A failed website migration is rarely the platform’s fault
In Short
- Broken ad links, missing redirects and lost search traffic happen on any platform when the cutover is handled badly.
- You don’t need to be technical to vet a migration, check who profits from the platform recommendation and call three businesses the agency has already performed a similar service for.
- Never accept a fixed budget during a sales pitch; demand a rigorous upfront discovery phase first to map out the real risks.
- Highlight in your migration plan who is watching key numbers like sales, from hour one, and keep the old site ready to switch back.
- Watch the 7NEWS report: “Cosmetics empire crumbles after Adobe website disaster”
Earlier this week, Seven News ran a story about a Melbourne-based skincare business closing down after a website migration went wrong. The clip has racked up a lot of media attention, prompting various outlets and influencers to cover the situation, including Mumbrella, who asked me what I made of it.
Most of the commentary I’ve seen online has either simplified this to be about things like software, which platform the business was on, which one it should have chosen or which e-commerce tool is best.
Having watched many, many migrations, both small and relatively simple, through to large, complex, and mission-critical migrations, I think a lot of the commentary misses the most important point.
What the reporting missed out
The business at the centre of this story that hit the mainstream media was Adorn Cosmetics, founded by Briony Kennedy. According to Seven News, the business turned over about $6 million a year at its peak, selling its products online.
From the reporting, the business was persuaded by Adobe to move from its existing Magento e-commerce platform to the more premium, upmarket Adobe Commerce platform.
From what the Seven News report included, the business was told by Adobe that they would see revenue increases of between 34 and 94 percent, working with a local partner to build out and perform the migration. When the site launched, Kennedy told Seven News that sales dropped instantly by 30 percent. She also says the business was spending around $1 million a year on Facebook and Instagram advertising that, all of a sudden, stopped working.
After that, it sounds like the business went through what I would refer to as a rescue migration, moving over to Shopify. Since then, the business owner has had to lay off staff and sell the family home. By the time this segment aired, it sounded like the business was on the verge of closing. Adobe told Seven News it couldn’t comment for legal reasons, and mediation has reportedly failed.
What the broadcast doesn’t tell us is where it all went wrong, who was responsible for what, what the agency was contracted to deliver rather than Adobe, and perhaps most importantly, how long those problems existed before somebody stepped in.
None of us will know the answers to these questions, and I don’t think there is any point in trying to guess who is at fault.
The same mistakes can happen on any platform
When you move any sort of website, you aren’t simply moving pages and products, there are a lot of technical aspects that happen behind the scenes. You are recreating or accounting for every URL, configuring the payment gateways, setting up your analytics pixels and ad connections bolted onto the business.
If page URLs change without proper redirects pointing old links to new ones (something Google documents in detail), search traffic evaporates. If product feeds aren’t updated or conversion tracking isn’t re-tested, Meta and Google keep spending your ad budget while sending shoppers to dead ends or flying blind on what’s converting. Miss any of those steps and you get the exact nightmare Kennedy described, that’s whether you are moving to Adobe, Shopify or WordPress.
It happens far more often than people realise. Search Engine Journal analysed 892 domain migrations and found it took an average of 523 days for a moved site to recover its old organic traffic, with 17 percent still unrecovered after 1,000 days. Meanwhile, the fastest migrations bounced back in under five weeks. That study looks specifically at domain moves using third-party traffic estimates, so take the exact averages with a grain of salt.
What was particularly odd about this situation is that it appears this site moved from Magento (acquired by Adobe) to the Adobe Commerce platform. In a software context, migrating within a single vendor’s ecosystem should be about as low risk as you can get, as it’s moving from one known entity to another.
“From what’s been reported, I don’t think the Adobe platform killed this business, the migration did. And migrations fail the same way on every platform.”
Ben May, Founder of The Code Company
Why upfront discovery is one of your best insurance policies
Every agency will happily walk you through a 100-point migration checklist during a pitch. The problem is that unless you build websites for a living, a great checklist and a terrible one look almost identical.
This is why you should be cautious of any agency that hands you a fixed budget and a firm timeline during the initial sales phase. When an agency estimates a complex data migration or replatforming before they have actually looked under the hood, they are guessing.
To mitigate that risk, you need to separate the scoping from the sales pitch. At The Code Company, we refuse to provide firm delivery costs while we’re still just talking. Instead, we mandate a dedicated, upfront discovery and roadmapping phase first. We sit down with a client’s stakeholders and break down every single requirement, legacy data structure, and third-party API. We map it all out until there is as little ambiguity as possible about what needs to be moved and built. Only then do we attach a finalised budget and timeline to the delivery.
