White-glove deliverability
Deliverability is not a feature you switch on. It is an ongoing relationship between what you send, who you send it to, and what the receiving side makes of it - and when it goes wrong, a dashboard tells you that it went wrong rather than what to do next. Omnivery assigns a senior deliverability analyst to work with you from onboarding onward, on a regular call cadence, and includes it in the service rather than invoicing it as a support tier. That covers everything you send. Marketing and transactional sending build their own reputations and are reported separately, and the campaign side is where the volume and the complaint exposure concentrate - so an analyst watching only one of the two streams is watching half of your sending.
Updated: July 2026
Three positions that shape how the service works. Two of them are unpopular in this industry.
Deliverability is a shared resource. On a platform that accepts anonymous free signups, your password resets share sending reputation with whatever those accounts send, and you inherit the consequences without ever seeing the cause.
Omnivery has no free tier by design, and every customer and every sending domain is reviewed by staff before sending, normally within one to three working days. That vetting step is not administrative friction; it is the mechanism that makes a shared pool worth being in.
A dedicated IP is widely sold as the premium option. It only outperforms a well-run shared pool if you can keep it warm with consistent volume. Below that threshold you take on all of the reputation risk and none of the volume that establishes trust, and every gap in your sending is a fresh cold start.
It is also the weaker signal of the two that matter. Most of your reputation attaches to the domain in your DKIM signature rather than to the IP, and receivers weight it accordingly, so your brand's standing accumulates against your signing key and travels with it.
Omnivery's position is that dedicated IPs are an industry upsell that underperforms for most senders, and that a vetted shared neighborhood does better. If your volume genuinely justifies a dedicated IP, that is a conversation to have with a senior deliverability analyst rather than a default to buy.
Omnivery's own research puts roughly half of B2C email opens as non-human, and they have been since Apple Mail Privacy Protection landed. The B2C bot click rate rose from about 2% in 2023 to about 16% by 2026. Bot-click rates differ sharply by provider: around 30% at Outlook against 2% to 4% at Yahoo and GMX.
If your sunset rules, reactivation triggers and segmentation run on opens and clicks, they are partly running on machines. Bot Detection separates the two, which is a deliverability tool as much as an analytics one - because sending to an inactive segment that only looked active is how reputation degrades. The full findings are in The State of Email Bots 2026.
Most delivery platforms optimize for Gmail, Outlook, Yahoo and Apple. That covers the majority of addresses on most Western lists, and as a way to allocate engineering effort it is entirely defensible - right up to the point where your growth is coming from somewhere else.
Omnivery's position is that a destination is not less important because it is smaller. The infrastructure sends according to each receiving side's requirements rather than applying one global sending profile, and the team maintains provider relationships reaching niche markets worldwide. A national provider with a few million users, a regional ISP, a corporate gateway - each has its own filtering behavior, throttling, authentication expectations and escalation path, and to a platform tuned for the big four they are rounding errors.
A blended figure hides the case that matters. If placement is 99% at Gmail and 70% at the provider your fastest-growing market actually uses, the headline looks healthy while your password resets and order confirmations quietly fail in one country. Nobody notices, because the average absorbed it.
Reviewing placement per receiver rather than in aggregate is therefore the first thing a senior deliverability analyst does on a standing call. It is also the reason consolidating your sending matters: split across several providers, the receiver quietly deferring your mail appears in nobody's report.
Omnivery's bot research tracks 126 mailbox-provider families across six world regions and reports behavior per provider rather than in aggregate: Italy's Libero has held under 1% bot traffic every year since 2022 and Germany's web.de sits at 3 to 4%, against roughly 30% at Outlook and over 20% at Gmail. Producing figures at that resolution requires watching each receiver individually, which is what the same infrastructure does for customer sending.
Omnivery grew out of Mailkit, founded in the Czech Republic in 2006. That is why escalation is possible at receivers with no public contact route. Membership of M3AAWG covers the global side, and of the Certified Senders Alliance - the European sender accreditation scheme, which carries particular weight with German and wider European receivers - and Signal Spam in France, the European side.
