11

Web Analytics: You’re Doing it Wrong

Most stats we see when we look at our website analytics are just “vanity stats”. They elicit an emotional response and we can even get addicted to logging in to check our stats, and seeing another slight improvement in traffic or some other positive signal.

But beyond that, how much use do we actually get from checking our stats? How often is it a real, useful business tool and how often is it just another distraction?

I was using analytics the wrong way for years and you’ve probably done the same. In this post, let’s kick our bad habits and look at how to use analytics to actually improve your online business.

Watch the Video Below:

 

Links: learn about a critical limitation in the way Google Analytics measures bounce rates and check out this review of Google Analytics alternatives.

The Top Down Approach to Analytics

The right way to approach web analytics is to start with a result in mind: what is it that you want to improve on your website? What’s your main coversion goal? What are your secondary goals? Depending on your business model, the answers can vary. They might include:

  • Make more sales.
  • Get more leads/opt-ins.
  • Get more clicks on affiliate links.
  • Get more ad impressions (page loads).
  • Increase user engagement.
  • Get more visitor participation (user generated content).
  • Get a better idea of what people in your niche want and need.
  • Improve social sharing.
  • etc…

Once you know what your goals are, you need to ask yourself what kind of data you need, to be able to make changes to your website intelligently, in order to reach those goals. Think of it as performance-based analytics.

Finally, figure out which tools you need, that will provide the necessary data and help you make the right improvements to your site. The answer simply isn’t always (or even often) going to be “Google Analytics”. In general, you’ll probably need a good analytics solution as well as a split testing tool such as Google Website Optimizer (basic and free) or Visual Webite Optimizer (awesome).

My question for you is: what are the main goals for your website? What do you need to improve? What kind of data and tools would the perfect analytics solution provide for you? Leave a comment and let me know!

Shane's Signature

0
Shares
Shane
 

I'm Shane Melaugh and I'm the guy writing most of the posts on this blog. My goal is to provide you with useful, straight-forward insights on how to grow your business by creating compelling offers, driving traffic and increasing conversions.

Click Here to Leave a Comment Below 11 comments
Paul @ IMH Blog - April 15, 2012

Optimizely looks like a really good tool for split testing, i tested out the service for them and was amazed at the things you could do, it was so easy to setup a split a/b test. You can change pretty much any element on the webpage with a simple click.

I have a health site that i would like to improve sign ups for , i am going to test a sign up box in the sidebar and also as a full width banner just below the header and also one at the bottom of each post like you have on imimpact.

Have you split tested this blog Shane and found some interesting results yourself?

Reply
    Shane - April 15, 2012

    Optimizely is very similar to Visual Website Optimizer. I honestly don’t know if one or the other is better, but they both offer very easy page editing for your split tests.

    I’ve done some tests with the videos on this blog, but haven’t found any really exciting results. I’m going to be doing some more testing on the homepage soon and I’ll be posting any interesting results I get.
    Having said that, I mostly use split testing on sales-pages.

    Reply
      Paul @ IMH Blog - April 15, 2012

      i see maybe ill do a comparison of the two, see what is better

      yes i do not actually have any sales pages most of my websites just offer information. Perhaps when i have a sales page or something to sell something like this will be more beneficial

      It would still be interesting to see how certain ad positions and sign up forms perform, so ill be sure to use it in the near future

      Reply
Tom - April 15, 2012

I have Google analytics set up on my two adsense sites. I stopped checking analytics months ago. I would just check my adsense account and see how much money and made for the day. As I am now making no money I may check my adsense account once a week. I have made $33 in the last three months.

But I am now on Facebook! I am about to try “a quick cash method”. I know from the webinar and your previous postings and rantings how you hate that idea but we will see.

Reply
Rusty Ferguson - April 17, 2012

I’m having a little trouble analyzing one of my MAJOR keywords. With Clicky, I can see that my true bounce rate is 50%. The keyword gets me a lot of traffic. I’m ranked number two and no matter how hard I try, I can’t knock Wikipedia off number one. They have the keyword in their title. I did try.

My overall bounce rate is 25%. Since I live in a country where many people don’t speak English well, that would probably explain why those visitors would bounce more often.