If the partner you are vetting isn’t demanding a rigorous discovery process before you sign the master contract, you are the one carrying the risk of a blown budget.
You can’t judge the tech, but you can ring three references
Beyond demanding a real discovery phase, you can evaluate track record. Before you sign anything, ask the agency for the names of three businesses roughly your size that they migrated onto the platform they are pitching you.
Then actually pick up the phone and call them. Don’t ask whether they like the agency, ask how the site responded in the weeks after launch, what broke, who noticed it first, and how fast it got fixed.
If an agency struggles to give you three names, that tells you everything you need to know before a single line of code is written. If they do hand them over, a ten-minute conversation with another owner who has been in the trenches will tell you more than any proposal ever could.
Someone has to watch the money from hour one
The part of the Seven News story that nags at me most is the timeline we don’t get to see. A 30 percent drop in sales on day one isn’t a subtle blip; it’s an alarm bell. The real question is what happened in the hours and days immediately after.
Before any migration goes live, someone needs to own the commercial numbers, not the technical dashboards showing that servers are online and pages are loading, but things like actual revenue, order volume, conversion rates and paid ad traffic. Put a specific person’s name next to those numbers and make sure they are watching them from hour one.
Just as importantly, agree in advance on what happens if those numbers tank. A steep drop on launch day should trigger a rollback conversation that afternoon, not weeks or months down the track.
Of course, you can only roll back if you have somewhere to roll back to. A website migration shouldn’t be a process that can’t be reversed. Keep the old site on standby and ready to roll back to for weeks after launch, not just a few hours. Paying for an extra month of overlapping hosting feels like an unnecessary cost at the time, but it could be the cheapest insurance you will ever buy.
When we migrated US publication The Dispatch off Substack, there were more than 35,000 paying members, 100,000+ readers, 10 newsletters and six podcast shows riding on the cutover. At that scale, you don’t measure launch success by whether the homepage loads. You watch member logins, payment webhooks and newsletter delivery in real time from the minute the switch flips, with a clear fallback plan ready if anything stalls.
If the stakes are high enough, bring in an independent set of eyes to sanity-check the migration plan before you commit, ideally someone who doesn’t make a bigger margin if you buy the more expensive platform.
For a publisher, the numbers are different but the failure is the same
Most of the clients we work with are digital publishers rather than retailers, and while the metrics look different, the failure pattern is identical.
In e-commerce, a botched cutover shows up immediately in lost orders. For a publisher, you see it first in search and referral traffic, a few days later in programmatic ad revenue, and eventually in subscriber churn. When a masthead has fifteen years of archives, sometimes hundreds of thousands of URLs, a huge chunk of daily readership lands on older articles. Miss those redirects and that audience vanishes overnight. Even if traffic holds up, an ad stack that loads slightly slower or in the wrong slots will quietly gut your yield, and if existing subscribers hit a broken login screen on Monday morning, you won’t hear about it until cancellations spike.
When we migrated Pedestrian.tv off a fragmented custom CMS onto WordPress, we had to move more than 50,000 articles, 300,000 media assets, 90,000 user accounts and 80,000 job listings. Because the legacy data was structured inconsistently, we engineered a live API service to export, translate and test the content iteratively before cutover rather than hoping a bulk import would hold together on launch day.
The safeguards don’t change: know who is watching the commercial pulse, agree on what triggers a rollback, and keep the old site alive until the dust settles.
Before you sign
None of this is to say the choice of platform never matters. Sometimes a business genuinely ends up on the wrong software. An enterprise system with enterprise licence fees can financially crush a company turning over a few million dollars, and getting locked into a contract you can’t exit is a serious commercial risk. Those are fair questions to raise in a case like this, even if we don’t have all the facts.
In my experience, choosing the wrong platform is a problem that manifests slowly. It shows up over years as a site that becomes harder to change and more expensive to run. Losing 30 percent of your sales the day you launch is almost certainly a migration failure.
Before you sign a contract, spend less energy debating platforms and more time vetting the people who are going to get you from point a to point b. Expect a real discovery process, call their references, put someone in charge of watching the revenue from minute one, and make sure you can revert and go back to the drawing board.
If you’re researching or considering a large, complex or high stakes migration and want an independent second opinion, this is the work we do. We’d love to chat.
