Organic traffic falls and the immediate instinct is to assume an algorithm update. Most of the time it is something duller and more fixable — a deploy, a directive, a tracking change, or a SERP that quietly stopped sending clicks.
Work through the causes in order of likelihood rather than in order of drama. The sequence below is the same one we run in a traffic drop audit, and you can get a long way through it yourself with Search Console and an afternoon.
Step 1: Confirm you actually lost rankings
Before anything else, establish what kind of loss this is. In Search Console, compare clicks, impressions and average position over the drop window.
- Impressions down, position stable — you did not lose rankings. Demand fell, or a SERP feature is absorbing the query. Check seasonality against the same period last year first.
- Impressions stable, clicks down — you still rank, but something above you is taking the click. AI Overviews, a new ad block, or a competitor with a richer-looking result. If you suspect an answer engine is absorbing the click, this is how to check rather than assume.
- Impressions and position both down — a genuine ranking loss. Continue to step two.
- Everything down at once, overnight, to near zero — this is almost always tracking or access, not SEO. Check that your analytics tag is still firing and that the property is still verified.
That fourth case catches more people than anyone admits. A developer removes a tag during a release and the resulting chart looks exactly like a catastrophic penalty.
Step 2: Date the drop precisely
Get the exact day the decline starts. A drop that begins on a single day has a single cause. A slow slide over eight weeks is decay, competition, or content ageing — a different problem with a different fix.
Once you have the date, check it against two calendars:
- Confirmed Google updates. If your loss starts within a day or two of a core update, that is your leading hypothesis. Before assuming that, check Search Console for a manual action notice — the two are different problems with different remedies, and treating one as the other is the most expensive mistake available here.
- Your own deploy log. If it starts on a release date, the update is a coincidence and your own change is the cause.
The deploy calendar is the one people forget to check, and in our experience it explains more drops than algorithm updates do.
Step 3: Check indexation before anything else
If pages left the index, nothing else matters. In Search Console’s Pages report, compare indexed page counts before and after the drop date, and read the exclusion reasons.
The usual culprits:
noindexshipped from staging to productionrobots.txtblocking a directory after a restructure- Canonical tags pointing at the wrong URL, consolidating a hundred pages onto one
- A CDN or firewall rule newly blocking Googlebot
- Server errors during a crawl spike
This is the most common cause we find and the fastest to reverse. Confirming or eliminating it takes ten minutes, so do it before you commission a content strategy. The technical SEO checklist covers each of these checks in more depth.
Step 4: Separate sitewide loss from page-level loss
Export your top pages for the periods before and after. Then ask: did everything fall proportionally, or did a handful of URLs collapse while the rest held?
Sitewide, proportional loss points at something global — an algorithmic reassessment of the domain, a technical directive, or a link profile change.
A few pages falling off a cliff points at something local — those specific pages were hit by a competitor, cannibalised by your own new content, or affected by a change to that one template. Page-level loss is usually cheaper to fix, and often comes down to internal linking or an orphaned page problem.
Step 5: Check what changed off-site
If the first four steps come back clean, look outward:
- Lost backlinks. Compare referring domains against your pre-drop profile. A single expired link from a major domain can move a competitive term.
- New competition. Search your top three lost queries manually. If a well-resourced new page is sitting above you, that is your answer and the fix is content, not diagnostics.
- SERP layout change. Compare the current SERP against what it looked like before. New features push organic results down the page even when your position number has not moved at all.
Step 6: Consider that it might be nothing
Roughly one enquiry in five turns out to be normal fluctuation. Small sites bounce around constantly, and a 15% swing on a low-traffic site is noise rather than signal.
Before spending money on a fix, confirm that the drop is outside your normal variance and that it has persisted for at least three weeks. Reacting to a two-week dip usually causes more damage than the dip did.
When to bring someone in
Do it yourself if the cause turns out to be indexation, a deploy, or seasonality — all three are findable with the steps above and fixable by your own developer.
Get help when the loss is large, when it followed a site migration, when it followed work by a previous agency, or when you have worked through this list and still cannot date or explain it. At that point the cost of guessing exceeds the cost of a diagnosis.
If step two pointed at a manual action or a core update specifically, that is a different engagement — penalty and update recovery rather than a general traffic drop audit, because the remediation work barely overlaps.
