Questions? (877) 643-7244
FutureNow Post
Wednesday, Feb. 2, 2011 at 10:15 am

The Need for Speed: Your Website Optimization Plan

By Melissa Burdon
February 2nd, 2011

Monday, our Senior Digital Marketing Analyst, Brendan Regan, was interviewed by Tim Ash on WebmasterRadio. Tim and Brendan chatted about FutureNow’s testing experiences with clients. During the interview, Brendan details some common problems his clients experience, explains what data tells him about his clients’ customers, and reveals why it is important to test and implement winning variations quickly.

Listen to their chat here »

Many clients don’t know what to look for in their data.

Brendan notes that many of his clients have identified and track some basic KPIs but have a hard time really digging into the details. For instance, he mentions the fact that many clients look at their aggregate conversion rate (CR), but fail to break that down into conversion rates for various segments. That’s a problem because some segments of traffic are earlier in their buying process, and may convert at a low rate, while other segments of traffic have a higher level of buying intent, and convert at a very high rate. The aggregate CR is the average of all these segments and all their conversion rates. This average tells us very little about where the problems exist and therefore, sheds little light on how to make improvements that will make a difference in the bottom line.

He mentions that segmenting traffic into custom segments such as early stage buyers and later stage buyers, helps you look at the data in more detail and really dig into where problems are. Doing this will show you where the real opportunity exists. An example of some key things Brendan suggests to look at for each of these segments are:

1. Conversion funnels – identify where the largest traffic drop-offs occur

2. Top landing pages – identify which ones have high bounce rates

If you can find some of the big problems in the data, this is likely where you’re bleeding money. Brendan points to the Pareto principle as support: 20% of the challenges on your site are likely responsible for 80% of your pain.

So, once you find those problem areas on your site, how can you most effectively make improvements and increase revenue?

Tips to get moving with a testing plan.

Brendan mentions some things you can do to get started with testing:

Don’t view a losing test as a failure. A losing test still gives you insight into the buyer’s behavior and will result in effective follow up tests. Sometimes it takes one or two consecutive losing tests to figure out how to get a big win.

Test a range of circumstances and don’t play it safe. Running timid tests sometimes results in tests that run too long, or that don’t reach significance. Be brave, and test things that seem radical, or things that might not appeal to you. Realize that you are not the target audience. Radical variations lead to spectacular results; either clear wins or clear losses. Either way, you learn from each test result.

Run tests quickly. Win quickly and/or lose quickly, so that you can learn, implement, and move on to continue increasing revenue. If you have a successful test end quickly, you can implement the winning variation quickly and bank the extra revenue quickly. Build on the winning variation as a baseline and continue testing and improving.

Brendan helps reinforce this approach every day with his clients.  If you’d like to work with him, or one of FutureNow’s other wicked-smart Digital Marketing Analysts, consider a subscription to our OnTarget program of continuous improvement.

Add Your Comments

Comments (37)

  1. It’s amazing what small changes in a web site can do for your conversion. Just moving a button or changing a color or font can make a huge difference. Software that split tests between two different versions of your site is a life saver!

  2. you are exactly right, I try to change the font and color when doing the test, and the results is changing the conversion

  3. I totally agree with what you said. One of the useful tools I’ve found is Clicktail, it let’s you view actual videos of clients navigating your site. You can see where they are clicking and what they are reading, and most importantly what they are not reading. I use it on my website.

  4. Eye opener !
    How can somebody think that changing few things here and there make a drastic change in SEO performance.
    Thanks for sharing.

  5. This is a very interesting topic and a great article. I was reading a case study recently where a sales letter achieved a massive increase in conversions simply by changing the word ‘repair’ to ‘fix’ in the main headline.

  6. Very informative your blog is…I think you can solve my problem.Can you tell me the tool to analyze a website before promoting…

  7. There are many kinds of conversions and it’s important to measure them all separately. It will give you key insights into what your audience is looking for. Then, you can optimize based on the findings.

  8. @George – We’re not talking about impacting search engine rankings. We’re talking about impacting conversion rate (to sale or lead) on the website.

  9. @web development blackburn – that depends on what specifically you want to analyze about the site. You can look at the site from a usability standpoint. We actually look at our clients’ sites from a conversion standpoint (trying to get more of their existing traffic to take action. We use a variety of tools when working with our clients. The big ones we use are an analytics tool (we use Google Analytics), a testing tool (we use Google Website Optimizer) and our own tool for making recommendations (called OnTarget). We do this through our continuous improvement system for websites, also called OnTarget. Your website needs to be live to do this.

  10. @Nick Stamoulis – precisely. It is important to have a system for figuring out where to start though Nick. We like to start by looking at late stage traffic first, and trying to maximize the wins there…. that traffic is closer to handing over the money that can be used to optimize the other conversion rates on the website! ;-)

  11. I definitely echo the advice about not playing it safe. You’re not going to increase conversions by changing the font. (I mean..maybe you will, are you using white on white?) You have to take risks to make money. Change the layout, change the entire color scheme. Add some arrows here an there. Do things that you think will influence conversion. Making subtle changes will have you waiting a month to see inconclusive results.

  12. Build on the winning variation as a baseline and continue testing and improving.

  13. Amazing, by doing such little things can effect SEO performance. That’s brilliant!

  14. @Marijayne Bushey.Thanks for the reply..It nice to use Google analytics..Is it good to optimize the site with “DoFollow” keyword?

  15. @Amulets – Brendan and Tim were not talking about SEO (how you rank in search engines), but about CRO (how many people take the desired action on your website).

