The hour a day nobody put a price on

Somebody in your business spends an hour a day inside your website’s admin. Adding products, updating stock, processing enquiries, publishing news, fixing what somebody else typed wrong.

That hour is the largest recurring cost your site has. It is also the part of the project nobody specified, nobody reviewed, and nobody tested — because the person who commissioned the site is not the person who has to live in it.

How to tell it is bad without looking

You do not need to audit anything. There are symptoms, and they are reliable.

Somebody keeps a parallel spreadsheet. This is the loudest signal there is. It means the system does not do something the work requires, and a human has built a workaround in Excel. That spreadsheet is now the real source of truth, and the website is a copy of it that is always slightly out of date.

There is a written instruction sheet next to the monitor. Somebody documented a sequence of steps because the interface does not make them obvious. Every new hire is trained from that sheet, and the sheet is wrong in at least one place.

The same field is filled in wrongly every week, by three different people. Not carelessness. When everybody makes the same mistake, the interface caused it.

One person is the only one who can do a task. Not because it needs judgement, but because only they remember the order. That is a single point of failure created by an interface, and it takes a holiday.

Requests come to you by email. “Can you change the price on this one?” A site that requires a developer for routine changes is a site that will stop being updated, because eventually the client stops asking.

Why it happens

The economics of a web project push in exactly the wrong direction.

The public site is what the client sees, what gets shown to their board, what the agency puts in its portfolio. It gets the design attention and the revisions. The admin is whatever the CMS produced by default.

The person paying is usually not the person using. The owner approves a design; the office manager lives with the consequences and was never asked. In most projects nobody in the room during the requirements meeting has ever done the daily job.

And the cost is invisible in a way that a slow page is not. Nobody measures the time spent in the admin, so it never appears in any budget. Twenty minutes a day is eighty hours a year. That is two working weeks of somebody’s salary, spent on friction, every year, forever.

What actually fixes it

The method is embarrassingly simple: watch somebody do the work.

Not a requirements meeting. A screen share, twenty minutes, while they do a normal task while narrating it. That reliably reveals more than an hour of discussion, because people stop mentioning the frustrating parts once they have adapted to them. You will watch somebody open three tabs, copy a reference number out of one, paste it into another, and describe it as “yeah, that’s just how it works”.

What comes out of that session, in my experience, is always the same five categories.

List screens that show the wrong things. WordPress gives you title and date. The person doing the work needs status, owner, branch, last modified, the reference number they search by. Columns are configurable, filters are configurable, and almost nobody configures them. This is an hour of work and it is the highest-return hour in the project.

Fields in the wrong order. Editing screens tend to be organised by what was easy to build. They should be organised by the sequence of the work — the order the information arrives in real life. If a form asks for a delivery date before the item, somebody is scrolling up and down on every single record.

Validation that reports instead of prevents. If a field must be filled when the status is “published”, the interface should say so at the moment of the choice, not after saving. Better: make the invalid state impossible to reach.

Missing bulk actions. Anything done one row at a time, more than ten times, should be a bulk action or an import. Changing a price on forty products individually is not a small annoyance, it is an afternoon.

Roles that do not match the organisation. Everybody an administrator is how sites get broken by accident. Roles should mirror actual responsibilities, so that the person editing news cannot reach settings that can take the site down.

What this is worth

Use round numbers. Twenty minutes a day saved, one person, is about eighty hours a year. Two people, two working months.

Against that, a few days of work: configuring columns and filters, reordering fields, adding validation, writing one or two bulk actions, setting up roles.

I have never done this exercise and had it not pay back inside the first year. And the second-order effects are larger than the time: fewer errors reaching customers, less training for new staff, no single person who is the only one who knows the sequence, and a client who updates their own site instead of emailing me.

That last one matters more than it looks. A site whose owner can maintain it stays current. A site that needs a developer for every change is three years out of date within three years — not because anybody decided to neglect it, but because asking is enough friction to stop the asking.

Ask one question

If you take one thing from this: go and ask whoever uses your admin every day what the most annoying part of it is.

They will have an answer immediately. They have had it for months. Nobody has ever asked, so they assumed it was normal.