Warm Regards,

Scott Miller, CEO

Author of “The ConversionLab.com,” The worlds first “How to” guide to testing and optimization. Visit http://www.conversionlab.com/ for a special offer!

Traditionally, web tracking has been done with a third party cookie. Within the last couple years, third party cookies have come under increasing criticism due to privacy concerns.

Consider the tracking cookies you certainly have on your system. Owners of these cookies can essentially track and archive every site you visit on the web that contains their scripting (if you let them.) Every single move you make is recorded, and stored for eternity. This is kind of scary if you ask me.

Check out our friends at Google for instance. Genius aren’t they? Google is best known for being the king of search. Rightly so. I think they are also the king of the cookie. They have code on many, many, MANY websites that they are using to spy on us, and log every move we make. In fact their analytics program is reportedly used on 400,000 sites now.

It’s no wonder that anti-spyware makers and vendors of internet privacy software have begun a “seek and destroy” campaign against third party cookies. With the resulting war on third party cookies, they no longer represent a viable way to track test results or any results.

Long Live the [1st Party] Cookie!

First party cookies do not suffer the same challenges of being deleted or rejected that the more common third party cookie faces. The reasons are as simple as they are compelling.

The first party cookie can only be viewed by scripts on the domain that set it, making it nearly impossible to track a users movement from one website to the next. This is good for privacy as well as conversion tracking. As users, we get the peace of mind that big brother is not looking down on us. As web experimenters, we get more accurate results.

Long Live [Testing]

It turns out that third party cookies are horrible for testing, because they are deleted so frequently. Depending on what you are testing, this could really throw your stats off.

I recently tested a homepage, all the way to a purchase (not recommended by the way.) A full 50% of users converted latently, many after two weeks. If you were using a third party tracking cookie, that would mean losing 50% of your data. Or even worse. This makes getting a statistically reliable result impossible, especially when the difference between the pages may be as small as .5%.

Warm Regards,

Scott Miller, CEO

Author of “The ConversionLab.com,” The worlds first “How to” guide to testing and optimization. Visit http://www.conversionlab.com/ for a special offer!

Here is a summary of what you will find inside this book:

* The most important phase of a test, that doesn’t even involve your website!
* When to use Split Testing and when to use Multivariate tests to get the highest ROI from your activities.
* How to test without affecting your Natural Search Engine rankings.
* 25 Common sources of Experimental Noise that can render your results completely useless.
* Which is better: A script you install on your server, or a hosted service?
* The 7 key categories of page elements that literally define your success
* Why most marketers are incorrect when they optimize by looking only at the conversion rate
* A library of 50 of the most important testable items on any site
* Includes a review of 5 key attributes your testing system should have.
* See 25 real life screen shots taken from actual live websites.

This is a limited release preview edition, so it won’t be available for very long. I suggest you go and get it now!

1. New GPS for web design, tells you how to build your site

2. 3 Things that will improve your conversion rate

3. Time your tests to get best results

4. Why testing with Google Adwords yields flawed results

5. What is Taguchi Multivariate Testing

Look forward to more great posts in 2006. I hope to have at least one or two more “best of” type posts before the end of the year!

* The ad group that used dynamic keyword insertion recommended a minimum bid of $5.00!
* The ad group that used static keyword insertion recommended a minimum bid of $0.40.

Read the post here.

The best thing about testing and optimizing through October is that the results are likely to remain consistent through the rest of the year, yet it is before the bulk of holiday sales are made. What would a 20% higher sales conversion rate do for your 4th quarter numbers this year??

Things to test? Free shipping offers work particularly well. Try varying the threshold for qualifying for free shipping. How about free gift wrapping? I have not seen this used as a promotion myself, but if it came down to it, I am sure it would sway many folks.

You can view it here: http://www.vertster.com/newsletter/newsletter.asp?newsletterID=18

The book starts with a look at the size of the PPC advertising market, and how Google got into it. Next comes basic information on how to set up and manage your account. This is followed by the best part of the book in my opinion… the intermediate and advanced stategies. Things like how to write powerful copy, tracking users after they click, and increasing conversion rates. There is even a section on landing page fundamentals.