  16. If you make the landing page to fussy you will put people off, to simple and you will put people off! its a fine line

  17. Thats some sound advice for an online marketing strategy. Im becoming quite a fan of the A/B split testing methods too as over time they allow you to really gather some valuable feedback on your website/webstore.

  18. Useful info thanks…I’m always concerned about what visitors are doing on my site (or not doing), maybe it’s the voyeur in me, but having been in retail for many years, the one thing I miss is not getting that instantaneous feedback about what customers like or do not like!

    But…is there a danger in over analysing, over testing, over changing?

  19. sometimes small changes are enough for the site becomes more attractive. my site earn with adsense. the last week with 2500 hits with adsense profits were small, it was just change the position of banners and leaves them more attractive to income almost doubled, even on some days the site ranging in less than 2000 visits!

  20. I find that the most important thing is various call to action methods. Just as another commenter stated, it’s all about that small color or font tweak that can drive tons of changes.

    Google Analytics, with their new intelligence view. It’s awesome!

  21. I break my site down into “advertising space” Things that I want to push to the user I place strategically within the site. I rotate this several time per month and it helps establish a baseline to use in the future.

  22. many people think that buying a hosting plan, uploading some *.html files with some nice pictures, a good article and some meta full of keywords will push them on the 1st page of SERP… well.. its not quite so easy.. :) and this article is the proof! :) good job!

  23. @webdesign – seo – hosting: ironically, you are correct, even though this article isn’t about how to rank better in search engines. It’s about how to convert traffic better.

  24. “Don’t view a losing test as a failure.” This is good advice, people can get down on themselves when they see they put some much work into something then realize they did it wrong. Ive been there before.

  25. Nice article, I think too many online marketing folk and the clients themselves look at this far too late, but it’s one of the most important things of a successful campaign.

  26. I can’t agree with this more…

    1. Conversion funnels – identify where the largest traffic drop-offs occur

    2. Top landing pages – identify which ones have high bounce rates

  27. Great article – it’s very important the experts translate their knowledge into tracking correctly. Conversion funnels are key to streamlining your download/ecommerce process to ensure it’s user-friendly and effective.

  28. Interestign article, I’m going to review our campaign based on these points, the comments have been useful also. Is it industry standard to use Google Analytics?

  29. Not only is it important to test your web site quickly, it is very important to test how quickly your website responds. With Google placing more emphasis on page load times, this web site speed is one component not to overlook1

  30. @Cloud Computing: Google Analytics is free and does almost everything we need it to do for our clients. :-)

  31. It really doesnt help if you get a lot of traffic to your website but your conversion rate is low. The tricky part is to determine where you are going wrong in converting a visit into a lead/sale.

  32. I do not agree that small changes make a big difference in the cr. If you have a really bad design you can do alot to improve it but font size and button color will not make you rich :) Make it easy, nice and trusted.

  33. I totally would have to agree with this. One would have never thought that single update to a color or a font change would improve conversions. However, you also have to keep an eye out of loading times of the site, as that can hurt your visitors if your site loads too slow. I would recommend looking at something like a CDN network if thats an issue your having.

  34. @Mads – you must have missed some of these recent case studies of the work we have done (making a continuous series of small changes) with our clients:
    1) added product reviews to only one product page, additional $3,000+/month in sales
    2) 25% increase in conversion, additional $9,000+/month in sales
    3) changes to reassurances, guarantees, shipping notice and calls to action resulted in 40.7% increase in micro-conversion rate

    And even better is when changes we do with clients not only result in significant changes in the conversion rate, but those changes compound upon one another over time with continued testing, as we see here:
    1) changes to background color and copy in contact us form on homepage, 200% increase in conversion to lead
    2) making product images click-able and adding short descriptions under products on a product option page, another 20% increase in the number of leads generated
    3) change in layout of technology page, another 9% increase in leads, and almost $3,000 additional revenue each month.

  35. @Marijayne – Yes I missed them but will read em :)

    Congratulations on the good results. I do not think that it has no meaning. But my experience is just that if you have a really good site with a great conversion rate, you can not increase the conversion much by color of a button. I changed the background colors, etc., without any change in conversion.

    I think it depends on what you have to work with. Switching from pink background color with yellow text to gray background color with black text will help but shift from light gray to gray?

  36. Testing changes on your own site is really important. I know many people just copy wholesale from other sites, e.g. website layout, color combination, colors of buttons, etc. assuming that they would be equally successful. This isn’t necessarily the case. There are so many variables that contribute to conversion, e.g. industry, demographics of the site’s visitors, the product/service on offer, etc. These days A/B testing is so easy to do, there’s really no excuse not to test changes and see their results. I also like your point about being agile – testing means you have to be quick to respond to their outcomes. If the changes being tested lead to better results, then they should be implemented on a bigger scale. On the other hand, if they don’t seem to yield the desired results, they should be modified or dropped.

  37. By the way,according to a study released yesterday by Strangeloop Networks, it takes an average of 10 seconds to load the top 2,000 websites on the internet. The list of these top sites has been compiled from Amazon’s Alexa list of top websites. Though some analysts argue that these results could be biased as Strangeloop is in the business of helping customers speed up their websites, they can still help developers optimize their sites better.

    Olivia

Add Your Comments

 

Print this Article
Share

Melissa is a Senior Persuasion Analyst at FutureNow.

More articles from Melissa Burdon

Marketing Optimization Blog
FREE Newsletter Sign-Up
send it once every: