Information architecture

Information architecture is how the content of a website is organised, named and connected. It decides what the sections are, what each one is called, which page a given piece of information belongs on, and how somebody moves between them. It is the part of a website nobody comments on when it is right and everybody feels when it is wrong, usually as a vague sense that the site is hard to use without being able to say why.

In short

• It is how pages are grouped and linked, not how they look.
• Every page should have one job that a visitor can state in a sentence.
• Labels work when they use the words customers already use.
• Structure decides what search engines can find and how important they judge it.

The four parts of information architecture

Information architecture is easier to work on when it is split into the four parts it is usually broken into.

• Organisation. How pages are grouped, and what sits under what. Services in one group, service areas in another, the company information somewhere it does not compete with either.
• Labelling. What each group and page is called, in the menu, in the heading and in the address.
• Navigation. How somebody gets from where they are to where they want to be, including the main menu, the footer, links in the body text and breadcrumbs.
• Search. How somebody finds a specific thing when browsing is too slow, which matters on a site with a lot of content and not at all on a site with twelve pages.

Most small business sites are weak on the second and third of these. The pages exist and are grouped sensibly enough, and then they are given names only the owner understands and linked to from one place.

Why the structure decides whether anything is found

A visitor arriving from Google rarely lands on your front page. They land in the middle, on whichever page matched what they typed, and their next move depends entirely on what that page offers them.

If the page answers the question and points at the obvious next step, they continue. If it answers the question and ends, they leave, and you never learn that they wanted three more things you also sell. That second outcome is not a copywriting failure. It is a structural one, and it repeats on every page that was written as an island.

The same is true in reverse. A page nobody links to is a page that is hard to find and easy for a search engine to conclude is unimportant, because importance is inferred partly from how the rest of the site treats it.

One page, one job

Every page should do one thing, and you should be able to say what that thing is out loud.

A page covering five services answers nobody’s specific question well. Somebody searching for aircon cleaning in Quezon City wants a page about aircon cleaning in Quezon City, and a search engine looking for the best match for that phrase wants the same. A combined page is a compromise nobody asked for.

Splitting them does two things at once. Each page gets to be genuinely about its subject, with the detail, the photographs and the questions that only apply to that service. And each becomes a target somebody else on the site can link to, which is how the structure starts holding itself together.

This is where information architecture and on-page SEO meet. On-page work makes a page the best answer to one search. Architecture is the decision about which searches get a page at all.

Naming things the way customers name them

Labels fail in a predictable way. The business names a section after its internal department, its product line, or a phrase it finds clever, and visitors do not recognise any of them.

• Industry words. Nobody outside the trade searches for facilities management, and plenty of people search for what it actually covers.
• Invented names. A section called The Hive is memorable to the person who named it and meaningless to everyone else.
• Vague words. Solutions, Resources and Products tell a visitor nothing about which one to click.
• Company structure. Grouping by department describes your organisation chart, not your customer’s problem.

The test is simple. Write your menu labels down, show them to somebody who does not work with you, and ask what they would expect behind each. Where they hesitate, so does everybody else.

What structure does for search

Search engines have to crawl a site before they can rank anything on it, and they follow links to do it. The structure is the map they use.

• Depth matters. Pages buried many clicks deep are crawled less often.
• Orphan pages read as unimportant. A sitemap will get a page found. A page that nothing on the site links to still reads as one even you do not rate.
• Internal links pass weight. The pages you link to most from elsewhere on the site are the ones you are telling search engines to take seriously.
• Overlapping pages compete. Two pages aimed at the same search split the signals between them and neither ranks as well as one would.

The mechanical side of this, crawling and indexing, belongs to technical SEO. The decisions about what should exist and how it should connect belong here.

Navigation on a phone

Most Filipino visitors will see your site on a phone, often on mobile data, and that changes what a structure can get away with.

There is no room for a wide menu, so the whole thing collapses behind an icon and becomes a list. A structure with four clear groups survives that. A structure with eleven top level items and three levels of nesting becomes a scrolling wall that people close.

Two things help more than any menu design. Keep the number of top level choices small enough to read at a glance. And put real links inside the content, so somebody reading about one service can reach the related one without going back to the menu at all.

How to test whether yours works

You do not need research software to find the obvious faults.

1. Say each page’s job out loud. If it takes more than a sentence, the page is doing two things.
2. Count the clicks to your most important page from the front page. More than three is a signal.
3. Ask somebody outside the business to find something specific and watch without helping. The pause before they click is the finding.
4. List every page and check what links to it. Anything with no links in is orphaned.
5. Search your own site on Google with the site: operator and look for two pages competing for the same thing.

If several of these come back badly, the answer is usually to fix the structure before touching the design. A new website built on the same grouping will look better and behave exactly the same.