Europe is where this stops being a principle and starts costing money, because it is the most fragmented mailbox market there is: national providers hold meaningful share market by market and do not behave like the large ones. See EU data residency, or e-commerce for retailers specifically.
The large cloud platforms have commoditized email sending - competing hard on price per thousand messages - and recovered the margin somewhere else. Deliverability is where it went. Human expertise became a product line, sold back to you as consulting, technical account management and strategic review.
You can see it in how the offers are packaged. Mailgun splits its business into two products, Send and Optimize: the sending platform is one purchase and the deliverability tooling is another. Inbox placement and reputation monitoring are listed as Optimize features on Mailgun's own pricing page, and the dedicated technical account manager sits inside Optimize's deliverability services rather than in the sending plan you already bought. SendGrid sells deliverability help as Expert Services consulting packages, and its ongoing-consulting page now redirects to Twilio Professional Services - which is where that expertise has ended up organizationally, in a professional services unit that bills separately from the API.
Both are coherent businesses, and neither observation is a criticism of the engineering underneath. But it does mean the person who understands why your mail is failing sits on the other side of a purchase order, and that the incentive to keep your sending healthy is not the same incentive that sold you the plan.
Omnivery has no deliverability business unit, because deliverability is the business. A senior analyst is part of the service from the day a sending domain passes vetting - included, not invoiced. There is no tier to upgrade into, no day rate, and no separate statement of work. If your sending is healthy we have done the job we were paid for, and there is nothing further to sell you.
That is also why the vetting is strict and why there is no free tier. In a shared, vetted neighborhood every customer's behavior affects every other customer's delivery, so the analyst work and the admission policy are the same commitment seen from two directions.
Competitor packaging described as published on the vendors' own pages in August 2026. Vendors restructure their offerings; treat their current pricing pages as the authority.
The phrase "dedicated support" usually means a faster queue. This is a different thing: a named senior deliverability analyst who knows your sending program, has seen its history, and is in regular contact whether or not anything is wrong.
The analyst is involved from onboarding rather than appearing after the first incident. That means going through the sending domains and their authentication records, the message categories and whether they belong on the same domain, the volume shape you expect, and the segments you plan to send to. Most deliverability problems are decisions made at this stage, and they are far cheaper to change before there is a reputation attached to them.
Then there is a standing call - weekly, every two weeks, or monthly, depending on volume and how much is changing. The point of a recurring slot is that it does not require an incident to justify it. Reviewing a slow drift in placement, a segment that is quietly decaying, or a receiver that has started deferring is a fifteen-minute conversation on a Tuesday and a multi-day problem if you wait for it to become visible in the numbers.
Typical ground covered on those calls:
Sending is monitored around the clock, and the analyst reaches out when patterns look wrong rather than waiting for a ticket. That is the part that is difficult to buy separately: an outside consultant sees your data when you send it to them, and a support tier responds when you notice something. Someone inside the platform watching your traffic sees it first.
When a receiver needs to be contacted, an individual sender has no relationship to draw on. Omnivery is a member of M3AAWG, of the Certified Senders Alliance - the European sender accreditation scheme - and of Signal Spam, and maintains provider relationships reaching regional and niche markets. Escalation goes through those channels rather than a public contact form.
None of this is a separate line on the invoice. It is part of the service, which is a deliberate choice: charging for deliverability help means the customers who most need it are the ones who decline it, and their sending shares a reputation with everyone else's.
What working with a senior deliverability analyst looks like over time, rather than as a list of features.
Every customer and sending domain is reviewed before sending, normally within one to three working days. The analyst uses that window for the setup review: domains and authentication, which message categories belong together, expected volume shape, and whether transactional mode is needed - it changes suppression behavior and is enabled by the team rather than self-serve.
Early volume is where a program establishes its reputation, so it gets the most attention. Test-mode sends via X-OV-testmode exercise the path first, then real volume ramps with placement and bounce behavior reviewed per receiver rather than in aggregate.
A standing call is agreed - weekly, every two weeks or monthly, depending on volume and rate of change. It runs whether or not anything is wrong, so drift gets caught while it is still cheap to correct.
Sending is monitored around the clock. If a pattern looks wrong - a receiver starting to defer, a complaint rate stepping up, a segment behaving unlike itself - the analyst makes contact rather than waiting for the next scheduled call or for a ticket.
