Tag Archives: website analytics

Website Analytics: Turn Flabby Numbers Into Customers

The hungry marketer Many years ago, a client called me at 7:30 a.m. “Your f…..g ads don’t work, you f……g a……..e.” In retrospect this is hilarious because his store wasn’t even open yet. Some people make the same snap judgments about websites based on top line data only. This information (see below) may flag a problem on the site, but it’s not the diagnosis. Now what? Say your car doesn’t start. The problem could be everything from a faulty ignition switch to a potato lodged in your tail pipe. We need to get under the hood and get dirty.

Typical dispaly of top line website data
These are topline numbers on your Google Analytics (GA) account’s home page. If you don’t drill down to understand what’s really going on, they can bite you in your rear part.

Top line numbers fluctuate due to many factors beyond your control, seasonality, popular soccer match, day of the week, or job postings on your website. The numbers go up, and they come down. And they will not tell you important dimensions about the size of your core audience, or the surge in traffic driven by social media.

So slap on a pair of coveralls, grab a wrench, and let’s do a deep dive on your website until we reach the core. As for the hard charging, results driven, monkey on your back, you’ll have a few more bananas to pitch at him.

Can you tell where the 300% increase in traffic dropped off?
Can you tell where the 300% increase in traffic dropped off?

Last October, Ideopia redesigned a website. Traffic growth was almost immediate. Eventually the growth slowed, but growth in visits continued to increase. Then boom! Traffic spiked over 300%. Champagne corks popped like gunfire at Ideopia. And yes, we were quick to point out this accomplishment to our client.

We poked underneath the traffic to find the mystery traffic.  And the bump wasn’t do to earned traffic at all.  By checking traffic sources, we found the culprit, an Adwords account that was running amok. While it’s nice to take the credit, our job here was to shut down the rogue Adwords account. And we did.

One reason website analytics get wonky is because they’re based on averages. For example: web traffic might go up, but pages per visit go down. Has your content broken down? That’s possible, but it also might be a slow download speed for your site.

Other and possibly deceitful metrics – time per visit, and pages per visit – are based on averages of all your site’s data, too. Now hear this:  There is no average visitor. One visitor might hit one page on your site. Another may visit a hundred making the average number of pages per visit 500. Again, we love to report the good news. But it’s not helpful for decision making. So lets take an example using Pareto’s 80/20 distribution. For the sake of this example, our website pulls in a whopping 100 visitors per month.

Traffic Segment A 

  • Web Traffic = 20
  • Average page views per visitor = 10
  • Total Pageviews = 200
  • Time per visit = 5 minutes

Example Segment B

  • Total  traffic = 80
  • Average page views per visitor = 1
  • Total Pageviews = 80
  • Average time per visit = 1 minute

Averaging A & B segments together yields the top line data you would see on your dashboard.

  • Total traffic = 100
  • Total  page views= 280
  • Average page views = 190
  • Average time per visit = 3 minutes

The point here is to show how averages misrepresent the values in both A & B segments. Knowing that we have 2.8 page views is useless. What’s interesting is that 20% of our traffic views 10 pages, and 80% account for 80 page views. See how unhelpful averages are? Depending on your objectives, you might be more concerned with core visitors, or overall traffic. If only you could determine what’s sticky for the core group, and what’s turning off the fringe visitors.

Well, you can with custom segments from Google Analytics. It’s a power tool for defining your key audience segments, and a component of decision making about content, user experience, and SEO.  The answers aren’t on top. Drill down, and start finding customers.

Find Customers in the Core

Use Google’s custom segments to define your target audience. Describe it by demographics (18-24), keywords from search, affinity categories like sports and gardening, and traffic sources, like social media or Adwords. The core is the group of visitors you want to tantalize with your content, capture their email addresses, and eventually contact them personally. This report shows engagement by comparing numbers of users in different time segments.

Rules of Engagement

A metric like engagement should make you swoon. On Google Analytics, it’s called “Engagement,” located under Behavior on the left-hand navigation. Let’s take the table below as an example, and say that we consider anyone who has made a visit of 180 seconds, or 3 minutes, to define a core visitor.

Which would you say are the most engaged visitors?
Which would you say are the most engaged visitors?

380 or 21% of visitors accounted for 6,116 or 56%, of page views lasting more than 3 minutes. According to our definition, this is the core audience. To refine further, choose from a slew of pre-configured reports to import into your accounts. Custom segment in hand, you’ll want to apply it to other reports in GA. Find out how much of your core audience visits from social media, and what platform. What pages are most appealing to this group? Who is performing what actions on your site.

