"Where does engaged website traffic come from?"
At Megalytic, we’ve taken a crack at answering this question. We analyzed millions of visits to more than 1,000 websites to identify where, on average, the most engaged traffic comes from.
This is the first in a quarterly series where we will be investigating what really works when it comes to attracting quality traffic to a website.
When making decisions about where to invest time and money to promote their websites, marketers try to answer questions like:
- Which social networks send the most engaged traffic?
- What are the most effective acquisition channels (e.g., organic search, email, etc)?
- Which mobile devices drive engagement with website content?
We thought it would be enlightening to answer these questions, in general, by averaging over a collection of different websites. We measured the median engagement levels for website traffic along three dimensions: Social Networks, Acquisition Channels and Mobile Devices.
Based on these three metrics, we gave each social network, acquisition channel and device a composite score we call Mscore.
Yelp, TripAdvisor, and Google+ ranked highest among social networks.
Organic Search and Email topped the list of acquisition channels.
Apple iPhone and iPad deliver better engagement than many devices with larger screens.
We analyzed 12 of the leading social networks to determine which ones delivered the most engaged traffic to the sample websites.
Interestingly, Yelp came out as the Engagement Champion with an Mscore of 2.88; beating Google+, YouTube, Facebook, LinkedIn, Twitter and others.
|Rank||Social Network||Mscore||Bounce Rate||Pageviews/ Session||Session Duration|
Social networks vary widely in how well they deliver website engagement.
Points to Note
Perhaps Yelp and TripAdvisor did so well because these networks send traffic that is actively researching websites (e.g., restaurants, hotels). Reddit visitors, by contrast, are possibly just scanning for interesting content.
We also find it interesting that Google+ outscores Facebook, LinkedIn, and Twitter. A few Megalytic customers have been telling us, informally, that they feel they get their best social traffic from Google+. This data supports their hypothesis that Google+ is an effective referrer of engaged website traffic.
Next, we looked at the acquisition channels, as defined in Google Analytics. As with social networks, we compared engagement across a sample of websites that received significant traffic from all 7 channels.
Organic Search led the pack, followed closely by Email. Despite how much we all hate spam, the data show that email marketing effectively drives engaged traffic to websites.
|Rank||Referral||Mscore||Bounce Rate||Pageviews/ Session||Session Duration|
Email and Organic Search are the stickiest Acquisition Channels.
Points to Note
Social and Display perform noticeably worse than the top 4 channels. Perhaps this is because those channels contain a higher percentage of "impulse clicks" than the other channels. Maybe search traffic, for example, is more engaged because these visitors are looking for something specific after they land on a website.
You might be surprised by the performance of so-called "Direct" traffic. People tend to think of Direct traffic as being those visitors who typed in or bookmarked the URL of their website. If this were true, presumably these visitors would be highly engaged as they are intentionally headed directly to your site with some purpose. However, the truth is Direct traffic is rarely direct. Mostly, it has become a bucket for traffic Google Analytics can't identify. For a variety of reasons, traffic from apps, mobile browsers, and redirects from social networks often end up in the Direct bucket. The fact that it has become a "catch all" for unidentified traffic is consistent with its placement in the middle of the Mscore ranking.
Analyzing engagement by mobile device presents a few data quality challenges. First, the metric Average Session Duration does not provide a great measure of engagement across devices, because older, slower, devices with less network bandwidth take longer to load pages than newer devices, and this skews the results. For this reason, we computed the Mscore for Devices using only 2 metrics: Bounce Rate and Pages per Session.
Another challenge involves accurately identifying and separating out all the various devices based on the tags provided in the Google Analytics data. Some Apple devices (e.g., the iPad and iPad mini, the iPhone 3 and iPhone 4) cannot be distinguished from each other and we are forced to lump them together (see: A Pixel Identity Crisis).
