Fashion Nails the Parts Nobody Sees and Flies Blind on the Rest

Fashion builds the tidiest small business websites I audit.

Cross that sentence out, because the rest of the afternoon took it apart. It was true for about an hour, though, and I want to show you the page that had me convinced.

One page. I opened it more or less at random, and I could have opened forty others that look the same.

If you sell clothes, I’d put money on your homepage sharing most of what comes next.

The page as I found it

A womenswear shop. The fall drop runs across the top of the homepage, five photos deep, a model on a wet sidewalk in a coat I’d buy if I were six inches taller.

New In, Dresses, Knitwear, Sale, Stockists. A padlock sitting beside the address.

Scroll past the photographs and a form is waiting, four boxes under a line inviting you to ask about stock and sizing, which is more than most homepages in any industry bother to offer a stranger.

Then I opened the code underneath, and I stopped being smug about small business websites for a while.

92.9% of the fashion sites I audited carry a canonical tag, where 84.9% of everything I audit does.

One dress can end up living at three web addresses without anybody meaning it to. Filter the rack down to black, then sort it by price, then arrive from Tuesday’s email with a tracking code stuck on the end of the link.

A canonical tag settles which of those three counts, by pointing at one and calling the other two copies.

I’ve aimed one of those tags at the wrong address myself. A client’s phone went quiet for a month before I worked out why, and I’ve never billed anybody for that month since.

Nobody in that shop typed it, and nobody ever will. The store software writes the line the second a product goes live, which is how a trade full of owners who have never heard of the thing ends up 2 of 47 on it.

The rest of the code told the same story. Nothing missing, nothing broken, and nothing anyone in that shop had ever sat down and chosen.

Code that arrives free can leave the same way, and I deleted a line of it during a therapist’s redesign once without noticing for months afterwards.

Then the form. 43.8% of these shops put one on the homepage, against 28% of small businesses.

Someone decided a stranger with a question at eleven at night should have a way in, and that decision, which puts this trade 3 of 47, is the only human fingerprint on the entire page.

So I sat there thinking this trade had quietly solved the website problem. Have you ever been that wrong about something for a whole hour?

Now the same page, one afternoon later

I want to change one thing about that homepage and show it to you again.

The photos are the same. The menu is the same. The form is the same form, in the same spot, with the same boxes and the same button.

What’s different is a handful of lines in the code that you would have to go hunting for, and a setting telling the form to file a note every time a stranger hits send.

Not one customer will ever see the difference, and that’s exactly why the change never gets made, because nothing about it feels like an improvement to the person paying for it.

But at the end of the month, the second version answers a question the first one can’t. Of everyone who opened the coat, how many typed their name into the form?

Turning it on is the shortest job on my entire list. It happens once, it belongs to whoever set the store up, and after that it runs unattended for years.

It’s on the list I work through before a site goes live, and it takes less time than picking the homepage photo.

44.0% of fashion sites run it, where 58% of small businesses do.

So a trade that hands a search engine flawless code keeps no record whatsoever of the people, a check where fashion lands 46 of 47.

Try asking your web person which page on your site gets opened most often. The answer, or the length of the pause before it, tells you which version you own.

Or skip the asking, because five minutes on your phone tonight will tell you whether your shop is counting anything at all.

“But the store already shows me my sales”

It does, and the sales screen is the most comforting thing on the internet, because everything on it is money that already arrived in your account.

What it can’t describe is the week that produced it. How many people opened that jacket and put the phone face down on the couch, and where were they before they landed on you?

A shop floor answers that for nothing. You watch a woman take the jacket off the rail, hold it against herself in the mirror, turn the tag over, and hang it back up two sizes along from where she found it, and by Friday you understand something about that jacket that no receipt would have told you.

Your website is that same floor with the lights off and the staff sent home. The till still works. Nobody is standing near the mirror.

So which of the two pages I described is yours? You knew the answer before this paragraph, I suspect.

Why one half came free and the other didn’t

The tidy half was installed. The other half has to be chosen, by you, on a Tuesday, with nothing forcing your hand.

How many of the checks on your own site did you pick? Be honest with yourself, because your answer is the whole difference between the two pages.

Nobody in that shop had to believe in canonical tags for the shop to have them. They arrived with the theme, unasked for, the way the size chart and the checkout did.

Analytics needs a person to decide it’s worth an afternoon. Then it needs a second person, or the same tired one, to open the thing every so often and care what it says.

Deciding is the expensive part. It always is.

And nothing ever comes along to force the issue. A missing tag eventually turns up as a warning inside a tool that someone opens, but a shop with no idea where its customers came from looks identical to a shop that knows, right up until the season the customers stop arriving.

When did anyone last hand you a page off your own website and ask whether it was doing its job?

Weight my ten checks the way I always weight them and fashion comes out at 70 out of 100, 35 of 47, which is a middling place for a trade sitting near the top of two checks.

That number is doing something dull and honest. The trade banks every point the software hands out, then gives back the points that need an owner to make a call.

What the ones I like do instead

The fashion sites I admire aren’t better looking than the page I described. A few are worse looking, honestly, and one of my favorites has a homepage photo I’d have argued about.

I keep a shelf of examples here, and what they share is boring. Someone is watching them.

A watched page gets changed. The photograph nobody opened comes down in March, the dress everybody opened moves to the top, and the form loses a box because whoever was reading the reports worked out that the box was frightening people away.

One client of mine moved a single photograph in February, from the third row of her homepage into the first, because her report told her that particular dress was getting opened more than anything else on the page. Sales on it climbed for a month afterwards.

She didn’t redesign anything. She didn’t hire me to redesign anything either, and I’ve mostly stopped pretending that bothers me.

Six months of that beats any redesign I could sell you, and I say that as a man who sells redesigns.

Rebuilding stores stuck in the first version is a good slice of my work, and the first meeting is never about code. It’s me asking what happened last spring and the owner telling me what she remembers, which is not the same thing.

Memory is a terrible instrument. It keeps the loud week and drops the quiet one, and it always flatters whatever you were proud of at the time.

So here’s where I leave you

Your own homepage, last month.

Someone opened it on a bus, looked at one dress, and left. Or she bought the dress. Or she typed her email into your form and never heard a word back from anyone.

Which of those three happened, and how many times?

I can’t tell you from out here, and your page can’t tell you either. Can you?