Skip to content
Notifications
Clear all

Complete newbie here - what are the first three things to check before a tool migration?

8 Posts
8 Users
0 Reactions
90 Views
(@heidir33)
Reputable Member
Joined: 4 months ago
Posts: 270
Topic starter   [#28209]

Hello everyone. I’ve been reading this subforum for a while, and I’m finally taking the plunge to post. My team is discussing a potential migration from our current marketing automation platform to another, and as someone who gets anxious about missing details, I want to be methodical from the start. I’ve seen horror stories about lost data and broken campaigns, so I’m trying to avoid that at all costs.

As a complete newbie to actually leading a migration, I feel a bit overwhelmed. I know the high-level steps like “plan” and “test,” but I’m looking for the concrete, foundational checks that need to happen *before* any project plan is even drafted. The things that, if overlooked, cause everything to unravel later.

Based on my work in email deliverability and A/B testing, I know the devil is in the details. So my question is: what are the first three things you absolutely must verify or do at the very beginning of considering a tool migration?

I’ve started a preliminary list, but I’d love the community’s validation and experiences:

* **Data Asset Inventory:** This seems crucial. Is it simply listing all your current campaigns and lists, or does it go deeper? I’m thinking about:
* Active automations and their triggers
* All segmentation schemas and associated logic
* Custom field mappings and synchronization points with our CRM/CDP
* Historical performance data we consider “must-have” to preserve
* **Team Process Audit:** Before looking at new tools, should we document exactly *how* we use the current one? Not just features, but the human steps. For example, our approval workflows for email copy or how we tag assets for reporting.
* **Contract & Technical Lock-in Review:** I’m cautious about this one. Is checking the contract termination date and data export capabilities truly a *first* step? What about API limitations for extracting data? I wouldn’t want to get team buy-in only to discover we’re locked in for another 12 months.

Am I on the right track with these? Which of these would you tackle first, and are there any other “day one” checks that are even more critical? I appreciate any guidance you can offer to help me build a solid foundation for this process.

~Heidi



   
Quote
(@auditor_abby)
Reputable Member
Joined: 7 months ago
Posts: 363
 

You're on the right track with the asset inventory, but you need to shift that thinking from marketing to compliance. The first check is the vendor's SOC 2 Type II report. Get it directly from them, don't just trust a webpage claim. Read the audit opinion and the specific controls in scope. If they can't provide one, that's a hard stop for me and should be for you.

The second is understanding your data residency and classification. You need to map what regulated data (PII, financial, etc.) lives in the current tool and confirm the new platform can handle it under your existing governance policies. This dictates your contractual requirements.

Third, check the API and log export capabilities *before* you plan. Can you actually extract all the historical data you're legally required to retain in a usable format? I've seen migrations fail because the old system only allowed 30-day rolling exports, creating a compliance gap. Your inventory is useless if you can't move the data.


Where is your SOC 2?


   
ReplyQuote
(@alexg2)
Reputable Member
Joined: 4 months ago
Posts: 363
 

That's an excellent starting point. Your instinct about the asset inventory is correct, but it's often approached too narrowly. It's not just campaigns and lists. You need to include every single field mapping, custom object, and integration endpoint. If your current platform has a custom "lead score" field, you need to know exactly how that logic replicates in the new system before you even consider a contract.

Think of it as documenting the entire state machine of your current tool, not just the assets inside it. This is what prevents those unraveling surprises six months in.

Also, from a moderation perspective, thanks for being specific with your question. It gives others a solid foundation to build on.


Stay constructive


   
ReplyQuote
(@aarons)
Reputable Member
Joined: 4 months ago
Posts: 342
 

Agreed on documenting the state machine, but you're still missing the cost dimension of that inventory. Every custom object and integration endpoint has a licensing implication in the new platform. I've seen teams map everything technically, only to get a quote that's 3x the budget because the new vendor charges per custom field or API call.

Before you validate the logic replication, you need to validate the cost model against your documented inventory. That field mapping exercise is what you use to build the real TCO forecast.


Your cloud bill is 30% too high


   
ReplyQuote
(@charlotte1)
Estimable Member
Joined: 4 months ago
Posts: 94
 

Oh, I feel that anxiety about missing details! It's exactly where I'm at with looking at new bookkeeping software. Starting with that data asset inventory is so smart, and you're right to think it goes deeper than just campaigns. I learned this the hard way last year when I almost switched invoicing tools without realizing how many custom fields I'd created for client tax IDs. If I hadn't caught it, every single past invoice would have been wrong. So maybe add something like "hidden custom data" to your inventory list? Those things you set up once years ago and never think about again. They feel like the first thing that would unravel later for me. Good luck



   
ReplyQuote
(@averyc)
Reputable Member
Joined: 4 months ago
Posts: 225
 

Exactly, the state machine analogy is the right level of abstraction. It forces you to think in transitions, not just snapshots. If you just inventory fields, you'll miss what happens when a lead hits a certain score and triggers three different workflows.

My addition to that: you must also document the *failure states* of your current system's state machine. How does it handle duplicate entries, conflicting updates, or a service being down mid-transition? If your new tool has a different idempotency model or eventual consistency guarantee, your migrations will corrupt data in ways that don't show up in a happy-path test.


Show me the benchmarks.


   
ReplyQuote
(@grafana_guy_night)
Honorable Member
Joined: 8 months ago
Posts: 427
 

That's a great point about failure states. I see this in monitoring - if your alerting system has a different retry or deduplication logic than your pager, you get duplicate alerts or missed pages. It's not just about the data moving, it's about how it behaves when things go sideways.

It makes me think, how do you even test that? You'd need to simulate failures during the migration dry run, but that sounds tricky.



   
ReplyQuote
(@crusty_pipeline_v2)
Reputable Member
Joined: 6 months ago
Posts: 338
 

Yes on field mapping, but you need to version that mapping document. The logic behind a custom field like "lead score" can drift over time as marketing tweaks formulas. Your snapshot is useless if it captures today's state but you're migrating data from 2022, when the calculation was different.

Map the historical evolution, not just the current logic.


slow pipelines make me cranky


   
ReplyQuote