Fix anything that blocks somebody completely first, then anything on the enquiry path, then everything else. Cost decides ties, not the order.

Two ways to order the list

The usual advice is to do the cheapest things first, which builds momentum and clears the count quickly.

That is a good way to start and a poor way to spend a fixed two days.

The alternative is ordering by how many people are stopped entirely, which puts different things at the top.

With two days, use the second ordering and let cost break ties rather than set them.

Blocking against degrading

That single distinction reorders most lists. A cookie banner that cannot be dismissed by keyboard blocks everybody navigating that way from every page, and a heading level out of sequence makes one page slightly harder to scan. An automated report gives both the same weight.

The blocking list

Which is short and goes first.

A popup or banner that cannot be dismissed without a mouse.

A form that cannot be completed or submitted without one.

A focus indicator missing entirely, which makes keyboard use guesswork.

A menu that only opens on hover, hiding half the site.

And a form that fails silently with no message.

Those five are the whole of the first day for most small business sites, and several will not apply.

Then the enquiry path

Since that is where the business consequence sits.

Form labels that are visible, permanent, and correctly associated with their fields.

Error messages that name the field and say what is acceptable.

Contrast on the form, the buttons, and the telephone number.

The telephone link being real, tappable text rather than an image.

Anything on the path from arriving to making contact outranks everything elsewhere on the site, because that path is what the site is for.

That is most of the second day.

A worked example

A business had an audit listing forty-one issues and two days of developer time available.

Ordered by cost, the first day would have gone on image descriptions and heading levels across the whole site.

Ordered by impact, the first morning went on the cookie banner and the focus indicator, which between them had been blocking keyboard use of every page.

The rest went on the form labels, the error messages, and the contrast on the enquiry page.

Twenty-six of the forty-one items remain open and the site is usable, which it was not.

The remaining twenty-six are all in the degrading category.

Count by people, not by instances

Which is the arithmetic underneath the ordering.

A problem in a template affects every page and a problem on one page affects one page.

A problem on the home page and the contact page affects nearly every visitor, and one on a page with four visits a year affects almost nobody.

So multiply the severity by where it sits and by how many people meet it.

That reasoning puts a template-level failure above a worse failure on an obscure page, which is correct and is not how audit reports are ordered.

What to leave

Stated explicitly, since the point of triage is deciding what not to do.

Heading levels on pages nobody visits.

Image descriptions on decorative images in a gallery.

Anything on a page you intend to delete.

And anything requiring a rebuild of a component supplied by a third party, which is a project rather than a fix.

Those are legitimate to defer and should be recorded as deferred rather than quietly forgotten.

The difference between a deferral and a gap is entirely whether it was written down at the time.

Give each deferred item a rough target, even a vague one, since an item with no date attached is one nobody will return to.

Two days is enough for most small sites

Which is the encouraging conclusion.

A five to fifteen page site with a form and a menu has perhaps six genuinely blocking problems, most of which are template-level.

Two days of competent work removes all of them and most of the enquiry path issues.

What remains is a long tail that can be worked through over a year at no urgency.

The site goes from unusable to usable in that time, which is the outcome that matters rather than the number in the report.

Get the blocking list yourself before paying anybody

Since identifying it needs no expertise and no tools.

The keyboard walkthrough finds the popup, the menu, the focus indicator and the form, which is four of the five blocking items in one fifteen-minute exercise.

Submitting your own form badly finds the fifth.

So the whole of the first day's work can be specified by an owner in under half an hour, which means the two days of paid time go entirely on fixing rather than on investigation.

A developer given a specific list works considerably faster than one asked to make the site accessible, and the difference is frequently a full day.

Tell the developer the ordering

Which sounds obvious and is regularly omitted.

Handed a list of forty-one items and two days, a developer will reasonably start with the ones that are quick, because nobody told them otherwise.

Say plainly that the five blocking items come first, the enquiry path second, and everything else is out of scope for now.

That single instruction is what converts two days into a usable site rather than into a lower number in a report.

Put it in writing with the list, since it also records why the remaining items were left.

The counter-case

Cheap wins have their own value.

Clearing thirty small items in a morning builds the habit and the confidence to continue, which a difficult first day does not.

Impact is also a judgement rather than a measurement, and reasonable people order these lists differently.

And a business with unlimited time should do both orderings rather than choosing.

Separate blocking from degrading, fix the blocking items and the enquiry path in your two days, and record what you deferred.

The two days

  1. Split the list into blocking and degrading.
  2. Fix anything that blocks completely.
  3. Start with template-level problems.
  4. Then the enquiry path.
  5. Form labels and error messages.
  6. Contrast on the form and buttons.
  7. Record what you deferred.

Step one is the whole method, since an automated report gives a dismissable-popup failure and a heading-order failure identical weight and they are not remotely the same problem.

The ordering by cost is covered in the cheapest accessibility wins.


Frequently asked questions

What are the two orderings?

By cost, which builds momentum and clears the count, and by how many people are stopped entirely. With fixed time, use the second.

What is the key distinction?

Blocking against degrading. A banner that cannot be dismissed by keyboard blocks every page; a heading out of sequence makes one page slightly harder to scan.

What is on the blocking list?

A popup that cannot be dismissed, a form that cannot be submitted without a mouse, a missing focus indicator, a hover-only menu, and a form failing silently.

What comes second?

The enquiry path: form labels, error messages, contrast on the form and buttons, and a real tappable telephone link. That path is what the site is for.

How do I weigh instances?

By people, not by count. A template-level failure affects every page; a failure on a page with four visits a year affects almost nobody.

Is two days enough?

For most small sites, yes. A five to fifteen page site has perhaps six genuinely blocking problems, most of them template-level.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Audit listing forty-one issues and two days available?

Split them into blocking and degrading first. The order changes completely.

Start a Conversation