How Far Back to Support Old Browsers

The question is not what to support

It is what happens to somebody using something old.

There are three possible outcomes and only one of them is a problem.

The page works identically, which requires supporting the older browser fully and costs the most.

The page works, and looks plainer. Some visual refinement is missing and everything functions.

The page is broken. Content is unreadable, the menu does not open, or the form cannot be submitted.

The second outcome is the target. Aiming for the first everywhere is expensive, and the third is what happens when nobody thought about it at all.

Find out who actually visits

Before any decision, look at your own data rather than at general statistics.

Analytics reports browser and version, and for most small business sites the answer is overwhelming: current versions of a small number of browsers account for nearly all traffic, with a long tail of individually negligible entries.

What matters in that tail is whether any of it converts. A browser used by two percent of visitors who produce no enquiries is different from one used by two percent who do, and for a business serving older customers that distinction can matter.

The other number worth checking is operating system, since somebody on an old operating system may be unable to update their browser at all, which is a different situation from somebody who simply has not.

What supporting old browsers actually costs

Not just testing time, which is the visible part.

It frequently means shipping extra code to every visitor so that a small number get a working page. Compatibility layers, fallback scripts, and older syntax are downloaded by everybody, which makes the site slower for the ninety-eight percent in order to serve the two.

That trade is worth making explicit, because it is usually invisible. The cost of broad support is paid by your typical customer in load time.

The practical position

For a small business brochure or service site.

1. Support current and recent versions of the major browsers properly.
2. Ensure older browsers get readable content and working links, forms, and phone numbers.
3. Accept that visual refinements degrade, such as newer layout features falling back to something simpler.
4. Do not ship compatibility code for browsers your analytics show nobody uses.
5. Never block or scold somebody for their browser.

That last point is worth stating. A banner telling a visitor their browser is unsupported is frequently shown to somebody who cannot change it, on a work machine or an old device, and it converts a plain page into an insult.

Building so that this is cheap

The technique that makes the whole question less important.

Starting with content that works in anything, then adding refinement on top, means an old browser receives the plain version automatically rather than a broken one. Newer layout and styling features are designed to be ignored by browsers that do not understand them.

The failure comes from the reverse approach, where the layout depends entirely on a modern feature and there is nothing underneath it. Then an unsupported feature means no layout rather than a simpler one.

This is a build decision rather than a testing decision, which is why it is worth raising before a site is made rather than after.

What to actually test

A short list, since exhaustive testing is not available to a small business.

The current version of the browser most of your visitors use. The default browser on both major phone platforms. One older version, on whatever your analytics show as the meaningful tail. And any browser your commercial clients are required to use, which for anyone selling into institutions can be an older one than they expect.

Test the things that must work rather than the whole site: the phone number, the menu, the form, and the checkout if you have one.

The specific cases worth knowing

Two situations where old browsers are more common than the averages suggest.

Institutional and corporate machines, where updates are controlled centrally and can lag substantially. A business selling to government, healthcare, or large employers should check this deliberately.

Older customers on older devices, particularly on tablets that stopped receiving updates. For trades serving an older demographic this is a real segment.

In both cases the answer is not full support, it is making sure the plain version genuinely works, which costs almost nothing when the site was built that way and a great deal when it was not, and it is the same underlying question as what the site is built on in what slowness actually costs you.

Frequently asked questions

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.