Do companies, especially start-ups, evolve over time? What happens when, once a local company, you want to go BIG in the international market? Depending on your company’s profile, a great many things come with growth. But at the same time, the same company undergoes lots of changes in the product, people, and financial processes.
Hi! I’m Andrei, an Engineering Manager at FINN, and today I’d like to focus on a challenging tech topic: migrating to a whole new international domain—usually to .com
. This article is about how we successfully migrated from finn.auto
to finn.com
.
The need for a .com
domain: How did we get here in the first place?
Typically, internationalization isn’t a top priority for companies when they are founded. You want to focus on more important things, like growing your business and making it successful. For the most part, it’s actually helpful for your business to settle for the common domain ending of your country, especially in the early days. Most German start-ups use .de
. We chose .auto
as it matched our business profile. .auto
isn’t very international, and people tend to trust .com
more. This may be because of the early days of the Internet when the vast majority of websites would be found on .com
domains. Regardless of the reason, it became an expectation and might hurt your international business if not considered.
The complexity of domain migrations
How complex are these domain migrations? I’m personally aware of three important German start-ups that have faced this challenge in the past, and I wouldn’t be surprised if this turns up to be a common topic for successful businesses around the globe.
Here at FINN, we faced the same challenge. It wasn’t easy, and it never is. Domain migrations are a complex endeavor, even when the impact is isolated. For us, it impacted internal operations, authentication systems, and communication with the customers and vendors. We also had to migrate with no downtime.
Here’s another issue with domain migrations: they’re usually irreversible. The reasons range from announcing the change to your customer base and employees to huge data inconsistencies that, in the end, lead to the conclusion: it’s better to fix everything that goes wrong after the migration than to revert it. But reasons aside, there are so many moving parts that it’s safe to trust it’s an all-in kind of thing.
Our approach
And so, we at FINN went all-in. Not on a bluff, but with the safest hand we had. Here is why our bet made sense. First of all, a major project like this can’t take place in isolation in the IT Administration team or department. It affects the entire company from top to bottom, and without strong support, you can’t get anywhere. This is probably one of the most difficult things to obtain: everyone’s support! It made all the difference that our team had the time to research, test, and think about the best course of action, but most notably that we could rely on the support of others to answer our questions and react quickly when needed.
From the technical standpoint, we ended up splitting the migration into two parts:
- Moving all customer and vendor-facing communications from finn.auto to finn.com, starting with our website.
- Moving all the rest: the internal operations and authentication systems.
Our learnings
One of our key learnings is that it’s critical to have an excellent and deep understanding of all the software you use. With the help of key software tool owners across the company, we were able to compile a list that included all tools from the largest to the smallest. This served as our base of operations because all of these elements were inspected thoroughly in order to understand what would and could happen once the migration started. Whenever we couldn’t find the answer ourselves, we asked for the support of the software vendors.
All information was carefully documented with the goal of knowing exactly what to do and in what order. This included tasks that had to be delegated to other teams outside of IT Administration as well. All information was accessible to everyone at the company at all times. We wanted to ensure that everyone was aware of the current progress and next steps. Moreover, all action items were defined and assigned to their respective owners in the form of tasks. In the end, we gathered a total of 46 individual tools and about 150 tasks. I’m pretty sure we didn’t document every single one of them, just the major ones. Moreover, some teams used their own tracking systems so the final number could easily be doubled.
On the day of the migration, we managed to migrate all business-critical systems from finn.auto to finn.com in about five hours on a Friday evening. It was definitely a team effort, supported by a lot of pre-prepared automation that helped us save a lot of time on that day. Together with the pre-compiled documentation, we were able to organize ourselves excellently, to such an extent that nearly nobody noticed that the migration happened besides the clear domain name change. In the end, no matter how well you prepare, you need to brace yourself for the unexpected. We didn’t have any major issues, but we were delayed by around one hour. The reality is that in this type of situation, you’re constrained a lot by the software you’re working with as well. When pushing changes, for example, we faced API rate limitations, so we couldn’t push the changes as fast as we would have liked.
Success!
All in all, we planned to reach 80% migration completeness over the weekend but managed to do so on Saturday at noon. The remaining 20% required software vendor intervention, which could only happen during the regular working week, no earlier than Monday morning.
We also needed the full post-migration week to tie in loose ends, but we’ve been operating on finn.com ever since with no issues whatsoever. Did we enjoy the process? Absolutely! Would we do it again? No! It’s a massively complicated process that requires the excellent coordination of teams both internally and externally, as well as perfect communication and technical skills. Even so, it was still a big challenge to reach our end goal. It required an entire quarter of planning and another of execution. Was it worth it from the business perspective? Absolutely! Our service demand skyrocketed. Does it make sense technically without a good business forecast? The answer is probably no. It’s not one of those things to do just for the looks of it but it does pay back. For those of you courageous enough to accept the challenge of a domain migration, we’ve compiled the list below of key items that we believe you need to be successful. Cheers!
List of key items to get you started
- Obtain the top-to-bottom buy-in on the domain migration (from the CEO to the Internship levels).
- Commit to the migration. This is an “all-in” operation.
- Detailed migration plan (all software tools with clear status, steps, action items for everyone, and test observations).
- Pre-prepared automation for everything (as much as possible).
- Set and communicate expectations on all levels at all times.
- Establish clear communication channels during the migration day.
- Migrate! 🙂
- Fix immediate issues.
- Test, test some more, then test again.
- Fix remaining issues.
- Clean-up!