How much time do you spend looking at your Google Analytics data? My guess is, if you’re like the majority of business owners, not a lot. It might be you’re plotting the key audience stats each week or month, but nothing more. The reason we wrote this post is to ask everyone that’s doing this to take a step back and think about the data – and particularly, what might explain any spikes in traffic that may be occurring. By spikes we mean peaks.
We all love traffic spikes, it’s shows that people like our content right? Well maybe, but unfortunately all may not be as it seems. This post will help you ensure your spikes are due to surges in traffic and nothing else, and will help you understand what caused them. If we can understand what caused the spikes we can repeat them.
Google Analytics installation
Firstly we’re assuming you’re a website owner and that you use Google Analytics, the most popular analytics package, and have it installed and configured correctly on your website. Incidentally we run a Google Analytics workshop whereby we check your setup and offer some suggestions as to how to make the most out of your installation.
At the very least you should check that the Google Analytics code is present on every page of your site. We use a website crawler to check this, Screaming Frog as it happens, but there are plenty of other options online including GA Checker. If you’re not sure, just ask your webmaster.
Identifying traffic spikes in GA
Let’s say you’ve changed your date range to show traffic in the last 30 days. That was the default date range until a little earlier this year when they switched it to the last seven days. Spikes are hard to spot in seven days – broadening your date range should help you spot any.
Go to your Audience Overview report. You should notice on the graph a pattern that repeats itself each week. All websites are different, some will be busier during the week, others at weekends. Some sites will get more traffic during the day, others in the evening. What you’re looking for are unnatural peaks that look out of the ordinary.
Likely causes of traffic spikes
Ok, you’ve found an unnatural looking spike in the last 30 days of traffic. Before you get into the data it’s worth considering what the most likely cause could be. Firstly ask yourself what promotional activities were conducted in that period that might have had a positive impact on your traffic.
I’d strongly recommend that you get into the habit of annotating your Google Analytics reports to highlight any significant promotional activities or other actions that could have influenced how much traffic arrived on a particular day. You can do this by clicking on the small grey arrow in the middle of the x axis, as per the screenshot below. Doing so could save you lots of time when you’re analysing your data.
If you’d attended a trade show for example you’d expect your traffic to have increased on or just after the show dates. Or it might be you’d sent an e-newsletter to your list of subscribers with links back to your website.
I mentioned earlier that your traffic may be quite typical each week, perhaps it builds from Monday to peak on a Friday. There’s nothing worth investigating here if that’s pretty much the same each week. But I would suggest it’s important to check your traffic against the same period the previous year, provided you have the data available. This is really simple to do. Click on your date range and tick the ‘Compare to’ box and select Previous Year from the dropdown.
In December for example, it’s not unusual for some sites, particularly eCommerce sites, to experience very high traffic on Boxing Day and the day after, when many people have more time at their disposal and decide to search the sales.
If neither of these provide the answers you’re looking for then it’s time to dig deeper.
Looking for clues
You need to isolate the date on which the spike occurred. You can choose to show data for one date by clicking on the date range and then double clicking on the date you’re interested in, then hit Apply. You’ll then see how the traffic arrived at your website each hour. Then it’s just a case of looking for clues as to what might have caused the traffic peaks.
If all the traffic arrived in the same hour then that would set alarm bells ringing. If the top line stats suggest anything out of the ordinary, for example really high bounce rate or low average session duration, those alarm bells will be louder!
I’m going to go through a number of likely causes next and how to fix any issues with your Google Analytics setup that could be responsible. We’ll use the Acquisition Overview report as a starting point.
This report breaks down how your traffic for that date range arrived at your site, be it Direct, Organic, Referrals, Social or Paid. I’ll go through each in turn.
This is traffic that has no referrer. The visitor has arrived at your site most likely as a result of typing your url into the search bar or pulling up a bookmark.
Carlos Escalera wrote a useful post, which he updated recently, highlighting reasons for unusual direct traffic in your analytics reports. Unnatural spikes are most likely explained by bot-traffic. The warning signs alluded to a little earlier, for example high bounce rate, should confirm if this is the case or not. Then it’s a case of checking your Settings to ensure your account is setup for Bot Filtering, as it’s likely it isn’t.
You can correct this by heading to your Admin panel in Google Analytics, going to the section furthest right entitled ‘VIEW’ and clicking on ‘View Settings’, scroll down to ‘Bot filtering’ then check the box below it and hit Save.
That should solve the problem going forward but will not correct it retrospectively. I would recommend you annotate your reports to reflect this change in your configuration.
