Every migration has its own personality. Different platform, different integrations and different team dynamics. But the mistakes? Those are remarkably consistent. We've run enough to see the same patterns repeat across projects, industries and company sizes.
This isn't a theoretical list. These are mistakes we've watched happen, sometimes on projects we inherited, sometimes on projects where the warning signs were there but decisions were made to press ahead. The point isn't to alarm you. It's to give you a checklist of things to actively prevent.
If you're planning a migration, our **migration services page** covers what's involved.
Mistake 1: No redirect plan until the final week
This is the most common and most damaging mistake in ecommerce migration. A brand spends months building the new store, launches on schedule and then watches organic traffic drop by 30-40% because nobody mapped the old URLs to the new ones until the last few days.
URL structures change when you move platforms. If the old addresses aren't redirected with proper 301s, Google treats those pages as gone and the traffic disappears, with recovery taking months.
A full URL audit should begin at the start of the project, not the final weeks. The mapping itself is an ongoing process that gets refined as the new structure takes shape, with the redirect file finalised, tested in staging and validated before go-live.
Mistake 2: Going live on a Friday
A Friday go-live means that if something breaks, and something always needs attention in the first 24-48 hours, your team is dealing with it over the weekend. Meaning, customer service is reduced and the development team is harder to reach.
We've seen brands launch on a Friday afternoon and spend Saturday morning manually processing orders because an integration didn't reconnect properly. The fix took thirty minutes, but the issue happened at 8am on a Saturday when nobody was watching.
Launch on a Tuesday instead, you get a full working week to catch issues and stabilize before the weekend.
Mistake 3: Testing with production data too late
We've seen migrations where staging was built with sample data, ten products, three customers, no order history and the first time real data hit the new store was go-live day.
Real data surfaces different problems: variant pricing mismatches, accounts without order history, discount codes that didn't carry across or gift card balances that vanished.
Load staging with a full copy of production data well before launch, as the issues you find there are cheap to fix, whereas the same issues found after launch are expensive and visible.
Mistake 4: Forgetting about email and marketing integrations
The store migrates, design looks right and orders are flowing, sounds like success, right? Except your email platform is still pointing at the old store's API, your abandoned cart flows have stopped triggering and your Meta Pixel is firing on the old domain.
Marketing integrations are consistently the last thing teams think about and the first thing that causes problems post-launch. Build an integration inventory at the start. For every system that connects to your store: email, SMS, reviews, loyalty, analytics and advertising, document what needs to change. Test every integration in staging before go-live.
Mistake 5: No content freeze discipline
Most migrations require a content freeze in the final days before go-live. Any change made to the old store after the final data export won't appear in the new store.
The problem is that nobody tells the right people. The freeze schedule needs to go to every person who touches the store. Put it in writing or send it twice.
On a migration last year, the communication went to six people across three departments, and one person still updated product pricing the day before launch. We caught it because we were tracking last-minute changes but without it, those products would have launched at the wrong price.
Mistake 6: Rushing the timeline to hit a peak trading date
The board wants it done before Black Friday. The timeline is tight but "doable if everything goes to plan." Nothing ever goes to plan.
Compressed timelines lead to shortcuts, skipped testing, incomplete redirect mapping, untrained teams and integrations that get configured but never properly verified. The migration launches on time, technically and then the first two weeks are spent fixing issues that another month of preparation would have caught.
Migrations belong in quiet trading periods, Q1 or early Q2 is ideal for most ecommerce brands. If you can't avoid a deadline near peak, build in buffer that assumes at least two things will take longer than planned, because they will.
Migrations are complex. That's exactly why you need a partner who's done them before and knows where things go wrong. If you're planning a move to Shopify and want to talk through what's involved?
