WEEK 10
Testing & Tracking: How To Optimize Your Site & Make Sense Of The Figures
Welcome...
Hi again and welcome to Week 10 of my 12 week training course. In last week's installment, we talked about 10 ways to monetize and dramatically improve the profitability of your site. I hope by now you've taken action on some or all of the tips I’ve shared with you and are starting to see the results of the changes you've made to your site. This week, we are going to take a look at how to test and track every aspect of your online marketing. This is a topic that many people like to overlook because it seems difficult, but after I'm done guiding you through how to test and track properly, you see how easy and simple it can be. You'll also never want to leave anything untested, because you will see and understand how easily you can improve your earnings with the simple testing and tracking methods I'm about to share with you. Let's get started right now.
The Art & Science Of Testing & Tracking
One of the main reasons why many people tend to stay away from testing is the fact that it can be a very mathematical process with a lot of statistical number crunching involved. But the good news is that these days that shouldn't be an excuse not to test. There are many tools in the market that simplify this whole process down to a point where all you have to do is figure out what you want to test, create it, let it run, then look at the data. By doing so, you’re able see what the outcome is so you can make changes to different elements of your marketing process. As much as there is a scientific part to testing and tracking, there is also an artistic side to it as well. You'll need a healthy dose of creativity and right brain thinking to come up with new ways to present an offer, write headlines, and calls to action. You'll then have those ideas validated by testing and tracking them.
Why It's Important To Test & Track Everything
It's been said many times that you can't improve what you can't measure, this statement is even more true when it comes to the fast paused, competitive world of online advertising. I don't mean to scare you off, but in this industry if you are not constantly improving and optimizing your marketing some other marketer will eventually step in and take you out. The only defence is constant and ongoing improvements in every aspect of your marketing. You are going to have to run tests daily, track everything with a solid tracking system and make constatnt updates to the way you market based on the data collected. The Japanese word that sums this whole concept up is Kaizen, apply this to your online marketing lifestyle and benefit for the results.
Don't Fly Blind Imagine you are about to board flight with your favorite airline. As you step into the airplane and peek into the cockpit, to your horror, you notice the pilot and co-pilot both pull out blindfolds and put them on. I'm not sure about you, but that would be my cue to immediately alert my fellow passengers, deplane and head to the safety of the airport terminal. When you look at things this way it's obvious that while this scenario is a bit extreme and over exaggerated, it should give you an idea of what you are doing when you choose to not test or track your marketing. It's a certain recipe for disaster, especially when you realize that with PPC and other paid online marketing channels every click or impressions cost you money. It's your responsibility to make sure that every penny investing in your marketing is spent in the best and most optimized way possible
The Customer Is Always Right Another important reason why testing is so important is that many times you get to start thinking that you know everything you can about a particular market, and you believe you know what the market will respond to best. My kind word of caution is to never allow yourself to get into that mind set. Remember tat the market is rapidly and constantly evolving, growing and changing. Your marketing should also follow suit and keep up with what's happening in the real world. Over the my years of marketing online, the best way I've found to always be up to date is to keep testing new ideas and marketing messages, and never to assume I know what the market I'm serving wants. There are many times when I've been surprised when something I thought would never work turns out to be extremely successful, and vice versa. Some of my so-called brilliant ideas that were sure to work have failed miserably. It's great to have your opinions on what should and should work, but eventually you must accept that the wisdom of the crowds is the ultimate decision maker. It's the crowd that will eventually be purchasing the product you are promoting, or file out the lead gen form so give them what they want, talk to them in a manner they will respond to and you will be rewarded. But the only way to really know what the want is to test, test, test.
Different Types Of Testing Methods
Once you start looking into the possibility of testing your marketing, you'll quickly realize that you have a lot of options and ways to test and track. Each method has its advantages and disadvantages, so it's helpful to have a decent understanding of these so that you make the best choice. Keep in mind that you don't have to use all of these testing methods, So feel free to stick to the one that you are most comfortable with. The main important thing that you should realize is that you are ahead of the majority because you are actually testing your ideas instead of blindly believing something works when you haven't backed it up with concrete test data.
A/B Split Testing A/B or Split Testing is one of the simplest and yet effective way to test. At it's simplest form you are basically testing version A against version B. Version A is usually the control, with is just another way of representing you best performing page, ad etc. Traffic should be evenly distributed amongst the different variations of your split test, and When you first start testing your control will be what you currently have, however as you perform more tests, your control will be continually replaced by newer winning combinations. The main thing split test are good for is for testing single changes. For example you may want to test out a new headline concept against the one you currently have. A split test would be ideal for this, and would allow you to quickly see which headline converts the best. Multivariate/Multi-variable Testing A multivariate or multi-variable test is when you test multiple factors at the same time. For example you may have 3 headlines, 5 different graphics, and 4 call to action buttons you'd like to test. Running multiple A/B split tests would be too time consuming and confusing. A multivariate test could test all of these elements at
the same time and use powerful mathematical formulas to determine the combination of elements that is most likely to have the highest conversion rate increase. Multivariates tests work best when you have a large volume of traffic to test with, because each combination of elements needs to have enough people sent to it in order to accurately come to a valid conclusion. The Google Website Optimizer offer both A/B and Full Factorial Multivariate Testing capabilities.
What Should You Be Testing?
Now you know how important it is to test, and also the multiple different testing methods available to you. The big question now is "What should I test?" The quick answer is that you should test everything! But, like a lot of things, when it comes to testing a small handful of factors produce the biggest results and outcomes. In this section I share some of the things I like to test and how you can accelerate your testing process by following my testing blueprint.
Test At All Stages Of The Sales Funnel First it's import to realize that there are multiple stages of a marketing funnel or path. Each stage has many factors that can be tested. When creating your testing plan of attack, evaluate all subsections of the sales/conversion funnel and craft a series of tests to allow you to optimize each point in the process.
Traffic Source At the highest level, you want to test your traffic sources. Understand that for every niche or market you enter, your conversions and performance will most likely vary based on where you get your traffic from. Google, Yahoo and
MSN (Bing) all have different demographic profiles which in turn affects how each of them perform. You may find that for Niche A, it's more profitable to purchase traffic from Google Adwords, while for Niche B your best traffic may come from Bing. A good tracking tool such as Tracking202 will distinguish between all your traffic sources and be able to provide statistical and conversion data based on the traffic source. Using this data you can immediately know where to spend the bulk of your marketing budget. Keywords The next major things to test are your keywords. A few years ago very few people bothered to test down to the keyword level. As long as the adgroup or campaign was profitable everything was ok. If it wasn't then the whole campaign or adgroup would be pased or deleted. This was a very simplistic and ineffective way of marketing, but it happened a lot because there was really nothing readily available to make keyword level tracking easy and accessible to everyone. The bigger marketer who could afford it often created their own tracking systems to enable to do keyword level tracking and enjoyed a period of high competitive advantage over other marketers. Now the playing field has been leveled, and you have no excuse to not track your marketing down to the keyword level. Tracking202 is available in both free and paid versions. Both are extremely powerful and easy to setup and use. With it you can quickly see how each of your keywords are performing so you know which ones are winners and which ones are losers.
Instead of turning of all the keywords in a campaign or adgroup you can now turn of only the ones that are not working. With this simple ability to fine tune your optimizations based on keyword level profitability, it's now possible to take a campaign that's losing money or breaking even and turn it into a profit generating winner in no time. At the overall level, you want to test the performance of the set of keywords you've selected. Further more, I recommend testing multiple variations of your keywords, for instance you want to test both singular and plural version of the keyword is possible. Testing misspellings and typos of a keyword is also a very powerful way to boost your conversions. I have had situations where the misspelling of a keyword I was bidding on get way more traffic and conversion than the correct spelling of the keyword. If I hadn't tested this out, I would have left a lot of money on the table. One of the best tools out there for finding real misspellings is the Keyword Mutation Detection Tool from the Microsoft Adlab You can see in the image above, that i just ran this tool with the keyword insurance, It then generated a list of spelling mistakes that people make. For generating a long list of typos Arron Wall has a popular typo generator script online at http://tools.seobook.com/spelling/keywords-typos.cgi Once again, in the example below I've use the keyword insurance.
One tip is that after generating your misspellings and typos, verify that they have enough search volume by plugging the keywords into the Google keyword suggestion tool. Make sure to uncheck the "Use synonyms" checkbox and in the filters section check the box that says "Don't show ideas for new keywords. I only want to see data about the keywords I entered."
This will quickly give you an idea of the search volume for the misspelling or typo, and ensure you don't waste time on any low volume variations that may not generate much volume. The image below shows the output of the keyword tool after running it with the setting above. You'll notice that some keywords have no volume data, these are the ones you can cut out.
Ad Copy When it comes to ad copy always have more than one variation of your ads running at all times. Go though all your adgroups and add at least one more ad variation to every adgroup that currently has only one ad. The goal is to always try to beat your best performing ad, take the attitude that there's always room for improvement. Ad Headline Your headline is one of the biggest things you can test, because the headline is usually the first thing a searcher will read and pay attention to. Once his or her attention has been caught by the headline, the rest of your ad copy will be read. So if you only have the time or results to test one aspect of your ad copy, the headline would be the element I'd recommend you focus on the the most. Try to draw attention with different messaging, capitalization and special characters. Ad Description Coming up with the body/description copy of your ad can be difficult at times. But when you are stuck, some of your best inspiration can come from your competitors. Go to Google, Yahoo or Bing and perform a search for the keyword you are writing ads for, The ads on page one are a great starting point. Pay specific attention to the premium ads because those are usually examples well written ads with high CTRs. When you are eady to start writing ads, you can also use the ShoeMoney Tools PPC Ad Generator to pull in these ads you just saw from the search engines.
