The morning after the store moves, customers find what the project didn't test.

Tuesday, 8:15 a.m.

This is a composite, but anyone who has moved an online store will recognize it.

The new store went live on Monday night. It's faster, it looks better on a phone, and the checkout finally works the way the team wanted. The project lead went to bed tired and pleased.

At 8:15 on Tuesday, Priya, who runs customer support, opens the inbox with a coffee. The first message is from a customer of six years: "I can't log in. It says my email isn't registered." The second asks where her order history went. The third has a gift card from last Christmas that now shows a balance of zero. A subscriber writes to say his monthly coffee didn't renew, and another says hers renewed twice. Someone clicked a link in an old newsletter and landed on a page that says it doesn't exist. And a wholesale buyer, who has always seen trade prices after logging in, is looking at retail.

None of this was on the launch checklist. The checklist was about the new store working. These messages are about the old store's promises arriving intact.

That's the pattern with most platform moves. Your new customers meet the new store. Your best customers meet everything you forgot to bring with you. The people who write in on the first morning are the ones who have bought from you for years, which is exactly who you least want to disappoint.

"I can't log in"

Passwords usually can't come across in a platform move. They're stored in a scrambled form that the new platform generally can't use, so customer accounts arrive as records waiting to be activated. That's normal. What isn't fine is customers discovering it at checkout.

Plan it like a small campaign. Decide when activation invitations go out, in what order, and what they say. Send them in batches rather than one blast at midnight, starting with your most active customers. Make the login page explain what happened in one plain sentence, with a reset link right there. Give support a ready answer before the first message arrives.

Where history and balances go

Customers keep score in ways the project may not have listed. Before you migrate, write down every place a customer has something with you:

• Order history. Even if old orders can't be fully recreated, customers should be able to see them, and support needs them for returns and warranty questions.
• Gift card balances and store credit. These are money you owe people. Reconcile the totals before and after the move, to the cent.
• Loyalty points, saved addresses, wishlists. Small things, until they're missing.
• Subscriptions. The card details behind recurring orders are typically held by your payment processor, not your store. Moving them is a planned handoff that involves the processor, and it takes time. Start that conversation early, and decide what happens to renewals scheduled during the move week.
• Trade or wholesale pricing. Customer groups, price lists and payment terms need to come across with the accounts they belong to.

Old links

Your old store's addresses live in more places than your sitemap: past newsletters, ads, social posts, other people's blogs, printed inserts, and QR codes on packaging that's sitting in customers' cupboards.

Google's guidance on site moves is clear and worth following. Make a map from every old address to its new equivalent. Use server-side permanent redirects. Don't send lots of old pages to the home page, which confuses people and may be treated as a missing page. And keep the redirects for a long time, generally at least a year. You can read the full guidance on Google Search Central.

Build the map from data, not memory. Start with the pages that get the most visits and the ones other sites link to, then add product, category and policy pages. Check printed materials and old campaigns for any special addresses that only exist there.

Pay particular attention to products that changed during the move. If two old product pages became one with variants, both old addresses should lead to it. If a product was discontinued, send its old page to the closest sensible alternative or category, or let it say plainly that the item is gone. And test the map before launch by running every old address through it and checking where each one lands. A redirect that points to a page that doesn't exist yet is worse than no redirect at all.

The quiet integrations

Behind the storefront, a store talks to other systems all day. Order confirmation and shipping emails. Tax calculation. The feed to your warehouse or 3PL. Your accounting software. Reviews. Analytics and ad tags. Marketplace and shopping feeds. A wholesale portal.

Each of these was set up once, years ago, often by someone who has since left. In a move, they're easy to break and easy to miss, because they don't show up when you click around the new store. They show up when the warehouse doesn't receive Tuesday's orders.

List them, name an owner for each, and rehearse with realistic orders that travel all the way through: placed on the new store, confirmed by email, taxed correctly, received by the warehouse, shipped, and recorded in accounting. We wrote about planning the cutover itself in write the cutover on one page before the freeze.

The week before, test what customers touch

Here's the short list we'd run in the last week, with real staff accounts and real-looking orders:

• Twenty existing customers, including a wholesale account, activate their accounts and log in.
• The most visited old addresses, plus every address on printed materials, each land on the right new page.
• Ten gift cards and store credits show the right balance and can be spent.
• A subscription renews once, on the right date, for the right amount.
• An order placed on the old store is returned through the new one.
• A full order travels to the warehouse, ships, and appears in accounting.

That fifth test is our favorite. It catches problems with history, refunds, stock and support tools all at once, and it's exactly what a loyal customer will try in the first month.

We follow the order all the way through in our ecommerce work, from the storefront to the systems behind it. If you're planning a move and want a second pair of eyes on what customers will notice, talk to us before the freeze. Deciding whether to move at all is a different question, and we've written a letter about that, too: replace the old system in pieces.

Related work: Howcast.