permanent rejections in November 2025; and Google replaced its four-tier reputation grades with a pass/fail compliance view in Postmaster Tools v2. A program tuned for last year’s rules is already behind.
3. Engagement erodes. Lists churn, segments go stale, and the engagement signals providers weigh most heavily fade unless someone is actively re-segmenting and pruning.
4. Every change needs validating. A new IP, a DNS edit, an ESP switch or a domain change can each break placement quietly, and each needs checking against real inbox data rather than assumed to be fine.
5. Incidents are measured in hours. When a provider starts deferring, the window to respond before a soft limit becomes a hard block is short. Nobody schedules that window in advance, which is exactly why it needs an owner.

# Seed-list placement by provider, against last week\u2019s baseline
$ seedcheck --by-provider --compare last-week
gmail      inbox 91%   (was 94%)   ← 3-point slip, flagged
outlook    inbox 88%   (was 88%)   nominal
yahoo      inbox 95%   (was 95%)   nominal

# Gmail slipped — is it reputation or a complaint-rate move?
$ gpt-domain-rep example.com --window 7d
domain reputation: HIGH → MEDIUM   complaint rate 0.08% → 0.21%
# complaints climbing toward the 0.30% line — act on segmentation now

Fully managed, co-managed, or advisory?

The amount of the operation you hand over is a decision, not a fixed package. We run the engagement in one of three shapes, and we are clear about who is accountable for what before any work begins.

1. Fully managed. We run the daily delivery operation end to end — monitoring, corrections, incident response and reporting — and you receive the inbox and the report rather than the work.
2. Co-managed. You keep an in-house marketing or delivery team and we act as the control layer beside them: reading the data, calling the corrections, and stepping in when an incident needs hands on the infrastructure.
3. Advisory retainer. Your team executes; we monitor, direct and review on a set cadence. The lightest option, and the right one for senders who have the hands but not the specialist eye.

What you receive, and how often

Managed deliverability only works if you can see it working. The reporting rhythm is built so that nothing important happens in silence and nothing trivial generates noise.

Reporting you can act on

A report is only worth the call that comes with it. Each month you receive placement and reputation trend lines drawn per provider, the incidents handled and how, the changes made and their effect, and a ranked list of what to do next, ordered by the placement each item stands to win or protect. The working call walks the document rather than standing in for it, so the reasoning lives somewhere your team can revisit. Two things are deliberately absent: a single composite score that would tell you something is wrong without telling you what, and any upsell folded into the findings, since the report has nothing to sell that the retainer does not already cover.

How an engagement begins

A managed engagement opens with a baseline, because no one can manage a number they have not measured. In the first weeks we instrument the data sources — connecting Postmaster Tools and SNDS, enrolling the feedback loops, standing up blocklist and seed-list monitoring — and we record where placement, reputation and complaint rates actually sit, per provider and per stream. With that picture in hand we agree the thresholds that will raise an alert, the rhythm of the reporting, and which work sits with your team and which with ours. Only then does steady-state operation start. The baseline serves two ends: it gives you a starting line to judge the engagement against, and it gives us an honest reading to work from instead of a set of assumptions. Nothing in your sending changes during this phase without your sign-off.

The signals we watch

An ongoing operation reads several sources at once and weighs them against one another, because each answers a different question and none is the whole truth alone.

• Gmail compliance, through Postmaster Tools v2, as a pass or fail on authentication, one-click headers, spam rate and TLS — the frame that replaced the old four-tier reputation grades.
• Complaint rate read against Google’s two levels, not one: the 0.30% line where mitigation stops, and the stricter 0.10% we treat as the working target.
• Domain reputation ahead of IP reputation, since it outlives a change of provider and travels with the domain wherever it sends.
• The Microsoft view at IP level through SNDS, Yahoo and AOL complaints through the feedback loops, and blocklist status as a floor rather than a verdict.
• Seed-list placement as a directional read across a panel of mailboxes — a sample to interpret, not a count of your real recipients.

Alerting sits on top of these, so a threshold crossed at an awkward hour reaches a person rather than a log file. The skill is in reading them together: a clean complaint rate beside falling placement points somewhere very different from a blocklist hit beside a spike in bounces, and the correction each calls for is different too.

Where we differ from the rest of the market

Managed deliverability is an established category, and the senders shopping for it have usually met its dominant shape: an enterprise console, a team of consultants in a European capital, and a price that arrives only after a sales call. That model works for the brands it is built for. It tends to leave the serious mid-market sender — too large for a self-serve tool, too pragmatic for an enterprise contract — paying for a console they half-use and waiting on a queue for the answers they need. Three things set our service apart for that sender.

The first is that infrastructure operation is standard, not an upsell. Most managed-deliverability offerings monitor the sending and advise on it; when the problem turns out to be the engine — connection limits earning deferrals, streams sharing a reputation, backoff that loops instead of recovering — they hand you a recommendation and wait. Running PowerMTA and KumoMTA at volume is our core work, so we can act on the infrastructure layer directly. The data and the engine are read by the same people.

