How I’d investigate a drop in website traffic

When website traffic drops, it is very easy to jump straight to the explanation.

Search is changing. AI is taking traffic. Google is sending fewer clicks. SEO is not working. We need more content.

Any of those things might be true. I just wouldn’t start there.

The first thing I’d want to know is what actually changed.

Start with the pages

An overall website traffic number tells you something happened. It doesn’t tell you much about what happened.

I’d look at which pages lost traffic and where that traffic used to come from. Did blog traffic decline while the service pages stayed steady? Did one historically high-volume article fall off? Did organic traffic drop across the site, or is the issue concentrated in a few pages? Did a campaign end? Did tracking change?

Those situations would lead me in completely different directions.

I’d also want to know whether the pages losing traffic were especially important to the business in the first place. A 30% drop on an old informational article may look dramatic in the dashboard and have very little effect on qualified demand. A much smaller decline on a high-intent service page might matter a lot more.

Then look at what is happening in search

Search behavior really is changing, so this is part of the analysis.

Pew Research Center looked at Google browsing behavior in 2025 and found that people were less likely to click a traditional search result when an AI summary appeared. That is useful context for anyone watching organic traffic right now.

It still doesn’t explain what happened to a specific website.

I’d go into Search Console and look at impressions, positions, clicks, and the actual queries that matter.

If impressions are relatively steady and clicks are falling, changes in the search results or user behavior may be part of the story. If rankings fell sharply, I’m looking at something different. If impressions also declined, I’d want to know whether demand changed, the content lost relevance, competitors improved, or the result page itself looks different now.

You need the detail before you decide what the trend means.

Then I’d look past traffic

The website exists to support the business, so I’d want to know what happened after the traffic changed.

None of these measures is perfect on its own. Together, they give you a better picture than “sessions are down 18%.”

I’d also be careful with conversion rate. It is entirely possible for conversion rate to improve while the business gets fewer opportunities.

If traffic falls from 10,000 visits to 5,000 and inquiries fall from 100 to 75, your conversion rate improved from 1% to 1.5%. That sounds good until you remember you have 25 fewer inquiries.

I want to see both the rate and the actual number.

The answer may be to fix something. It may also be to keep watching.

Once I understand what changed, then I can decide whether there is something worth doing.

Maybe an important article needs updating. Maybe a service page needs stronger proof. Maybe there is a technical SEO issue. Maybe a page that used to bring a lot of traffic no longer supports the business particularly well. Maybe the website is fine and the bigger issue is that leadership is still using traffic as the primary measure of marketing health in a search environment that is becoming harder to interpret.

And sometimes the change is small enough that I’d rather keep watching than invent a project because a dashboard moved.

Traffic still matters. I’m not ready to throw it away as a metric.

I just want a better question than “Why is traffic down?”

I’d rather ask which traffic changed, what we know about why it changed, what happened to the people we actually care about reaching, and whether there is anything useful to do about it.

That usually gets us somewhere.