A skincare brand says its new store cost it 30% of sales overnight. What to check before you sign a replatform
Adorn Cosmetics’ founder says sales fell about 30% when its new store launched, with paid ads failing to reach the site. Her disputed account raises questions to settle before you sign a replatform.

Treat a replatform as a revenue event, not a design project. Before signing, write down why you need to move, find out what the agency recommendation actually means, and tie sign-off, pause and exit rights to launch criteria you can measure. Before launch, map every old URL to a 301, test every ad, email and social destination on the new site, stage the switch with a rollback ready, and agree in advance the conversion drop that sends you back.
What Adorn's founder says happened
Briony Kennedy has spent 17 years building Adorn Cosmetics, a Melbourne skincare brand that 7NEWS reports was turning over about $6 million a year at its peak. In the segment available on November 8, she said that, barring a miracle, the company would close “tomorrow.” That was her warning in the interview, not confirmation that it has closed.
Her account, told on air and in her first LinkedIn post, goes like this. In 2021 the store was running on open-source Magento and, in her words, thriving. She wasn't looking to rebuild. She wanted to be sure customer data was secure. She says Adobe pitched its paid platform, Adobe Commerce, put a projection of 34% to 94% revenue growth in writing, and recommended a local agency, Convert Digital, to build it. She puts the Commerce cost at about $15,000 a month.
When the new site launched, she told 7NEWS, sales dropped by about 30% almost immediately. She said the Instagram and Facebook ads she was paying for, roughly a million dollars a year of them, weren't linking back to the website. She says she asked repeatedly to leave or pause the contract without success. In another post, she says revenue fell 17% in the 2023 financial year and nearly 37% the following year, leaving it below half its earlier level. 7NEWS reports that she laid off staff, sold the family home, and rebuilt the store on Shopify. The sources use dollar figures without explicitly naming the currency; no currency conversion is applied here.
These are Kennedy's allegations and 7NEWS's reporting, not findings that establish what caused the losses. According to 7NEWS, a spokesperson from Adobe headquarters in the United States said they would not comment on this case for legal reasons. The report says mediation failed; Kennedy says her legal team is commencing court proceedings. The cited material contains no response from Convert Digital. We don't know what happened inside that project, and this note isn't a verdict on anyone.
The practical question is broader than this dispute: when a working business moves platforms, how do you keep a website project from becoming an unplanned revenue risk?
Ask whether you need to move at all
Kennedy's advice on 7NEWS was to question a change if things are working, rather than chase the new shiny thing. That is worth taking seriously before anyone talks platforms.
Write down, in a sentence or two, why the store has to move. "Our current platform can't do X, and X is costing us Y" is a reason. "The new one is more modern" or "a vendor told us we'd grow" is not yet a reason. It's a pitch.
If the worry is security, as hers was, there are usually cheaper ways to deal with it first: get the current platform patched and up to date, move to better-managed hosting, or pay for a security review. (When the PHP2Shell flaws landed last month, we wrote about how to tell whether a WordPress or Magento store is patched. That kind of work usually costs far less than a rebuild.)
If someone hands you a growth projection, ask who wrote it, what it assumed, and whether they will put any of it into the contract. A number in a sales deck is something to discuss. Something they'll be held to has to be in the agreement. If the answer is "it's only an estimate," that tells you how much weight to give it.
Decision guide
What the pitch says, and what the contract should say
| Consideration | What the pitch says | What the contract should say |
|---|---|---|
| Growth | A projected revenue range on a slide. | Launch criteria measured against your own baseline, tied to sign-off and final payment. |
| Partner | We recommend this agency. | Named team, named references at your size, and what the recommendation does and doesn't mean. |
| Problems | We'll fix issues after launch. | A right to pause work and payments if the criteria miss an agreed date. |
| Leaving | Not mentioned. | Exit with handover of code, data, admin access and documentation. |
| Cost | One monthly platform price. | When the licence starts, how it relates to the build contract, and what it costs to stop. |
Before launch day, open every door your paid traffic will arrive at.
Find out why that agency was recommended
A platform recommending a build partner feels like a safety check. Kennedy says that is exactly how it felt to her. Before you lean on it, ask the platform plainly what the recommendation means. Did they vet this agency for a store your size, or is it a listed partner in their program? Those are different things, and it is fair to ask which one you're getting.
