Check whether impressions fell alongside clicks. If both dropped, something changed in how Google sees your site. If impressions held and clicks fell, you still rank — you're just no longer being chosen. Then check your own change log before Google's. Most drops blamed on algorithms trace back to an internal change nobody recorded.
Start with the shape of the drop, not the cause
Every traffic decline looks the same in a dashboard: a line going down. But there are only a few shapes it can take, and the shape narrows the cause faster than any amount of speculation about updates.
Open Google Search Console, go to Performance, and do one thing before anything else: turn on both Total clicks and Total impressions, then compare the last 28 days against the same 28 days last year rather than the previous month. Month-over-month comparison catches seasonality and reports it as a crisis. Year-over-year doesn't.
Now read the two lines against each other.
Impressions fell and clicks fell together. Google is showing you less often. Something changed in how it sees your site — indexation, a technical problem, a genuine ranking loss, or a competitor who improved. This is the case where an algorithm update is a plausible suspect.
Impressions held and clicks fell. You are being shown exactly as often as before and chosen less. Your ranking didn't move. This is displacement: something is now answering the question before the reader reaches your listing. An AI Overview, a featured snippet, a competitor with a more compelling title.
Both rose but conversions fell. Not a search problem at all. Look at the page, the form, or what changed in your offer.
Most firms skip this step and go straight to reading about the most recent update. That's backwards — the shape of the drop tells you whether an update is even worth investigating.
Check your own change log before you check Google's
The most common cause of an unexplained traffic drop is not an algorithm. It's something the firm did to itself and didn't write down.
Ask, in order:
- Did anyone edit page titles or meta descriptions? Titles are the single biggest lever on click-through rate at unchanged rankings.
- Did a plugin update, a theme change, or a caching change alter how pages render?
- Did anyone add a
noindexwhile staging something, and not remove it? - Did the site move hosts, change URLs, or add redirects?
- Did an SSL certificate lapse, even briefly?
- Did a page that used to rank get "improved" — rewritten, shortened, or restructured?
That last one catches people constantly. A page ranking well is a page Google has decided is a good answer. Rewriting it is a bet that you understand the query better than the accumulated evidence does.
If you're on WordPress, your revision history and plugin update log give you most of this in ten minutes. If you have Search Console linked to Google Analytics, annotate the date the drop started and work outward from there.
The seasonality trap
Some declines are simply the calendar. A CPA's traffic falls off a cliff in May. An estate attorney sees a summer lull. A real estate team's search volume tracks inventory.
The test is not whether traffic fell — it's whether the floor is lower than last year's floor. Chart the same weeks across three years. Normal seasonality returns to roughly the same baseline. Erosion doesn't: each year's low sits below the last.
If your floor is dropping year over year, you have a real problem wearing a seasonal costume.
What you can and can't conclude
Directional — You can prove clicks fell against stable impressions. That's real, measurable, and yours to see.
What you can't tell from Search Console alone is why the click went elsewhere. An AI Overview absorbing the answer, a competitor's better title, a SERP layout change, and your own untracked title edit all produce an identical line on that chart. Distinguishing them requires SERP-level observation over time and, for the AI-answer case, direct prompting of the engines — neither of which is visible in the dashboard.
Unconfirmed — If you suspect an algorithm update, be honest that you're inferring. Confirming an update affected your site specifically means correlating your date with the confirmed rollout window, checking whether the drop is site-wide or concentrated in one content type, and ruling out the internal causes above. Most published "this update hit X industry" analysis is aggregate data that may not describe your site at all.
What to do this week
- Pull the impressions-versus-clicks comparison, year over year. Write down which shape you have.
- Build the change log you don't have. Every edit, plugin update, and structural change for the eight weeks before the drop.
- Chart three years of the same weeks to rule seasonality in or out.
- Only then read about updates.
That order matters. Reading about updates first gives you a story, and a story will make the data fit it.
If you've run all four and the shape is clear but the cause still isn't, that's the normal stopping point — not a failure. Confirming causes usually requires server logs and cross-engine testing you can't reach from the outside. The full symptom map is on how do I tell if my website is losing me clients.