Increasing Website ROI

While it’s fun to gloat over top line numbers, it can lead to rash decisions based on them, like trashing your existing site, revamping the home page, or stuffing copy for SEO. Keep in mind that your site still needs  to cater to the other 80%. They will become the new core, and fodder for content marketing programs. What are your top analytic tricks? Please add them to the comments.

Bill Abramovitz is CEO and Creative Director at Ideopia, a Massively Integrated Idea Company.

Comments Off on Website Analytics: Turn Flabby Numbers Into Customers Back to Top

How your Awesome Looking Web Site is Tricking You

When it comes to web design and development, what you see is most definitely not what you get.

Before writing any code, the client reviews design layouts to show what the website will look like. Unfortunately people tend to treat these the same way they treat print proofs. They think once they sign off on the web design comps, that is EXACTLY what the website will look like. The problem is that not everyone is looking at your website with the same operating system, web browser, screen size, or pixel density.

All browsers are not created equal.
Although browsers should conform to W3C standards, they are not required to. Complicating matters further, the W3C standards are always evolving and expanding. But the browser companies decide when and how they implement these changes. Each browser interprets the same set of code slightly differently. For instance, Internet Explorer will optimize it for Windows while Mobile Safari will optimize it for the iPhone. Sometimes these different interpretations are barely noticeable and other times they are very dramatic.

Complicating matters even further, there are always multiple versions of a browser being used at any given time. For instance, Internet Explorer 8, 9, and 10 are currently being used. Each version displays the same web page very differently.

Test. Test. And test some more.
It’s crucial to test your website in multiple browsers and versions of each browser. It can be a daunting task, but with a good strategy and some technology, any pain is totally worth it.


Step 1) Know yourself.

Empty your mind, be formless. Shapeless, like water. If you put water into a cup, it becomes the cup. You put water into a bottle and it becomes the bottle. You put it in a teapot, it becomes the teapot. Now, water can flow or it can crash. Be water, my friend.”
― Bruce Lee

The first step is to change your perspective. Separate design and content. What happens to your precisely designed desktop website when it’s displayed on the small screen of a mobile phone? Does the content reflow to use the small space? What happens if the user has vision problems and increases the font size to 30pt? The design must be flexible. You can either create a website that flows like water to take on the shape and size available or you can demand a rigid, pixel perfect design that will “crash.” Don’t waste time making sure that a line break hits properly in all 20+ browser iterations. It might work for a while…until a new browser version comes along. But think of the time and money you wasted and what you could have invested in better marketing or content. This doesn’t mean compromising your design. But you will need to stop treating website design like print design. It’s a different medium with different challenges and requirements.

Step 2) Know your audience. Look at your current website analytics and see what browsers your customers are using. How much traffic comes from mobile devices? How many customers are using Internet Explorer?

You should also view global usage breakdowns per browser. It’s important to know the most popular browsers and what version of those browsers are used most. Compare those stats to your web analytics.

We’ve found that global trends don’t always match up to vertical market usage. For instance, Internet Explorer 8 usage is 5.5% globally. But you may find that in your vertical market that Internet Explorer 8 usage is over 30%. Armed with this information, you’ll be able to easily determine how much effort you need to put into dealing with browser quirks.

Step 3) Know thy enemy. Every browser has its quirks. As stated earlier, every browser translates and displays the same code differently. Know the differences or find someone who does.

Step 4) Know the tools of the trade. There are a number of tools (free and pay) available for testing your website on multiple operating systems and browsers. Some will provide a screenshot of your page in each browser.

Quirktools
Browser Shots
Browserling

Others diagnostic sites test your site live. You log in and select the operating system and the browser version to test the page. For instance, you could select Windows 7, Internet Explorer 9. This type of solution is great for testing customer processes like a shopping cart checkout.

Browsercam
Sauce Labs
Adobe Edge Inspect

It’s essential that your website looks great for all your customers, and with the amount of fragmentation in the browser market, browser testing is a must. It can be difficult and time consuming for the novice. If testing is not built into your development budget, it should be. The ROI is very high if you have an experienced developer.

See more of our web design posts on WavyBrainy; and leave comments below!

 

Comments Off on How your Awesome Looking Web Site is Tricking You Back to Top