Despite these challenges, we were able to pull together some interesting data.
|Rank||Device||Mscore||Screen Size (inches)||Bounce Rate||Pageviews/ Session||Session Duration|
|1||iPad (all models)||1.96||9.7/ 7.9||45.2%||3.06||2:56|
|2||Galaxy Note (10.1)||1.96||10.1||47.6%||3.18||3:09|
|3||Galaxy Tab 3 (10.1)||1.93||10.1||48.5%||3.15||3:35|
|4||Galaxy Tab 2 (10.1)||1.83||10.1||49.8%||2.89||3:19|
|5||Galaxy Note (8.0)||1.78||8.0||50.3%||2.79||2:35|
|6||Galaxy Tab 3 (7.0)||1.68||7.0||54.4%||2.68||2:29|
|7||Galaxy Note 3||1.64||5.7||54.3%||2.57||1:56|
|8||Moto G (XT1032/33)||1.61||4.5||54.2%||2.46||2:09|
|9||Galaxy Tab 2 (7.0)||1.61||7.0||55.9%||2.54||2:57|
|10||LG Nexus 5||1.59||4.95||55.2%||2.47||1:41|
|12||Galaxy Note II||1.58||5.5||54.5%||2.37||1:59|
|13||iPhone 5 Series||1.57||4.0||55.3%||2.41||1:38|
|14||HTC M7 One||1.56||4.7||56.2%||2.41||1:39|
|15||Galaxy S IV||1.55||5.0||56.6%||2.42||1:47|
|16||Galaxy S4 Mini||1.55||4.3||55.2%||2.34||1:55|
|17||iPhone 3/4 Series||1.53||3.5||56.8%||2.36||1:46|
|18||Galaxy S III||1.45||4.8||59.0%||2.22||1:51|
|19||Galaxy S III Mini||1.38||4.0||61.0%||2.13||2:01|
Some devices (like the iPhone 5) "punch above their weight" - delivering better engagement than peers with larger screens.
Points to Note
There is a strong correlation between screen size & how well a device delivers engagement with website content. However, it is not the only factor as you can see in the table above, several devices “punch above their weight” with respect to screen size. The iPhone 5, for example, has a 4.0 inch screen, but edges out the HTC M7 One (4.7 inch screen) and Galaxy S IV (5.0 inch screen).
Since mobile device screen sizes are growing, our data show that the opportunity for engaging mobile visitors is improving. Optimizing a website for mobile may be the best investment you can make for increasing engagement.
A few other points to take away from this data are:
- Your website's engagement numbers may differ from these. A lot may depend on the type of content - a variable beyond our scope here. We hope to look at content type in the near future.
- Our data show wide variations in engagement based on traffic source. Know the engagement numbers for your website. Focus marketing efforts on the social networks, channels, and devices where your website gets the most engagement.
- Lastly, there have been a few other studies of this type, although focussed exclusively on social networks. Here are a couple that stood out for us:
About Our Analysis
To perform this analysis, we started with an anonymous random sample of approximately 1,000 websites using the Megalytic reporting tool. We pulled the Google Analytics engagement metrics (Bounce Rate, Pages per Session, and Average Visit Duration) across three sets of dimensions: Social Network Traffic Source, Acquisition Channel Traffic Source, and Device Type.
We wanted to look at sites that receive a balanced variety of traffic. So, for each set of dimensions, we whittled down the sites in the sample, to ensure that most of the dimension values in that set occurred in the data for all the sites. For example, from the 1,000+ starter websites, we selected about 200 that received traffic from all the social networks in our study (e.g., Yelp, Google+, YouTube, etc). This whittling down process helped to prevent sites receiving a preponderance of traffic from only a few of the sources from skewing the results.
From the whittled down subset of sites, we calculated the median Bounce Rate, Average Session Duration and Average Pages per Session. The Mscore was then calculated as the weighted sum of these engagement metrics. This is not precision data science, for a variety of reasons. But, we believe it provides a reasonable guide to understanding where engaged traffic comes from.
What Do You Think?
Know something we don't? Think our numbers are wrong? We love feedback. Let us know below.
Let's Stay in Touch
This is the first study in a series. Stay in touch to learn more about what drives website success. Bookmark the blog, grab our RSS, join us on Twitter or Facebook. Also, signup for our Megalytic service to learn more about your data.
About the Author
Mark Hansen is the Founder and President of Megalytic, the leading tool for building web analytics marketing reports.
Megalytic is used by digital agencies, marketers and business owners for faster, more insightful and better looking analytics reports.
Mark loves math, programming, and marketing – a combination of interests which eventually led to web analytics. He holds a PhD in Computer Science from MIT.