The second is independence. We resell no ESP, no IP space and no mailbox-provider relationship, so nothing in our advice is steered by a product we need you to keep buying. In a market where a great deal of deliverability guidance is attached to the thing being sold, a reading with no commercial pull behind it is worth more than it sounds.

The third is honesty about price. Published 2026 figures show the spread the category lives in: self-serve reputation tools around 149 US dollars a month, managed engagements commonly from about 1,100 a month, strategy audits near 2,500, and enterprise work quoted privately. We sit in the mid-market band on purpose, scope the retainer to your actual sending, and put the number in writing before you commit. There is no console licence you are quietly funding and no enterprise minimum to clear.

Behind all three sits the credential stack that lets us back the claims: work grounded in KumoMTA partnership, CSA membership and M3AAWG participation, the same standards bodies the serious end of this market answers to. And because we operate in Spanish and Portuguese as readily as in English, a sender expanding into Latin American or Iberian markets does not have to translate its deliverability problem before it can be solved.

How does it compare to an in-house hire?

Senders weighing this service usually weigh it against hiring, and for a large enough program the right specialist is a sound choice. The honest comparison has three parts. Cost is the first: an experienced deliverability hire is a senior salary on top of a tooling budget — reputation monitoring, seed testing and analytics that run into four and sometimes five figures a year — before the role has headed off a single incident. Coverage is the second: one person takes holidays, moves on, and cannot watch the queue through the night, while the knowledge inside a managed engagement does not leave with an individual. Breadth is the third: deliverability runs across authentication, data, content and infrastructure, and few people are equally strong in all four, least of all the infrastructure corner most careers never reach. A retainer buys a team that already spans those layers for less than the loaded cost of the hire who would still need the tools. For some senders the hire is right regardless, and we will say so; for most mid-market programs the arithmetic favours the retainer.

What managed deliverability will not promise

The category attracts overpromising, so the limits are worth stating outright. We do not sell a guaranteed placement number, because the mailbox providers decide placement and rewrite the rules as they go; any firm figure is one no provider will underwrite. We do not buy or rent reputation, and we stay away from engagement schemes built to trick filters, which a modern filter reads as manipulation and which usually end in a worse position than they began. We will not keep billing a retainer a sender has outgrown either — if your program reaches the point where an occasional check is enough, scoping down is the honest move. What we stand behind is the work itself: done by people who can act at every layer, and reported plainly enough that you can judge it for yourself.

What happens in the first ninety days?

Plain expectations serve a sender better than a sales promise. The first ninety days of an engagement are usually about recovery and instrumentation rather than records — closing authentication gaps, pruning the data that drags engagement down, and pacing any warm-up so reputation has room to climb. Placement that has slipped recovers gradually, because reputation is a moving average and providers reward a sustained pattern over a single strong week; Google’s own rule, where a domain above the spam line must hold below it for a run of clean days before mitigation returns, is that same principle written small. By the steady state the work turns from repair to defence: holding the gains, catching drift early, and absorbing the rule changes and infrastructure edits that would each otherwise cost a dip. The figure that matters is whether the people you are trying to reach are reliably seeing your mail, and that is the figure we report against.

Common situations we are brought in for

The trigger for a managed engagement is rarely abstract. A handful of situations account for most of the senders who arrive, and naming them tends to help a reader place their own.

• A migration or a new IP range went live, the early sends looked fine, and placement has been sliding ever since with no single obvious cause.
• A compliance deadline landed — the per-provider bulk rules, a DMARC enforcement step — and the team needs the change made correctly and watched afterwards, not merely shipped.
• An incident already happened: a blocklisting, a reputation drop, a run of deferrals from one provider, and the priority is recovery without making it worse.
• A sender outgrew an ESP’s shared infrastructure, moved volume onto its own MTA, and found that owning the engine means owning the deliverability that comes with it.
• A portfolio reached the size where eyeballing a few dashboards a week stopped being a plan, and the reputation of several brands now rides on someone watching properly.

None of these has to become a crisis first. The cheapest version of every one of them is the one caught while it is still a line moving on a chart, well before it shows up as a number the business can feel.

Who is managed deliverability for?

Managed deliverability earns its retainer for a particular kind of sender, and it is honest to say it is overkill for others. A small list mailing occasionally does not need it. The senders who do tend to share a profile.

• High-volume senders past the 5,000-a-day, per-provider threshold, where a compliance slip means bounced mail rather than a softer landing.
• Email service providers and agencies whose own clients’ reputations ride on the infrastructure underneath them.
• Multi-brand portfolios where one brand’s incident can pull the others down without per-brand isolation and watching.
• Programs where email drives enough revenue that a sixth of it going unseen is a number the business can feel.
• Senders moving into Spanish- and Portuguese-speaking markets, where the local provider landscape rewards operators who know it.

If that describes your program, the path is short. The free 25-point audit shows where placement stands today and whether a retainer is even warranted; if it is, we scope the engagement to fit. If it is not, we will tell you that too, because a managed service you do not need is not a sale we want.