Once you've pulled in ads to work with, you can start modifying them using some of the Ad Copy CTR Enhancers I'm about to share with you.
The Famous ShoeMoney Sideways Arrow - Create a sideways arrow with your main keyword and watch your CTR increase Use Numbers - Numbers are easier to read than words, they also take up less space leaving more room for ad copy. Take a look at these example snippets, 7 Great Tips, 50% Off, Save $34, 2009 Capitalization – Try capitalizing the first letter of every word, it improves readability and clicks. Special Characters - This one is big, the Trademark, Copyright and registered symbols do wonders for your CTR. So does the exclamation, quotation, and question mark Try them in your ad copy and headlines. "Look", What?, Wow!, Original©, Unique®, Exclusive™ Call To Action - Ending your ads with a powerful call to action which usually include words such as Buy, Order, Get, Try, Visit, Sign Up, Reserve, Download, Find, Obey Make it Time Sensitive - Create a sense of urgency by using works that have some sort of "deadline" built in. Examples include: Today Only, Right Now, Today, July Special, 2009 Sale
Display Url One thing that a lot of beginners fail to realize is the importance for the display url, and how a simple change can boost CTRs and conversions. It's important to not that adwords will not let you test completely different domain names in the same adgroups. So for example if I wantedt to test out ads that had www.shoemoney.com and www.shoemoneyx.com in the display url, I'd have to creat two seperate adgroups. This relativly new rule make it a bit harder to split
test completly different display urls, however you can kind of get around this by duplication the adgruop for each display url you like to test. The results may not be the same because in Google if two or more adgroups share the same keywords, the adgroup with the best performance will get the majorit y of the traffic. This makes it difficult to test multiple varaitions of the display url at the same time. Another idea to try would be to test each varaition over the course of a week. This make it take a little longer to get your test results, but due to the new limiations there's not much we can do about it. In the table below, I've listed some display url variations I test out when I'm split testing my display urls. This is not even a full list, but alredy you will notice that there are 60 different ways to present you display url. Some of these may not work out for you due to display url length limitations.
domainname.com/keyword-here domainname.Com/keyword-here DomainName.com/keyword-here DomainName.Com/keyword-here www.domainname.com/keyword-here www.domainname.Com/keyword-here www.DomainName.com/keyword-here www.DomainName.Com/keywordhere keyword.domainname.com/keywordhere keyword.domainname.Com/keywordhere keyword.DomainName.com/keywordhere keyword.DomainName.Com/keywordhere domainname.com/Call_To_Action domainname.Com/Call_To_Action DomainName.com/Call_To_Action DomainName.Com/Call_To_Action www.domainname.com/Call_To_Action www.domainname.Com/Call_To_Action www.DomainName.com/Call_To_Actio n www.DomainName.Com/Call_To_Actio n keyword.domainname.com/Call_To_Act ion keyword.domainname.Com/Call_To_Ac tion keyword.DomainName.com/Call_To_Ac tion keyword.DomainName.Com/Call_To_Ac tion domainname.com/keyword_here domainname.Com/keyword_here DomainName.com/keyword_here DomainName.Com/keyword_here www.domainname.com/keyword_here www.domainname.Com/keyword_here www.DomainName.com/keyword_her e www.DomainName.Com/keyword_her e keyword.domainname.com/keyword_h ere keyword.domainname.Com/keyword_ here keyword.DomainName.com/keyword_ here keyword.DomainName.Com/keyword_ here
domainname.com domainname.Com DomainName.com DomainName.Com www.domainname.com www.domainname.Com www.DomainName.com www.DomainName.Com keyword.domainname.com keyword.domainname.Com keyword.DomainName.com keyword.DomainName.Com
domainname.com/Call-ToAction domainname.Com/Call-ToAction DomainName.com/Call-ToAction DomainName.Com/Call-ToAction www.domainname.com/CallTo-Action www.domainname.Com/CallTo-Action www.DomainName.com/CallTo-Action www.DomainName.Com/CallTo-Action keyword.domainname.com/C all-To-Action keyword.domainname.Com/C all-To-Action
keyword.DomainName.com/C all-To-Action keyword.DomainName.Com/C all-To-Action
How To Test Ad copy In Adwords
When using Google Adwords, there are few important steps to take to ensure your split tests are no skewed or invalid. By default, Google will automatically try to determine the best performing ad sn show it more often than the other ads. The problem with this is twofold, first you now have no control over how long to run your tests before a winner is chosen. The second problem is that Google optimizes for CTR which may not always be your main goal. You may want to optimize for ROI and profits instead. To turn off the automatic built in ad optimization algorithm, follow these steps.
