You Should Test That!

Daily eBusiness Optimization Ideas

Legitimate SEO Cloaking Technique (with Google’s Approval)

February 17th, 2010 by TraiaN No comments »

Warning: if you’re going to implement the technique describe below, do it at your own risk!

That’s right. You can you Google’s own tool, Website Optimizer to implement a nice cloaking (well, technically it is not cloaking, but the results are the same).

So what SEO cloaking mean? Simply put, you display some content to search engine bots and another one humans, with the purpose of manipulating the search engine rankings. Most of the cloaking scripts are identifying the IP of the user agent (humans or search engine bots) and based on a predefined list of IPs of search engine bots will try to guess if the visitor is a bot or human. Others, have bots traps to find robots. Based on the findings, you will setup your web server to serve the tricky content to SE and nice looking content to humans.

What does Google Website Optimizer do? While the technicalities are not the same, the result is almost the same: it displays two (or more) versions of a page to some visitors, and other content to another visitors. For a simple A/B test 50% of your visitors will see the page A and the other 50% will be automatically redirected to page B (test page). The result of GWSO is almost similar to cloaking: a redirect; while Google will keep the original page in their index and therefore will rank it high, your visitors will see a test page (not necessarily indexed or ranked in search engines, but high converting). Make sure search engine will not index the test page.

Scenario

You rank high with a page, but the page is not converting well. You design another page and run a test to see if it’s performing better. Let’s say it does convert 50% better than the original. However, when you replace the old page with the new one, your ranking will drop and you don’t have enough visitors anymore.

So, how could you balance this? How do you rank high at the same time with having a high converting page? Well, you can use Google’s own tool for the trick (at your own risk, don’t blame me for getting banned).

First, you will run a test and identify a better converting page. Next, run a follow up test and split the traffic 95% to the converting page and 5% to the page that is ranking high. Let the test run for a while (until it conclude and a little more)

This way your high ranking page will show up in SERPs and your converting page will persuade people to act. That’s it. As simple as that!

Honestly, if you continuously test and search for a better converting page, at the same time with keeping the old one live, I don’t see why Google will ban your website. After all that’s the purpose of website testing: continuously testing! However, keep in mind to stop the test and rerun/run another one once you’ve identified a.

Disclaimer: please read Google’s policy on testing and cloaking!

GD Star Rating
loading...

No comments »

Posted in Conversion Optimization, Others

Tags: cloaking google website optimizer

Why Bounces is a Better Metric than Bounce Rate

February 12th, 2010 by TraiaN No comments »

This is the follow-up post on the bounce rate poll from Can You Answer this Simple Google Analytics Question?. If you did not vote yet, please do it before you read further.

The results are the following:

The correct answer is marked in the green box

As you can see, only 20% the voters got it right (only 25 votes). It’s not surprising as some of the reports in Google Analytics can be misleading, i.e. top content, content by title or content drilldown.

To find out the number of bounces (or the real bounce rate) one have to look at the landing page report, not at the content performance reports (top content, content by title or content drilldown). The reason is for that most of the time, a page will have a bounce rate only if it serves as landing page (single page visits).

For the example used in this poll

Misleading Bounce Rate

if one multiplies 826 (unique page views) by 52.17% (bounce rate) it results in 430 bounces, which totally wrong. Let me explain.

If you look at the landing page report (image below) for the same page, same time interval (dec 11, 2009 – jan 10, 2010), you’ll see that while the bounce is the sameas on any of the content performance report (52.71%) the real number of bounces is 24, which is not by far 52% of 826, but a merely 3% of it.

The Real Number of Bounces

If one will look at the Top Content report and see 10.000 unique page views (almost equal to visitors) with a 50% bounce rate, it might wrongly conclude that 5.000 visitors bounced.

So where is this confusion coming from? While the data in the Top Content, Content by Title and Content Drilldown reports are based on page views (unique or not) the bounce rate column is actually taking into related with data from other report and it’s computed based on the entrances to that specific page.

I think that the bounce rate in the aforementioned report should be supplemented by the Bounces column or. For that you can use the following custom report to have the bounce rate and the bounces in a single report.

If you wish you can use this custom report to look at the page views (I was not able to combine Unique Page views with Bounces in Google Analytics custom reports). I am looking at the page view columns as a relative importance of the page, then I look at the bounces and if the number is high compared to the page view I take a quick look at the bounce rate.

Custom report with pageview, bounces and bounce rate

Rate this post!

GD Star Rating
loading...

No comments »

Posted in Don't Do Like Them, Google Analytics

Tags: bounce rate Google Analytics

Google Is Pulling the AdWords Plug for Canadian Online Pharmacies

February 9th, 2010 by TraiaN No comments »

There is no surprise on this, based on the online pharmacy policy changes during the last months. Starting the end of February 2010, Canadian online pharmacies will not allowed to advertise in the U.S.

However, based on my experience with several clients in the industry, I think that Google will lose at least $50M/year. Something must’ve happen with FDA and Google.

Read the announcement here.

Rate this post!

GD Star Rating
loading...

No comments »

Posted in Others

Tags: canadian pharmacies and adwords

Minimize the Impact of Website Testing on SEO

January 28th, 2010 by TraiaN 3 comments »

A real concern anyone involved in A/B/n or multivariate testing should be worried about is what’s going to happen to the search engine optimization (SEO) efforts while, and after the test is done.

If you don’t set up and follow up your tests correctly – and I don’t mean how you setup your testing platform, but the technicalities of how you implement the test on your web server – you can encounter two major problems:

1. Duplicate content issues in search engines. In this case, the control version can be totally excluded from the index or the ranks can decrease

2. Browser bookmarkers. In case if your variation pages change URLs and they are bookmarked by visitors with the changed URLs, once a winner is found and you removed the variations URL, visitors will reach a 404 not found page

To address these issues I do the following:

1. Add the following meta tags in ALL variations of the tested page:

This will tell search engine spider not to index the page, not to follow the links and not to archive the test page(s) in their index. However, this is not working all the time so,

2. If your tested page changes URL (i.e. testing /olympics.php against /olympics_test.php) you should put the test page(s) under a different directory (i.e. www.mysite.com/testing/) and

3. Exclude tested URLs (complete directory of specific files) in your robots.txt file

To exclude directories use:

User-agent: *
Disallow: /testing/

To exclude just the file use (not recommended):

User-agent: *
Disallow: /testing/olympics_test.php

4. If you’re testing variations using the same page URL with different parameters (i.e. /olympics.php against /olympics.php?test_ver=5) use the rel=”canonical” meta tag, to tell Google and the other search engines that the control page is the page they should index.

I add the line below in the <head> section for all variation pages (i.e. /olympics.php?test_ver=5):

<link rel=”canonical” href=”http://www.mysite.com/olympics.php” />

5. Tell you programmers and link builders not to link to your testing directory/page at all (IT should be careful about internal linking, and link builders should not link from external websites)

6. If you plan on testing minisites (more pages under the same testing directory) use “rel=”nofollow” on all internal links of the minisite. Additionally you can put all the links under Javascript so they won’t be followed by search engine bots

7. Once the test is concluded and you found a winner, do a permanent redirect (301) for all tested URLs to the winning URL. For example, in your .htaccess you can add:

RewriteRule ^(.*)mysite.com/testing/olympics_test.php$ http://www.mysite.com/olympics.php [R=301,L]

If you’ll implement all of the above you will have a better chance of not getting into duplicate content issues or ranking fluctuations, due to testing and experimentation.

However, if your pages do get indexed by search engines, you should take the necessary steps to minimize the impact on rankings, as soon as possible:

1. create a Google Webmaster account and keep an eye on your 404 page errors. Eventually ask for the test pages to be excluded from their index

2. create a custom 404.html page, then track and fix 404 errors with the help of Google Analytics. On your 404 error page, add this to your page tracking code:

pageTracker._trackPageview(“/404.html?page=” + document.location.pathname + document.location.search + “&from=” + document.referrer);

Your code should look something similar to:

try{

var pageTracker = _gat._getTracker(“UA-xxxxxx-x“);

pageTracker._trackPageview(“/404.html?page=” + document.location.pathname + document.location.search + “&from=” +

document.referrer);}

catch(err) {}

</script>

I hope this will address your boss’ concerns on SEO and it will provide more flexibility for testing!

GD Star Rating
loading...

3 comments »

Posted in Conversion Optimization, Google Analytics, SEO

Tags: testing and seo

Avoid The Pitfalls of Computer Generated Heat Maps

January 22nd, 2010 by TraiaN 21 comments »

I see a rising trend for computer generated heatmaps of websites, used for landing page optimization and conversion rate optimization. I am updating an older post regarding such tools.

If you’re using services like feng-gui or attention wizard you should be aware of the following:

computers are machines and even if they can detect skin, faces, smiles or eyes within pictures, they can’t create heatmaps for emotions associated with the image. Be honest, are you looking at the picture below exactly like computers do?! (scroll to see how a computer will look at Pamela Anderson).

Pamela Anderson

Computer Generated Heatmap on Pamela Anderson

human eyes (the direction they point at) and smiles (large, white teeth smiles) also play an important role on where your visitors will look at. An image depicting model’s eyes looking directly at the user will attract his eyes on the model’s eyes, while an image with the eyes of the “hero” pointing to a product (looking to the left or right) will “heat up” the product image. From GrokDotCom, here’s a real (made on humans) eye tracking heatmap:

Model looks at you (left) and at the product (right)

a website seen on a 1024×600 resolution will have a different heatmap compared with the same website seen on 1920 x 1200. What’s above the fold still matters, not matter what anyone will say. You need to analyze and optimize the website for most used resolution of your audience. Below are screenshots for the same website taken at different resolutions:

The heatmap at 1024x768 res

Same website heatmap taken at 1920x1200.See the differences? Not sure if it's even technically correct

foveal and peripheral vision can also affects the eye movement and generate heat maps. The way you place elements on a page (headlines closers to an attractive image) will play a difference and I doubt current computer algorithms are taking this into consideration (via entrepreneur.com)

The hotspot seen by human eyes as per foveal view

banner blindness also has some influence (how would a computer know that an image is actually a banner and take banner (via useit.com)

Banner Blindness Eye Tracking Study

animation will change the hot spots

Rate this post!

GD Star Rating
loading...

21 comments »

Posted in Conversion Optimization, Landing Page Optimization, Usability and Design

Tags: heatmaps

Who Else Does the Same Mistakes as Search Engine Land?

January 20th, 2010 by TraiaN No comments »

Here’s a good sample of a poor concept of a registration page (I did not expect searchengineland.com to do such mistakes):

search engine land mistakes on the register page

Huge Mistakes on Search Engine Land Registration Page

I will point the major problems only:

1. primary action (register button) is below the fold

2. the registration button doesn’t look like a button at all

3. secondary actions have more prominence than the desired purpose of this page (to register and create an account)

Funny thing, when I took the screenshot, the captcha was the bounces :)

To address the issues above you need to:

1. always have the main call to action above the fold

2. CTA buttons should look like buttons

3. identify the primary purpose of each page and make it easy for visitors to reach that goal. Secondary goals should not overwhelm the primary purpose

Rate this Post!

GD Star Rating
loading...

No comments »

Posted in Conversion Optimization, Don't Do Like Them, Usability and Design, Who's Testing

Wordpress Comments Not Working? Here’s the Fix!

January 19th, 2010 by TraiaN No comments »

If you’ve recently upgraded Wordpress to version 2.9.1 and you are not able to post comments on your website, disable Akismet plugin (try activating it again with a later update) plugin and you will be able to do it again.

Rate this Post!

GD Star Rating
loading...

No comments »

Posted in Others

Tags: wordpress comments fix

Who’s Testing What #30 – www.vacationtimesharerentals.com

January 18th, 2010 by TraiaN No comments »

This is an MVT test from www.vacationtimesharerentals.com:

Vacation Time Share Rentals - Original

Vacation Time Share Rentals - Variation 1

Rate this test!

GD Star Rating
loading...

No comments »

Posted in Conversion Optimization, Who's Testing

Tags: Conversion Optimization Landing Page Optimization Usability and Design Who’s Testing

When Government Website Design Goes Wrong

January 15th, 2010 by TraiaN No comments »

There’s nothing worse than a government website using free stock photos for it’s website. Especially when using images with people.

Why? In the case of the website below, the “pharmacist” in the picture is found on quite a few of bogus pharmacies. And the website is about medical services plans in BC, Canada. The association of the Government, thru that picture, with fake pharmacies is not a good think for them:

BC's Governmental Website and The Bogus Pharma Guy

Rate this post!

GD Star Rating
loading...

No comments »

Posted in Don't Do Like Them, Usability and Design

Tags: bogus pharma government design

Trust Seals Collection and How to Use Them to Increase Conversion Rates

January 14th, 2010 by TraiaN 1 comment »

Trust seals is the generic phrase used to depict those images (see collection at the bottom of the article) which online businesses use on their websites to address Fear, Uncertainty, and Doubt (FUDs).

Seals can be categorized into privacy seals, business validity seals, PCI compliancy seals, hacksafe seals and other types. It may be useful to know that there are seals which are applicable to certain industries only (like the Pharmacy Checker for Canadian pharmacies) and seals which can be used on most of the websites (business verification seals).

Smart online marketers use seals on websites to transfer trust from a well-know company (i.e. Verisign) to visitors and to make them feel comfortable about their privacy, confidentiality or other concerns. In our terms, we use them to increase the conversion rate of websites.

Do they really have an effect on conversion rate? Yes, they do, but each seal will perform different, simply because your audience is different than the other website. While I do recommend using some of the seals, namely those which anyone will recognize, I advocate for multivariate testing. Most likely, if used correctly the impact on conversion rate should be either positive or neutral. If improperly used (clicking is not working, fake seal, etc) your conversion rate could suffer.

Having 20 trust seals on each page of your website will hurt. Having none, may hurt also. For site wide use, you should display 2 or 3 of them, which are important and recognized by your audience. However, if you’re would like flex you muscles with 10 or more (which I don’t advise), I would recommend creating a dedicated page on your site, where you can lay all of them.

Here are some ideas when it comes to using and testing trust seals:

start testing with 2 or 3 seals: one wide known (i.e. McAfee) and your industry-specific seals
always display the seal above the fold; if your copy is way longer, repeat it near your call to actions
test overlaid seals (i.e. comodo, remains in the same position at scroll) against layout-integrated seals which have fixed position
seals should be displayed site-wide (on all pages)
except for checkout pages, where you may have a different layout, keep the seal in the same spot for the whole website

Additionally:

a good place to test seals is near the logo (top left) or on the top right corner of the site
test the logo near the main Call-To-Action (CTA) button
display seals in the checkout process
at click on the seal, open the merchant site in a new window

Those are only some ideas, but you I am sure that you can think of your own tests. One thing in particular (never tested by me thou) is to test the effectiveness of a no-spam seal in email blasts :)

Here’s the collection listed in alphabetical order – click seals to enlarge – (back to top):