These are visitors that arrive at your site as a result of clicking on a link to content on your site in the organic, ‘unsponsored’, section of the search engine results pages, be it Google, Bing or another. This is what SEO is all about.
If the percentage of Organic Search traffic that arrived at your site on the date you’ve isolated is greater than it would be typically, you may have found an answer. It may be that your SEO efforts paid off on that date. It could be that you published a post on or just before the date which successfully targeted a specific keyword with a decent search volume. You could prove or disprove this theory by looking at your Behaviour Overview report, but finding out what happened isn’t always this straightforward unfortunately.
You’ll find, if you haven’t already, that you’re not going to get a lot of information if you drill down into the Acquisition Reports. The majority of the data is guarded by Google with the keyword(s) responsible for any traffic being categorised as ‘not provided’.
However all is not lost. You’ll need to head over to your Google Search Console. If your website is not set up to be tracked here then you should make that an absolute priority. The Search Console offers fantastic insight into how your website is being used, and via the Search Analytics report itself, some fascinating data about what keywords visitors used to arrive at your site as well as how prominent your content is on the SERPS for said keywords.
It’s highly likely that if you run a custom date range for the date in question – when the spike was experienced – the date returned will highlight an explanation.
This is traffic that arrives at your site as a result of clicking on a link in content on another website.
Your referral reports can provide really useful insight into which other sites mention you, positively or otherwise, on theirs. If you don’t check this report then you may be surprised by what you find. Referrals from other sites to yours are sent in the main as a positive endorsement of your content and as such these are extremely important in establishing your link profile.
However when you drill down into your reports you may be slightly underwhelmed if you’re not filtering out spam bots on a regular basis. In the same way as with direct traffic, bots can skew your data so you’re not seeing the complete picture. If you don’t recognise the sites liking to you and they look ‘spammy’, it’s highly likely these are spam bots. Excluding them is easy enough – but as with the other bots, they cannot be excluded from your historical data.
It’s really important to go through this process as they’re not only skewing your Acquisition Reports, but your overall Audience Reports too.
Here’s how you do it. You’ll need Administrator access to your account.
Go to your Admin panel. Click on ‘All filters’ in your Account folder. Click the red ‘Add Filter’ button. If you can’t see this button you’ve not got the right access to your account.
Give the filter a name ‘Spambots’ for example – I tend to add a last checked date to the name too so I know when I last carried out the process. So ‘Spambots 2 May 2017’.
Then configure the settings as per the image and step-by-steps below.
- Filter Type = Custom
- Filter Field = Exclude, Campaign Source
- File Pattern = Enter domain for the spam bots you’re looking to filter out, separated by pipes with no spaces e.g. buketeg.xyzbuketeg.xyz
- Ensure all the Views you want to apply the filter to are selected
- Hit the blue Save button.
You’ll need to set a reminder to add any new spambots on a regular basis.
And congratulations, you’ll have taken control of your Referrals report. You should take time to check the referrals you do receive and acknowledge them as and when you think necessary.
The traffic that arrived at your site as a result of the visitors clicking on a link to your content you’ve shared via your social media channels.
If the report highlights an abnormally large percentage of your traffic arrived ‘socially’, it’s looking likely that your content has been engaged with and/or shared more than usual on social media. If you didn’t notice it at the time on your channels, you might not find it too difficult to look back and pinpoint what happened. Or you’ll need to invest in social media monitoring software to establish what happened.
Getting a full understanding of what content ‘went viral’ and how, is of course hugely valuable so do make sure you invest some time and resource into this. And attempt to repeat whatever worked often as possible. As with your referrals, make sure you stop by and thank whoever helped amplify your content.
These are the visitors that arrived at your site as a result of clicking on a link to content on your site in the sponsored section of the search engine results pages. Also known as paid search or pay-per-click (PPC).
I’m not going to spend too much time talking about this here as this post is about Search Engine Optimisation. If you are running paid search campaigns I’d strongly recommend you use a url builder to ensure the pages you’re promoting via your campaigns, are unique and as such will be highlighted as such in your Google Analytics reports. It’ll then be much easier to understand if your campaigns had a positive impact on your traffic at any given time.
Hopefully this post will have given you some valuable insight into what you should and shouldn’t believe when you look at your Analytics data, particularly the spikes. To sum up, please don’t take what you see at face value, you’re going to need to delve a little deeper to fully appreciate what’s going on.
We welcome your thoughts on all this below.
Main image by Unsplash