Where to start when analysing a website's SEO

When someone tells me "I need SEO", they almost never need forty new articles. They need to know what they have, what Google sees and what's broken. This is the order I work in.

1. What Google sees

Connect Search Console. It's free and it's the only source that tells you which pages are indexed, which were dropped and why. Without it you're blind.

Look at two things: how many pages are indexed versus how many should be, and which queries already bring you visits. The second usually surprises people: they get found for things they never considered.

2. What your clients actually search for

Before chasing high-volume keywords, look at the ones you already have. Search Console gives you real queries with position and clicks. Three seams of gold there:

• Queries where you sit at position 5-15: that's margin for little effort.
• Queries with lots of impressions and few clicks: the title and description aren't convincing.
• Queries you didn't expect: they tell you how people really look for you.

3. Which pages work and which are surplus

List every page and check how many visits each brings. It's nearly always the same: a handful bring almost everything, and there's a long tail contributing nothing. That list tells you what to improve, what to merge and what to delete.

4. The technical part that does matter

• Fast loading and readable on a phone. It's a ranking factor and it decides whether people stay.
• A unique title and description per page, written for people.
• One h1 and headings you can follow.
• Internal links. The most underused lever: link from what already works to what you want to push.
• No duplicate pages. Canonical set properly and one version only (www or not, trailing slash or not).
• If you've migrated, check old URLs redirect to the matching new page, not the homepage.

5. Who you actually compete with

Google your three most important queries and see who appears. That's not your business competition: it's your competition in the result. Open those pages and see what they answer that you don't.

And then

With that you have a real plan: fix what's broken, improve what's close to ranking, and write only what's missing. It's far less glamorous than a twelve-month content strategy, and it works sooner.