I want to start with a conversation I have had WAY too many times.
An organisation switches email platforms, the sales process was smooth, the demo looked great, the onboarding team were helpful. They go live. And then, a few weeks later, the cracks start showing. Automations that were running fine are not firing. Contacts who should be excluded are getting emails. The forms are not routing correctly. The deliverability has taken a hit nobody warned them about. And the engagement metrics in the new platform look completely different from the old ones, in ways nobody can explain!
They are not back to square one — they are further back than square one, because now they have a broken programme, a team under pressure, and a platform contract they have committed to.
This happens constantly (all the flipping time). And almost all of it is preventable — not by choosing the right platform, but by understanding what choosing a platform actually involves and doing the right things in the right order before, during, and after the switch.
This blog covers all of it. How people get the selection wrong. The questions to ask before you sign anything. What "plug and play" actually means versus what people assume it means. What a migration actually involves. And the ten things most likely to go wrong — because I have watched them go wrong, repeatedly, and the pattern is almost always the same.
The first mistake happens before anyone has touched a new platform. It happens in the selection process and it is almost entirely driven by how ESP sales and demos are designed to work.
The demo shows you the best version of the platform. Features are demonstrated in ideal conditions, with clean data, perfect integrations, and a presenter who knows exactly which buttons to press. The platform looks simple. Powerful. Exactly what you need.
What the demo does not show you: how it performs when your actual data is imported with all its inconsistencies. How it behaves when your specific integrations are connected. What the experience is when something breaks and you need support. What the platform looks like six months after you go live and the onboarding team has handed you over.
Most buyers walk out of a demo thinking about features. The right question is not whether it has the feature you need. The right question is whether it handles your specific situation, with your data, your integration dependencies, and your team capability level.
Buying based on what a platform can do rather than what it can do for you
Every mainstream ESP can do most of what you need. The features are not the differentiator. What differentiates them is how those features work in your context, with your CRM/CDP, your forms, your contact fields, your automation logic, your team's technical comfort, and your commercial calendar.
I have seen companies choose platforms because they have a feature they were excited about in the demo, then discover that feature requires a plan tier they cannot afford, or depends on a data structure they do not have, or works completely differently in practice from how it worked in the presentation. The feature existed. It just did not exist for them, in the way they needed.
The phrase that causes more problems than any other in ESP sales
"Plug and play" in ESP sales language means: this integration exists and has been built. It does not mean: this integration will work seamlessly with your specific data setup, in real time, without configuration, for your specific use case.
There is a world of difference between a static integration — one that syncs data on a schedule, or requires a manual export/import, or only pushes data in one direction — and a real-time, bidirectional, dynamic integration that reflects live CRM data in your email platform at the moment an automation fires.
The distinction matters enormously for how you design automations. If your integration is a sync rather than real-time, any flow that depends on a contact's current CRM status — deal stage, lifecycle stage, a field that changes when someone does something — may not work the way you expect. The data it is checking may be hours old. The conditional logic fires against a snapshot, not live reality.
Most teams only discover this after they have built their automations and started testing. At which point the flow needs to be redesigned, sometimes fundamentally.
The questions that matter are not in the feature list. They are in the operational detail. Here is what to ask every ESP before you commit.
What a migration actually involves and why it takes longer than you think
A migration is not moving your email list from one place to another. That is the smallest part of it.
What you are actually moving is an entire email infrastructure: every contact, every field, every suppression list, every form connection, every automation, every template, every integration, every Zapier connection, every trigger, every piece of conditional logic, plus setting up authentication on a new domain, warming up a new IP, rebuilding benchmarks, and monitoring everything daily for weeks after go-live.
When I scope a migration with a client, the work falls into six phases. Each one has to be done before the next one starts. Skipping ahead, building automations before you know your integration is real-time, for example, or importing contacts before you have imported suppressions — is where the expensive mistakes happen.
Before you touch the new platform, understand exactly what you have
The first thing to do is not open the new platform. It is document everything about your current one. Export all your campaign performance data now — because once you lose access, it is gone. Note your current open rate, click rate, unsubscribe rate, and complaint rate by email type. These are your benchmarks — you cannot know whether the new platform is performing well or badly unless you know what "normal" looked like on the old one.