Helpful Tip: Even with the campaign set to even rotations, you may notice that certain ads show more often than other ads. This is a result of how the Google ad auction system works. In order for an ad to be displayed, it has to meet all the requirements for even being eligible to be a part of the ad auction. Higher CTR ads are more likely to meet this requirement and therefore a shown slightly more often than lower performing ads. Even though the impression could may end up a little lopsided the numbers are usually close enough to run a successful test..
How To Test Ad copy In YSM
Yahoo Search Marketing is another network that automatically optimizes your ads to select the best ad copy. Once again to disable the optimization algorithm, follow these steps. First select the campaign which has the adgroups you'd like to set even rotation on. On the next page select all the adgroups with ads you'd like to enable for even rotation. Next click the "Optimize Ads" button, and set "Optimize Ad Display" to "No"
How To Test Ad copy In Adcenter
Unfortunately at this point in time there is no way to turn on even rotation for adcenter ads, you'll have to rely on the ad optimization algorithm in adcenter for testing of your ads. To do this you'll have to create a brand new ad group. This is to make sure that historical performance on ads do not skew results. Next add in the ad copy and the keywords you want to test.
Landing Page Testing Testing on the landing page level opens up a whole new possibilities that are not available at the ad copy level. This also means that a new level of complexity gets thrown into the mix. For example when split testing in a PPC network such as adwords, the testing functionality is already built into the system, however once you start to test your landing pages you now need to have an external tool of some sort that will perform the tests and report back on the results. The are a number of tools out there, but let's keep things simple and best of all free. The #1 free tool for this is the Google Website Optimizer, available online at http://www.Google.com/websiteoptimizer or through the Tools menu in your Adwords account.
I covered the website optimizer in detail during Week 5's advanced PPC lesson, so please review the information covered in that chapter if you need a refreshers. I'll quickly go over some of the important elements to test on your landing page again: Headline - It's such an important aspect of your landing page and can sometimes be the single biggest issue affecting your conversion rates. Calls To Action - Believe it or not, people need to be told what to do. You call to action is your opportunity to make your visitors take your most wanted action
Colors - Colors play an important part in creating feelings and emotions. Get your color pallet optimized and you'll see an increased conversion rate Graphics - A picture is worth a thousand words, but you have to get the right picture so the right words are said. When testing, always make it a point to test the performance on multiple graphics, you'll be surprised at how much your conversion rates are tied to your choice in pictures. Layout - Have you considered how important the placement of all your page elements are? Certain layouts will produce lower conversions because of how difficult it is to find or see the important conversion generating elements such as the order button or a headline of some sort. Test your layouts, and keep in mind monitor resolutions can affect how a page looks to the end user Offers - Not all offers are created the same. Always split test offers to find the one the works best for your traffic sources. This tips alone can more than double your conversions.
A list of Cool Resources to help you test and track better: Google Website Optimizer Tracking202 Google Analytics Crazy Egg ClickTale ShoeMoney Tools
Don't Make These Testing Mistakes
To wrap things up, I'd like to highlight a few things to watch out for as you begin testing and tracking: Jumping to conclusions - Take your time and don't get tempted to stop running your tests too soon. Make sure you have enough data to back up your test results before you even think about choosing a winner Changing too many things in one go - If you are running a split test, don't try and make too many changes, it’s ideal to stick to only one change so you know exactly what you did to product your results. If you change a headline and a graphic and end up with a 50 % increase in conversions it's hard to tell whether it was the headline or the graphic that helped boost your conversion rate. It could have also been both. Picking the wrong kind of testing method - Split tests work best with lower volume traffic, don't try and run a multivariable test on a site that gets hardly any traffic. Due to this nature of these kinds tests, you'll need higher amounts of volume to be able to run them in a timely manner
ShoeMoney Action Plan: Week 10
Make sure that you do not run any more OOC traffic until you are tracking everything. If you do not have a tracking tool sign up for Tracking202 at http://www.tracking202.com Identify a current campaign you like to test Write down 2-3 goals you'd like to achieve from this test Make a list of all the point in the sales/conversion funnel you like to test Use ShoeMoney Tools PPC Ad Generator (https://tools.shoemoney.com/documentation/PPC-ad-generator-2/) to generate some ideas for you ad copy Identify a few of the CTR enhancement techniques I shared with you and decide how you'd like to use them in your ad copy For each ad group write at least two ads to test Make sure you set up your campaign to evenly rotate your ads to ensure the data isn't skewed Run your ad copy test for at least a week before picking a winner Remember to use http://www.PPCtools.com/splittestz to analyze your results to ensure statistical significance Identify elements you'd like to test on your landing pages Setup a Google Website Optimizer (GWO) campaign to test out your landing pages Run the experiments until the GWO indicates a winner has been found.