The blog is about that country. The stats for visit within the country recently took a huge nose dive. And that’s really okay because I’m going to sell them my eBook about living in their country. HAHA

According to SEOmoz, I have 10,000 pages. There is close to 1000 post and pages.

I just can’t see me removing that content. I’m going to try to find a way to keep people on the page longer. Could be hard. i need to see if Clicky will let me get that bounce rate via country. With Statcounter I could see most of the traffic is coming from the USA now.

The page has been ranked number 2 for a long time, over a year. So it seems Google is happy with it. So far that’s the worst preforming page.

What is a bad bounce rate. haha 25% seems wonderful to me because I’ve been looking at a bounce rate of nearly 80% with other analytics.

If you had a page ranking well that seomoz gives a ranking difficulty of 52% but had a bounce rate of 50% would you pull it?

This question might make a good blog post for you. How do you decide is a bad page and what do you do about it? What are things we could do to improve engagement? Polls, picture gallery? I have a youtube video on it. I think I’ll try to make my own and add that too it. Any other suggestions?

You know, one day, I’ll write a short comment for you Shane. at least my engagement must be very high for your site. :)

Reply
    Shane - April 17, 2012

    No, don’t pull a high-bounce page that’s getting a lot of traffic! Improve it.
    The first thing to do is to change the copy at the top of the page. Change the headline and the leading paragraphs to see if you can find wording that grabs people’s attention better.

    Also think about whether you might be offering something else than what people are expecting, on this page.

    And to answer the question about Clicky: go to “Content” and select the page you want to analyze. Then, click on the “Filter” item and select “Country” or whatever else you want to filter the data by.

    Reply
      Rusty Ferguson - April 19, 2012

      I did fine that in Clicky, I need more stats. What I did find is that my traffic for the Philippines is way down. But I couldn’t tell how down it was for the USA. Before more than 50% was from the Philippines and now its less.

      I was surprised to find though that the bounce rate for Filipino is much lower than that of the USA from search.

      I’ve got a Youtube video in there but I will defiantly look at the first part of the article.

      I’m not too worried about talking about the niche because if I knew four years ago what I know today, I might not have ever gone into the niche. lol

      Low traffic and high competition. Bad combination but I’m at the top of the SERPs now so its working. Its been a good learning ground for me.

      Just wish Google had left my affiliate review sites alone. Looks like they are slapping all my sites like that now and one is massive or on the way to be massive. I might just have to remain it, but on other hosting and start over. I’ll give it some time.

      You’re a fountain of information and you give me more ideas to blog about than I have time to write about. :)

      Thanks!

      Reply
    Paul - April 17, 2012

    Ditto what Shane said – I wouldn’t pull a page that’s getting loads of traffic.

    To reduce the bounce rate consider content calibration! There’ll be a post about this really soon ;)

    Outranking a wiki page can be a tough one – the Google brand bias is very strong at the moment.

    Reply
Rusty Ferguson - April 20, 2012

I beginning to think GA is a total waste. I pulled most of my accounts out two days ago.

The only reason I don’t pull the rest is because I’m an SEOmoz subscriber and it is nice to see the traffic numbers in my account. But i wonder how accurate those are.

I looked at page load speed times and saw 25 seconds to load my home page on average. They appear to be getting those numbers from people as they visit. Today I look and see some pages at 450 seconds. Give me a break. They take about 5 seconds to load on my slow Philippines connection. it gets worse if I don’t use my US based VPN though.

Google is starting to feel like a foe to do battle with rather than a partner to work with. Bad stats, penalties for “over optimization.” I don’t know, I don’t think that’s a good reason to penalize someone. Optimized for data that doesn’t fit the content, now that seems like a valid reason.

And now they have the ultimate spy tool, Google Plus which I don’t feel like I can ignore.

Reply
Rusty Ferguson - April 20, 2012

And what are they doing monitoring my editing actions. I was editing that page with high bounce rate as you suggested last night. The server got slow at that time do to a backup running. (I’m changing how the backup runs.) But it didn’t take 3 minutes to load it. Worse, how do they even freaking know what I’m doing in the admin panel?

/wp-admin/post.php?post=11429&action=edit 178.96

Reply

Leave a Reply:

Notify me of followup comments via e-mail. You can also subscribe without commenting.