Run an inbox placement test before you move. If you have an active deliverability problem, migrating will make it worse, not better. Moving to a new IP while your sending reputation is damaged is like moving house to escape a bad credit score — the debt comes with you. Resolve spam issues before you switch, not after.
Check your SPF, DKIM, and DMARC records. If they are not correctly set up now, you are starting the migration from a weaker position than you should be.
The step most teams skip and the one that causes the most problems
Before you build a single automation in the new platform, you need to know exactly how your integration works. Get the technical documentation — not the sales page, the actual integration docs — and read it.
The critical question is whether your integration is real-time or a sync. If a contact's CRM deal stage changes, when is that reflected in the ESP? Immediately? In an hour? At the next daily sync? The answer determines what kinds of automations are possible and what kinds need redesigning.
Map every form on your website that captures email addresses. For each one: how does it currently connect to the ESP? Will that connection still work with the new platform? If you are using Zapier to connect forms to your current ESP, those Zaps point to the old platform — they will need to be cloned and updated. Clone them, do not edit them. You need the originals running until go-live day.
The order is not optional!
When you import your contacts, the order matters more than almost anything else. The first thing you import into the new platform is not your active subscribers. It is your suppression lists.
Your unsubscribes must be imported as suppressed contacts before any active contacts go in. Your hard bounces must be imported as suppressed. Your spam complainants must be imported as suppressed. If you import active contacts first and then your suppressions, there is a window — even a brief one — where the platform does not know those people are suppressed. If an automation fires in that window, you risk sending a marketing email to someone who previously unsubscribed, which is a compliance problem and a legal risk under PECR and UK GDPR.
Build your field mapping document before you import anything. Go through every contact field in your current ESP — every native field, every custom field — and map it to the equivalent in the new platform. Fields that do not have a direct equivalent need to be created in the new platform before import. You cannot map data to fields that do not exist yet.
You are not just moving lists. You are rebuilding an IP reputation from zero.
When you move ESP, you almost certainly move to a new sending IP — or a new IP pool. Inbox providers have seen behaviour from your old IP over months or years. They have no history of your new IP at all. Suddenly sending your full list volume from a brand new IP, with no sending history, looks from the outside exactly like the pattern associated with spam operations. Even if every contact is legitimate and permissioned, the providers do not know that yet.
The warm-up plan is the process of building that reputation gradually. You start by sending only to your most engaged subscribers — the people who open and click reliably, from the new platform, while continuing to send to the rest of your list from the old platform. Week by week, you increase the proportion sent from the new platform, monitoring deliverability as you go. If placement stays healthy, you increase. If it drops, you pause and investigate.
A standard warm-up takes eight to twelve weeks to reach full volume safely. During that period, you need to keep your old ESP active. This surprises people, they expect to switch cleanly on day one. But the overlap is essential. You cannot safely send your full programme from the new platform on week one without risking your deliverability.
The thing that most often derails a warm-up plan is the commercial calendar. If you have a major campaign, a peak period, or a high-volume appeal coming up within twelve weeks of your go-live date, either delay the go-live or plan to send that campaign from the old platform. Sending high volume from an unwarmed IP during a critical commercial period is a deliverability risk that can take months to recover from.
These are not hypothetical. Every one of these is something I have seen happen, more than once, in real migrations.
Platform migrations go wrong because they are treated as technical projects with a go-live date, rather than as programme transitions that require planning, sequencing, monitoring, and patience.
The checklist I have built for this, linked above, covers everything in the right order. But the mindset matters as much as the sequence. The questions to ask before you sign. The integration audit before you build. The suppressions before the active contacts. The warm-up before the full volume. The old platform kept running until you have proof the new one is working properly.
Every shortcut in a migration becomes a problem later. Sometimes a small one, sometimes a large one. The teams that navigate migrations cleanly are not the ones with the best ESP — they are the ones who took the process seriously enough to do it in sequence, ask the uncomfortable questions early, and resist the pressure to go live before they were ready.