A seasonal peak, a new sending system, a migration off another provider, or a big segment change is raised in advance. Domain rate limits auto-increase by 20% at 75% utilization, which covers organic growth; a step change is better handled by raising the ceiling before the event.
Escalation to a receiver goes through Omnivery's M3AAWG and Certified Senders Alliance channels and its direct provider relationships, not through a public form. The analyst already has the context, so the conversation starts from what changed rather than from who you are.
Three support models, not three vendors. All three are legitimate, and they fail in different places.
| Platform support tier | External consultant | Omnivery analyst | |
|---|---|---|---|
| Cost model | Priced tier or per-incident | Day rate or retainer | Included in the service |
| Knows your sending history | Whatever is in the ticket | What you send them | Yes, from onboarding onward |
| Sees your data | On request | Exports you provide | Directly, continuously |
| Contacts you first | Automated alerts | No | Yes, when patterns look wrong |
| Regular scheduled review | Rarely | If retained for it | Weekly, biweekly or monthly |
| Involved before first send | No | If engaged early | Yes, during vetting and onboarding |
| Can escalate to receivers | Sometimes | Via own relationships | M3AAWG and CSA channels |
| Can change platform behavior for you | Sometimes | No | Yes, including rate limits and domain settings |
An external consultant with deep specialist expertise is a genuinely good option for a one-off diagnosis. The differences are continuity and access to the sending data. Platform support tier reflects how the large cloud providers package deliverability help as published on their own pages in August 2026: Mailgun sells it as the separate Optimize product, and SendGrid sells it as Expert Services consulting through Twilio Professional Services.
The most common case. Email works until it does not, and then it becomes urgent for whoever is nearest. A standing call with someone who already knows the program replaces a scramble.
Migration is when reputation is most fragile, because sending identity and volume shape both change at once. Kiwi.com moved off SendGrid and Mailgun in 45 minutes and recorded a 17% improved click rate afterwards.
Retail, travel and anything with a campaign calendar. A volume step is manageable when it is planned and a problem when it is discovered.
Where the national and regional mailbox providers a global platform treats as tail traffic are actually your audience, each with its own filtering behavior and escalation path.
Sunset rules, reactivation triggers and lead scoring built on opens and clicks are partly built on machines. Separating human from non-human interaction is a deliverability decision, not just an analytics one.
Where a delivery failure is a compliance or operational event rather than a metric, and the useful question is what changed rather than what the dashboard says.
Included. Omnivery assigns senior deliverability analysts as part of the service rather than selling them as a support tier or a per-incident add-on.
Charging for deliverability help means the senders who most need it are the ones who decline it, and on a shared sending pool their reputation problems become everyone else's. Making it standard is cheaper than the alternative.
A named person who knows your sending program and is in regular contact whether or not anything is wrong. They are involved from onboarding - reviewing sending domains, authentication records, which message categories belong together and the volume shape you expect - and then hold a standing call at a cadence agreed with you.
The distinction from "dedicated support" is that a support tier responds when you notice something. A analyst sees your sending data continuously and makes contact when patterns look wrong.
Weekly, every two weeks, or monthly, depending on volume and how much is changing. A program in migration or approaching a seasonal peak warrants more frequency than a steady transactional stream.
The value is in it being a standing slot rather than incident-driven. Reviewing a slow drift in placement or a segment that is quietly decaying is a short conversation if it happens on schedule and a multi-day problem if you wait for it to show up in the headline numbers.
Probably not, and that is usually the right answer. A dedicated IP outperforms a well-run shared pool only if you can keep it warm with consistent volume. Below that threshold you carry all of the reputation risk with none of the volume that establishes trust, and every gap in sending is a fresh cold start.
There is a more fundamental reason the question matters less than it used to. Most of your sending reputation is not attached to the IP at all - it is attached to the domain in your DKIM signature, which receivers weight more heavily than the address a message arrived from. Your brand's standing accumulates against your signing key, and it travels with you. An IP is the weaker signal, which is why buying a dedicated one rarely moves placement the way it is sold as doing.
