Pointing Fanzy AI at a storefront you have already been running

Your storefront has been live for about a year. You have not really looked at it in four months — you send the link, people buy things, and the page has become furniture. Then someone screenshots it in a reply and you see it properly for the first time since February: the card at the top is the one that stopped selling in spring, there is a product you no longer offer, and the bio line still describes a business you have quietly moved on from.

Why pages drift

Storefronts do not decay because the design ages. They decay because your business moves and the page does not. You add a product in a hurry and it lands at the bottom. You stop offering something and leave the card up because removing it feels final. Your prices go up but the description still implies the old ones. Each of these is a two-minute fix that nobody makes, because none of them ever becomes urgent.

The harder problem is that you cannot see your own page any more. You have looked at it several hundred times, so your eye goes straight to whatever you changed last and skips everything else. That is the specific job an assistant is good for here: not taste, but attention. It has no memory of what the page used to look like, so it reads what is actually there.

The genuinely difficult part is knowing what to keep. A year-old page contains decisions that were right for reasons you have forgotten. If you hand over a vague brief and accept whatever comes back wholesale, you will lose those along with the drift. So the refresh has to be argued rather than accepted, which takes longer than a rebuild and is worth it.

Rebuilding, or asking for a second opinion

Most people redesign by starting again, because starting again is emotionally simpler than auditing. The cost is that you throw away everything the page has learned, including the copy that quietly does the persuading and the card order that reflects a year of messages.

How to run the review

Set aside an hour. Do not open the designer for the first twenty minutes of it — the numbers come first, because they decide what you are going to ask for.

1. 1Screenshot the page as it is today. On a phone, not a laptop. This is your before, and in six weeks you will not otherwise remember what it looked like.
2. 2Write down the per-product numbers. Views, buyers and revenue for each item over the last 90 days, plus the traffic source breakdown. Full analytics come with Starter, and this step is the whole reason the refresh will be better than the original build.
3. 3Mark each card as earning, learning or dead. Earning pays for its position. Learning is new and needs more time. Dead has had views and no buyers for a quarter, and its problem is the offer, the price or the words — decide which before you redesign around it.
4. 4Brief Fanzy with the evidence, not the aspiration. "Three products, this one carries 70% of revenue, this one gets views and no sales, my traffic is mostly from one platform on weekday evenings" gives it something real to arrange.
5. 5Ask for layout and order first, theme second. There are 11 layouts, and moving from a dense grid to a stacked list changes behaviour far more than swapping between the 28 themes does.
6. 6Take exactly two changes. Write down the rest on a list for next quarter. Two is enough to move a number and few enough to attribute.
7. 7Update the copy on the dead card last. Rewriting words on a product you might discontinue is a good way to spend an evening achieving nothing.
8. 8Diary a check for two weeks later. If you do not book it, you will either forget or fiddle in the meantime.

A worked example: three products, one carrying the page

Imagine a running coach with a 15 training-plan PDF, a 90 monthly membership and a 40 one-off video call. The numbers say the membership is most of the revenue, the PDF sells steadily in small amounts, and the call has had views all year and hardly any bookings. The page currently leads with the PDF, because it was the first product built.

The briefed second opinion returns a stacked layout with the membership first and a longer description under it, the PDF second as an entry point, and the call third with the order-form questions surfaced so people can see they will be asked what they are training for. Two changes get taken: the order, and the membership description. The call is left alone deliberately, because the hypothesis is that its problem is the price rather than its position, and testing two things at once would tell you nothing. A fortnight later the numbers say whether the reorder did anything. If the membership share moves at all, that is worth more than a new font, and memberships billing monthly or yearly compounds a small change over time in a way a one-off sale does not.

The mistake most people make

The mistake is redesigning when the actual problem is upstream. A quiet quarter feels like a design problem because design is the part you can control on a Sunday afternoon. Usually it is traffic — the platform that used to send people has changed what it shows — or it is an offer that made sense a year ago and does not now. Check your traffic sources before you touch the theme. If views are down and conversion is flat, the page is fine and the problem is distribution, which is a different job entirely and is not solved in the storefront designer.

A page that has been live a year is not a page that needs replacing. It is a page with a year of evidence attached to it, which makes it the easiest thing in your business to improve — provided you bring the evidence to the conversation rather than your feelings about the font. An hour, two changes, a fortnight of holding still. And if the audit suggests the page should be doing more than it currently does, the features list is a reasonable place to see what else could sit on it.