Then interview the agency the way you'd interview anyone you're about to hand your revenue to:
- How many migrations have you done from our current platform, at roughly our size, in the last two years? Can we call two of those clients?
- Who is actually on our project, and how much of the work goes to subcontractors?
- What did launch week look like on your last migration, and what broke?
- From day one, do we hold the admin logins, the hosting account and the code repository, or do you?
That last question matters more than it sounds. Kennedy writes that only when she finally got full access to her own site could she have another agency look under the hood. If you can't get someone independent to look at your store whenever you like, you can't check what you're being told.
Write the way out into the contract before you need it
Most rebuild contracts are clear about what you pay and when. Fewer say what happens if the new store makes less money than the old one. Before you sign, get a lawyer to read it (this note isn't legal advice), and make sure these are in it:
- Launch criteria you can measure. For example, checkout conversion within an agreed range of your own baseline for a set number of days after launch. You pick the numbers. Final payment or sign-off waits until they're met.
- A right to pause. If the criteria aren't met by an agreed date, you can stop new work and stop paying for it while fixes are made.
- A right to leave, with a clean handover. Code, data, admin access, documentation, and an agreed amount of help moving to someone else.
- When the platform bill starts. A platform licence and an agency build are often two separate contracts. Check both. Paying full price for a new platform while the old store still takes the orders is a slow leak.
None of this is aggressive. A good agency will have answers ready, because a store with its revenue protected is a client that stays.
Treat launch day as a traffic event
The allegation in Kennedy's story that should worry every store owner is the ads. Paid traffic doesn't care what the new site looks like. It cares whether the link it was given still opens a page that sells something.
We've written before about the morning after a store moves: logins, store credit, old newsletter links. This list covers the money you are still spending while the switch happens.
- Map every old URL to its new one with a permanent (301) redirect: products, categories, collections, blog posts, and the odd landing pages someone built for a campaign three years ago.
- Pull the actual destination URLs out of every ad account (Meta, Google, TikTok, Pinterest, whatever you run) and test each one against the new site before cutover. Do the same for email automations, links in bio, affiliate links and the QR codes printed on your packaging.
- Check that tracking survives. Redirects shouldn't strip campaign tags. Purchase events should still fire. If the ad platform stops seeing sales, its algorithm starts spending your budget blind.
- Launch in stages if you can. Go live on a quiet day. Keep the old store running and ready. Write down who can call a rollback and what triggers it before anyone flips the DNS.
Sequence of work
From signing to the first quarter
- Before signing
Written reason to move, agency references called, exit and pause terms in the contract.
- During the build
You hold the admin, hosting and code access; an independent reviewer can look in.
- Two weeks out
URL map with 301s; every ad, email, social and QR destination tested; tracking checked.
- Launch day
Quiet day, old store kept ready, rollback owner named, baseline on the wall.
- First 90 days
Weekly check against baseline by channel; agreed drop pauses spend or rolls back.
Agree on the number that sends you back
Before launch, write down what normal looks like: the last eight weeks or so of conversion rate, revenue per visit and orders per day, split by channel (paid social, search, email, direct). That's your baseline. Without it, "sales feel slow" turns into weeks of arguing about whether it's the season.
Then agree, in writing, on the drop that triggers action. Something like: if paid social conversion is down more than an agreed amount for two days running, we stop the campaigns pointing at the new site, or we roll back. Name the person who makes that call. Someone senior who isn't on the agency's side of the table.
Keep watching after the first week. Kennedy describes a decline that, after the first hit, kept grinding on for two financial years: fewer visitors buying, more budget going to fixes, less left for ads. A launch-day crash is loud. A slow slide is easy to blame on the market, so look at the baseline every week for the first quarter.
Put this into practice
If you're being pitched a replatform, or you've already signed one, start with three things this week:
- Write the one-sentence reason you're moving, and the growth number you were promised, next to each other. See whether the contract says anything about either.
- Export every ad, email and social destination URL you're currently paying to send people to. That list is your launch test.
- Pull your eight-week baseline by channel and agree the number that pauses spend or rolls back.
Our ecommerce work starts from the order rather than the homepage: discovery, checkout, payment, inventory and the people answering customers. If you want a second pair of eyes on a migration plan or a contract before you sign, send us what you've got and we'll start with the parts that carry your revenue.
Related work: Paddle8.
We build alongside the people doing the work.
Bring us the problem you’re thinking about.