Omnivery's position, then, is that dedicated IPs are an industry upsell that underperforms for most senders, and that a vetted shared neighborhood does better. That depends entirely on the vetting: every customer and domain is reviewed before sending and there is no free tier, so the pool is worth being in. If your volume genuinely justifies a dedicated IP, discuss it with a senior deliverability analyst rather than buying it by default.
Because the vetting is the product, not the paperwork. Every customer and every sending domain is reviewed by staff before sending, normally within one to three working days, and Omnivery has no free tier.
On a platform that accepts anonymous signups, your transactional email shares sending reputation with whatever those accounts send. The review step is what makes a shared pool a benefit rather than a liability, so it is worth planning around in a migration timeline rather than treating as friction.
That is the part Omnivery is built around. Delivering into Europe means the regional and national mailbox providers a global platform treats as tail traffic, each with its own filtering behavior and escalation path.
Omnivery grew out of Mailkit, founded in the Czech Republic in 2006, maintains provider relationships reaching niche markets, and is a member of M3AAWG and of the Certified Senders Alliance - the European sender accreditation scheme, which carries particular weight with German and wider European receivers. Escalation goes through those channels rather than a public contact form.
Assume a large share of it is not. Omnivery's own research finds roughly half of B2C email opens are non-human and have been since Apple Mail Privacy Protection landed, and that the B2C bot click rate rose from about 2% in 2023 to about 16% by 2026, with bot clicks running around 30% at Outlook against 2% to 4% at Yahoo and GMX.
That matters for deliverability, not just reporting: if a sunset rule or reactivation trigger runs on inflated engagement, you keep sending to addresses that are not actually engaged, which is how reputation degrades. Bot Detection separates the two. The findings are in The State of Email Bots 2026, also available as a PDF.
You should already be compliant when it does. Omnivery holds sending standards above what mailbox providers currently require, on the reasoning that the direction of travel in email security is predictable and a requirement met in advance costs nothing to meet. When the 2024 Google and Yahoo bulk sender rules landed, requiring DMARC enforcement, one-click list-unsubscribe and a spam rate under a defined threshold, much of the industry issued advisories with a deadline attached. Omnivery already required all three, and its customers reconfigured nothing.
That extends past the mandated requirements. Omnivery validates DANE on outbound delivery, so where a receiving domain publishes TLSA records the receiving server has to present a matching certificate or the message is not delivered - transport security that most senders never have to ask for because it is handled underneath them.
The same posture applies to what is coming rather than only to what has arrived. DKIM2 is being developed as the successor to today's DKIM, and it is a larger change than the version number suggests. Where DKIM signs the headers and the body, DKIM2 also signs the recipients, which closes the replay problem: a DKIM signature today can be captured and reused to send somebody else's mail under your authenticated domain, and signing the recipients makes that class of attack impossible rather than merely harder. It also gives each delivery status notification a secure and traceable path back through the servers that handled the message, and adds a way for a receiver to reconstruct a message to its earlier form when something in the path has modified it, so a legitimate change stops breaking the signature.
Most of that is the receiving side's problem to implement. From a sender's point of view the useful part is shorter. DKIM2 requires no DNS changes, because it reuses the DKIM keys you already publish. And because it is clearly better than what it replaces, receivers are expected to push for adoption quickly rather than over years, which makes it the next requirement worth meeting in advance. Omnivery is preparing support now so it can be rolled out as soon as the standard is settled or receivers begin testing against it, rather than after they start expecting it. Being early to a standard is the pattern rather than the exception here, and for a sender it means an authentication change arrives as a platform update rather than as a project.
Not necessarily. A specialist consultant brought in for a deep one-off diagnosis - a full audit, a reputation recovery, a complex authentication problem across many domains - is a genuinely good use of money, and an external perspective has value.
What an analyst provides that a consultant structurally cannot is continuity and direct access to the sending data: they see your traffic as it happens rather than in an export, they were there when the program was set up, and they can change platform behavior on your behalf. The two coexist comfortably.
If your current arrangement is a dashboard and a ticket queue, the useful question to ask a provider is who will be on the call and how often. Vetting takes one to three working days, and the senior deliverability analyst is involved from that point rather than after the first incident.