Order your copy today from Amazon.com and save 32%. This is a very smart purchase at only $16.99.

Get more information on our redesigned product pages at http://www.vertster.com/pricing.asp! Note that this offer is only good for new customers, subscribing for the first time.

The response form is always important, since it is the way conversions come in. For this test, we are actually looking at using the same form, in three different locations- in the right column, the left column, and underneath the text body. I have always believed that putting the form above the fold was important. On most landing pages I have designed, you can see this. The right column is a personal favorite of mine, using a 65/35 lineup.

Which form position do you think is best? I bet you never thought about testing it

I have always questioned the deliverability of our other provider StreamSend. So this is a test, and we’ll find out which works better or if it is a wash. Streamsend has prettier graphs, and their interface is quite a bit different from Vertical Response’s. Our list is squeaky clean double opt in, but I have seen lots of people subscribe several times in a row on Streamsend, apparently unable to get the confirmation email.

I will let you know my results in a few days.

1. Self Interest rules on the web: People want to know whats in it for THEM, and them alone. If your headline does not accurately predict what the visitor is interested in, AND convey it in a short compelling manner, you are losing visitors and money. Instead of: “This enterprise class system features 16 key functionalities,” try “Your organization can enjoy increased productivity and peace of mind.” Including the word YOU can make or break a headline, and an entire ad. We’ve seen it in numerous tests.

2. Be specific: Nothing sounds like canned boilerplate copy worse that general, non specific descriptions. Adding real numbers makes your copy much more beleivable. Instead of “Our system has many valuable features,” use “Our system has 97 great features that will save you at least 9 hours and 16 minutes each week.”

3. Use “How to”: There are few words which are as powerful as these in copywriting. The reason this works so well is that using “How to” to start off a headline puts you right into a self interest style of writing. People online are seeking information, or in other words… how to do something.

The esteemed John Caples, author of “Tested Advertising Methods,” estimates the headline makes 50%-75% of an ad. Must be why he spent and entire 5 chapters on headline writing and testing alone. If you don’t have this book on your shelf now, and you are serious about improving your online sales copy, I suggest you get it immediately. It’s probably the smartest fifteen bucks you could spend.

During the interview, Anne provides some highlights of the information in their new Landing Page Handbook, including the top mistakes people make on their landing pages.

View his blog entry now, with a link to the podcast!

In this case, the single offer paradigm may be bettered by using more than one conversion option. Create conversion options for each possible stage of the buying cycle.

Consider this example. Prospects the farthest out in the sales funnel may be averse to filling out a full registration form to get a whitepaper. For them, a simple opt-in newsletter may work better. But the mid-stage prospects are beyond the newsletter and will gladly give their information for the information rich document.

How to handle these very different prospects with one page? Put both options on the same landing page. How you count conversions here is important. You can measure each element separately, lump them together, or do both at the same time.

The other option is to set up a “two stage” conversion process. This basically presents the conversion options one after another. First get the user to sign up for one thing, then on the confirmation page, present them with the next. This can work very well because the person has already crossed a mental barrier against providing information after the first step.

Of course the question is, which offer to put first. And that is one that needs to be tested too!

By incorporating website addresses in a television spot, marketers can quickly guage an ad spot’s effectiveness. Companies like Cadillac are running ads and immediately correlating the numbers over to visits on the landing pages created for the TV campaign (see www.cadillacunder5.com for an example.) According to the Journal, Papa John’s is able to track online orders by the hour and can immediately judge a spots effectivess.

The part of the article I found the most compelling was this quote from John Freeland, managing partner at Accenture: “Do they focus on marketing as a management science, or do they treat it as an art form?” And of course the fact that all this stuff is measureable also means that it can be tested and optimized. I wonder if they are doing that too. They certainly should be.

For this example, lets look at a newsletter. It’s possible to greatly increase your newsletter subscriber list, with a few simple tweaks and optimizations. Imagine if you could generate 1000 new subscribers a month instead of the current 500. Is this realistic? Absolutely. Almost everyone can achieve these kind of results with a carefully administered test and optimization strategy.

What you do with these ezine subscribers is up to you, but most people use ezines to build their brand as well as promote products and services. Take a typical offer in the newsletter, which pushes people to a landing page on your site. Have you tested and optimized this offer? Imagine you have, and you increased your clickthrough rate from 2% to 4%. Again, totally realistic.

Once they get to your site, they hit a landing page. This is where things get fun- the money is actually made here. Optimization of the landing page has yielded a conversion increase from 2% to 3%.

Lets see what happens to the big picture:

WOW! We went from 4 to 13 sales, by optimizing three steps in the process.

Rather than jump on the bandwagon and report how great their Google research is, we’ve found some other nuggets on their site and blog. The best is probably this case study of an actual home page improvement project. Conversion was not their specific goal, but I would like to think that most of their findings could improve conversion rates, especially for internal marketing messages. People can’t click things they don’t even see, which after reviewing their research is quite a bit.

Keep an eye on these guys, their doing some really interesting stuff!

Sitepal avatars provide an animated flash character intended to “engage” the user. This being the case, Dr. Wilson used Vertster combined with Clicktracks to compare changes in site stickiness caused by the Sitepal avatar. View his full assessment and results.

With paid traffic, the best way to convert is often by sending the visitors directly to a targeted landing page, with few options other than the buy button.

Maximizing conversion results from organic search traffic is altogether different. It is difficult to write strong direct response copy, and optimize it for top search engine rankings. You just don’t have control of which of your pages will show up for a given search term.

The bottom line is, you can’t control where people will enter your site, and you have little control over what they will see. When working with organically search traffic, we move to using “internal site marketing tactics.”

Internal site marketing refers to elements of the site which drive people to act- such as promoting a whitepaper, prompting people to request more information, visit the forum, sign up for a trial, register for the newsletter, etc. How these are presented and promoted will have a huge impact on the actual benefit of organic traffic.

The best part about them, is that they can be tested and optimized! Although testing with “global” site traffic introduces noise, if you run your test out longer, you’ll still be able to get a pretty good indication on what works and what doesn’t.

This is definitely something that would be worth testing for 90% of people out there. In most cases, my guess is it would increase conversion rates for many.

Web site visitors tend to be very different on the weekends and evening, than mid day during the work week. I have seen a lot of activity and responses come from senior management of very large companies. Rarely during the day. Weekends are when these time-stressed executives get to surf.

Say you run a test starting on Monday. By Thursday, you could have a substantial amount of data, pointing to a clear “winner.” The inexperienced tester would begin celebrating prematurely. Along comes the weekend, and what you thought was the winner could end up falling behind.

How to avoid this? Always run your tests by the calendar week. This essentially removes the weekday/weekend fluctuations from affecting your end results.

How these different items are promoted on your site is something that should be tested and optimized. Whether your site uses clever banner ads, product images, or simple signup forms, getting the most out of them should be a top goal. This particular example is one that is currently in use on the Vertster Site. In the coming months, we plan to launch and optimize more internal marketing offers, promoting a free demo, webinars, and more.

Things we could and will test include: layout, headline, use of the word free, inclusion of an image of the item, or inclusion of a “lifestyle” image. There are obviously many more ideas you could use as well. But the point is, get testing!

Bloglines also allows you to export a list of all the blogs you read for inclusion on your website. This makes perfect sense. It would stand to reason that anyone reading my blog is likely interested in some, if not MOST of what I like to read.

You can view the list when you enter the blog section of the Conversion Rate Marketing and Split Testing Resources Conversion Rate Marketing Resource Center. This will be constantly updated with the best Conversion rate and marketing blogs out there. If you think I may have missed one, please send me an email at scott@vertster.com.

Unfortunately, a common tendency is to complete one test and call it “good.” This is a rookie mistake. A true optimization campaign is made up of a battery of tests, some designed to determine best overall creative direction, some to fine tune, and some to provide confirmation.

You should only rarely make conclusions based on a single test. Even if you have a clear winner, it is smart to run at least one identical confirmation test following your first one. This is particularly true if your activity is low (IE low confidence,) or if you know you have some uncontrollable noise factors.

To verify your own tests, feel free to use my free Google Adwords Statistical Validity Checker

Using Adwords as a “poor man’s” testing tool for landing pages has also been advocated by people who think they know more than they really do. In fact, if you peruse the marketing forums, you’ll often see this discussed, along with haphazard guesses at how much activity denotes statistically valid results.

When you use Google to split traffic between two or more landing pages, you are inadvertently adding a HUGE random noise factor. This WILL screw up your analysis and could cost you MASSIVE amounts of lost sales and advertising dollars. DON’T DO IT.

Basically it goes like this. Google counts a view every time someone clicks on your ad. Even if they come back 4 days in the next ten before buying (and coming back through the same paid link is common.) When this happens guess what it does to your apparent conversion rate. It goes down. In fact, for this guy that converted… he is only showing a 25% Conversion Rate (CR). The next guy buys on the first visit, but from one of your other landing pages. Well guess what, that one just scored a 100% CR for that customer. If this happens a few times, you’ll start thinking that this other creative is better. Bad Idea. Who KNOWS if it really is?

There is no way to be sure, and no amount of volume is going to provide enough confidence to make this test “trustworthy.” The only way to tell for sure is to only count the view once… On the first time a prospect is exposed to your offer. The vast majority of conversion tracking tools out there are flawed- counting all views against all conversions. So be careful. Testing incorrectly is worse than not testing at all- because it takes just as much work, and yields data which will never verify in the bank account.

So now what? How long are your test results good for? Well, you should plan on testing some new ideas every so often, because there is almost always room to improve. The answer to this question is highly dependent on your market and your offer. But unless your response rates suddenly take a nosedive, you probably don’t have too much to worry about.

Certainly, many things happen in the marketplace that can effect your response, both positively and negatively. If there is a shift in the market landscape that will stick around, you should definitely plan on doing some more testing. Some examples of this could be new laws passed which affect your product, major lawsuits, or natural disasters. If everything remains “status quo,” there is a good chance your response will remain relatively static.

You see them on on the web in all shapes and sizes, not to mention what words are used on them. Non-marketing type web designers usually stick with the default text- “Submit.” Does this have an effect on peoples likelihood of actually pressing them? Only testing will tell you whether it makes a difference or not.

Things you could test:

• Button Text… “Submit” vs. “Subscribe” for example
• HTML button vs. Graphic Image
• Where on the page the button appears
• Action vs. Passive text
• and many other ideas

The success metric:
Determining the best submit button takes time, and the metric you will want to watch to determine success or failure is pretty obvious- number of forms submitted!

• introduce your customers to your offer
• intrigue them
• turn them off
• qualify them
• get them excited
• make them sad
• create desire
• create demand
• hook your customers
• scare your customers
• and many more things
  

What to test

What your headline says flows into the rest of the offer. But without a captivating headline, your chances of actually getting your prospects to read the rest of your offer diminish drastically. Things you could test?

• Word Length
• Style
• Aggressiveness
• Font Size
• Appeal Type
• Color
• Font Face
• and more…

What about your success metric?

A successful headline will not automatically close the deal. Your indication of success might be sales revenue, or leads generated. Or maybe not. Maybe just time spent on the site. Are you “captivating” the audience? That might be the best measure.

Choosing when to test is important… After answering how and what to test, the question of when is the next most important. Seasonal fluctuations surrounding the holidays can have a major influence on web users.

Many sites note a significant drop in traffic during November and December, while retail sites often see the opposite.

Most people are just too busy to bother with the internet. Think about it, they are out shopping, standing in line, fighting traffic, travelling, spending time family and old friends, etc. Not sitting at the computer.Even if they do spot something online they’re interested in, chances are quite good that they will put off their action until “after the holidays.”

For you retailers, folks are much quicker to buy and make decisions during the “mad dash” to buy presents prior to the holidays. This is hardly an accurate representation of what your customers will be like during the other 300 days of the year.

These are a few of the reasons why testing during the holiday months is tough, and make it hard to get an accurate picture of what motivates your customers. You might find that things which work during this period actually hurt you during the rest of the year.

The good news is, that it’s now January 2005, and things are rapidly getting back to normal. This means you have 5 solid months to test until the next problem area comes along… summer!