Almost every consultant or in-house SEO will be asked at some point to investigate an organic traffic drop. I’ve investigated quite a few, so I thought I’d share some steps I’ve found helpful when doing so.
Is it just normal noise?
Before you sound the alarm and get lost down a rabbit hole, you should make sure that the drop you’re seeing is actually real. This involves answering two questions:
A.) Do you trust the data?
This might seem trivial, but at least a quarter of the traffic drops I’ve seen were simply due to data problems.
The best way to check on this is to sense-check other metrics that might be impacted by data problems. Does anything else look funky? If you have a data engineering team, are they aware of any data issues? Are you flat-out missing data for certain days or page types or devices, etc.? Thankfully, data problems will usually make themselves pretty obvious once you start turning over a few rocks.
One of the more common sources of data issues is simply missing data for a day.
B.) Is this just normal variance?
Metrics go up and down all the time for no discernible reason. One way to quantify this is to use your historical standard deviation for SEO traffic.
For example, you could plot your weekly SEO traffic for the past 12 months and calculate the standard deviation (using the STDEV function on Google Sheets or Excel makes this very easy) to figure out if a drop in weekly traffic is abnormal. You’d expect about 16% of weeks to be one standard deviation below your weekly average just by sheer luck. You could therefore set a one-standard-deviation threshold before investigating traffic drops, for example (but you should adjust this threshold to whatever is appropriate for your business). You can also look at the standard deviation for your year-over-year or week-over-week SEO traffic if that’s where you’re seeing the drop (i.e. plot your % change in YoY SEO traffic by week for the past 12 months and calculate the standard deviation).
SEO traffic is usually pretty noisy, especially on a short time frame like a week.
Let’s assume you’ve decided this is indeed a real traffic drop. Now what? I’d recommend trying to answer the eleven questions below, at least one of them will usually identify the culprit.
Questions to ask yourself when facing an organic traffic drop
1. Was there a recent Google algorithm update?
MozCast, Search Engine Land, and Moz’s algorithm history are all good resources here.
Expedia seems to have been penalized by a Penguin-related update.
If there was an algorithm update, do you have any reason to suspect you’d be impacted? It can sometimes be difficult to understand the exact nature of a Google update, but it’s worth tracking down any information you can to make sure your site isn’t at risk of being hit.
2. Is the drop specific to any segment?
One of the more useful practices whenever you’re looking at aggregated data (such as a site’s overall search traffic) is to segment the data until you find something interesting. In this case, we’d be looking for a segment that has dropped in traffic much more than any other. This is often the first step in tracking down the root cause of the issue. The two segments I’ve found most useful in diagnosing SEO traffic drops specifically:
- Device type (mobile vs. desktop vs. tablet)
- Page type (product pages vs. category pages vs. blog posts vs. homepage etc.)
But there will likely be plenty of other segments that might make sense to look at for your business (for example, product category).
3. Are you being penalized?
This is unlikely, but it’s also usually pretty quick to disprove. Look at Search Console for any messages related to penalties and search for your brand name on Google. If you’re not showing up, then you might be penalized.
Rap Genius (now Genius) was penalized for their link building tactics and didn’t show up for their own brand name on Google.
4. Did the drop coincide with a major site change?
This can take a thousand different forms (did you migrate a bunch of URLs, move to a different JavaScript framework, update all your title tags, remove your navigation menu, etc?). If this is the case, and you have a reasonable hypothesis for how this could impact SEO traffic, you might have found your culprit.
Hulu.com saw a pretty big drop in SEO traffic after changing their JavaScript framework.
5. Did you lose ranking share to a competitor?
There are a bunch of tools that can tell you if you’ve lost rankings to a competitor:
If you’ve lost rankings, it’s worth investigating the specific keywords that you’ve lost and figuring out if there’s a trend. Did your competitors launch a new page type? Did they add content to their pages? Do they have more internal links pointing to these pages than you do?
GetStat’s Share of Voice report lets you quickly see whether a competitor is usurping your rankings
It could also just be a new competitor that’s entered the scene.
6. Did it coincide with a rise in direct or dark traffic?
If so, make sure you haven’t changed how you’re classifying this traffic on your end. Otherwise, you might simply be re-classifying organic traffic as direct or dark traffic.
7. Has there been a change to the search engine results pages you care about?
You can either use Moz’s SERP features report, or manually look at the SERPs you care about to figure out if their design has materially changed. It’s possible that Google is now answering many of your relevant queries directly in search results, put an image carousel on them, added a local pack, etc. — all of which would likely decrease your organic search traffic.
Celebritynetworth.com lost most of its SEO traffic because of rich snippets like the one above.
8. Is the drop specific to branded or unbranded traffic?
If you have historical Search Console data, you can look at number of branded clicks vs. unbranded clicks over time. You could also look at this data through AdWords if you spend on paid search. Another simple proxy to branded traffic is homepage traffic (for most sites, the majority of homepage traffic will be branded). If the drop is specific to branded search then it’s probably a brand problem, not an SEO problem.
9. Did a bunch of pages drop out of the index?
Search Console’s Index Status Report will make it clear if you suddenly have way fewer URLs being indexed. If this is the case, you might be accidentally disallowing or noindexing URLs (through robots.txt, meta tags on the page, or HTTP headers).
Search Console’s Index Status Report is a quick way to make sure you’re not accidentally noindexing or disallowing large portions of your site.
10. Did your number of referring domains and/or links drop?
It’s possible that a large number of your backlinks have been removed or are no longer accessible for whatever reason.
Ahrefs can be a quick way to determine if you’ve lost backlinks and also offers very handy reports for your lost backlinks or referring domains that will allow you to identify why you might have lost these links.
A sudden drop in backlinks could be the reason you’re seeing a traffic drop.
11. Is SEM cannibalizing SEO traffic?
It’s possible that your paid search team has recently ramped up their spend and that this is eating into your SEO traffic. You should be able to check on this pretty quickly by plotting your SEM vs. SEO traffic. If it’s not obvious after doing this whether it’s a factor, then it can be worth pausing your SEM campaigns for specific landing pages and seeing if SEO traffic rebounds for those pages.
To be clear, some level of cannibalization between SEM and SEO is inevitable, but it’s still worth understanding how much of your traffic is being cannibalized and whether the incremental clicks your SEM campaigns are driving outweigh the loss in SEO traffic (in my experience they usually do outweigh the loss in SEO traffic, but still worth checking!).
If your SEM vs. SEO traffic graph looks similar to the (slightly extreme) one above, then SEM campaigns might be cannibalizing your SEO traffic.
That’s all I’ve got — hopefully at least one of these questions will lead you to the root cause of an organic search traffic drop. Are there any other questions that you’ve found particularly helpful for diagnosing traffic drops? Let me know in the comments.
