RSS Ray interviews online testing, conversions, and marketing analytics guru Bryan Eisenberg about his new book “Always Be Testing: The Complete Guide to Google Website Optimizer.”

RSS Ray recently sat down with past show guest and online testing and conversion guru Bryan Eisenberg to discuss theimportance of website testing, Google Website Optimizer, and Bryan’s new book: Always Be Testing: The Complete Guide to Google Website Optimizer.

The interview provided some fantastic advice for anyone interested in getting started with online testing or improving their existing campaign. Take a look:

When you refer to testing, what do you mean and why is it so important?

What we mean is ‘trying to improve’ — taking something already existing on the site and making educated, thoughtful changes to it for the purpose of improvement: conversion rate, gross sales, sign-ups, etc., whatever the business goals might be.

What other kinds of business decisions does testing help us make?

Great questions. You know at first blush one might think the only stuff to test is what I just mentioned above: conversion rate or gross sales — whatever the company wants to achieve. But that really speaks to the late-stage goals of the visitor, which often overlap what the business wants to achieve.  

But if you think about it, once you get some optimization testing going at that point, a smart business also starts thinking of other ways to improve interaction with the customer, perhaps earlier in the buying process — how do we supply the potential customer with better info while she’s still “just looking” — and how to we optimize the post-sale process. Both of these impact future sales, looking beyond just this quarter’s profits, and that’s where true long-term growth comes from.

Is testing something only big companies with huge marketing budgets can do?

No, not at all. In fact, it’s often the smaller companies that test more efficiently.  Big Companies are often challenged by their very size, and in the same way that a small improvement in a key KPI — say, conversion rate — can have a huge effect at big companies in an absolute dollar amount terms, it also means that much more pressure to test well, which in turn can lead to fear of failure responses by the team involved. In smaller companies they are often a little more aggressive about getting it right as soon as possible exactly because the risk of not testing is just as great as testing in efficiently.

I often think of this analogy:  the big company is a like a passenger airliner: it carries a lot, somewhat comfortably, but all that bulk means it’s slow to turn in new directions.  Smaller companies are often more akin to fighter jets. They’re nimble and fast, very light but don’t have a huge range and aren’t built for comfort.

What kinds of things should we be testing?  What parts get us the quickest return for the least amount invested?

Well you’ve touched on Part 2 of our book {chuckles}. There can be so much to test that you end up seizing up in analysis paralysis. We’ve devoted the entire second part of the book to giving you a massive set of ideas to draw from, and we organize based on what context you’re testing in.  Is your business retail-based and you’re looking for ideas to improve product detail pages? We got a bunch of tests for that. Are you a lead-gen sort of site and you are looking to impact landing pages? We got a bunch of tests for that.

Here’s something to keep in mind: don’t always shoot for ‘quickest return for the least amount invested’.  If you knew ahead of time what would get you that, you wouldn’t even test, you’d just go for it. Instead, think of getting “quickest improvement over what I’m currently doing for the least amount invested.” That gives you room for mistakes — which are going to happen no matter what you do — while still improving.

How does someone get started setting up a testing program?

One of the reasons we focused on Google Website Optimizer in the book, is because it’s free. Sure, there are very good analytics platforms out there you can buy, but that’s an issue for marrying testing and measurement. A really good testing idea will improve your business whether you’re using a free or a paid analytics platform.

And, a lot of folks’ objection to testing is that they don’t have the money for testing — as GWO is free, that pretty much kills that excuse!

How do we know that we are actually getting statistically valid information from which to make business decisions?

Fortunately, the statistics for doing that are built right in to testing and analytics platforms. Without going into the statistics — and let’s face it, if you need a Masters in Stats to use a testing platform, how often will you actually use it? — but GWO does all that for you behind the scenes and gives you various levels of statistical confidence that what you’re testing is having an impact or not.  IN this political year, it’s like watching a bunch of polls. One given poll may return Candidate A in a lead over Candidate B, but it might be within the margin of error. And that’s one poll. But if you see 5-10-20 polls all showing Candidate A ahead, you start to get the idea that as close as the election might be, Candidate A likely has an edge.

How often should you test and when do you know you are done?

We try to tell clients to, well, “Always Be Testing”.  Always have a 3-6 month plan for testing and continuously be adding new tests as old ones are completed.  As for how long to run the test and how to know when you’re done, we think the minimum is about 2 weeks, and that only if you’re getting really high enough amounts of traffic to get actionable results. But if the test is not getting a lot of traffic, or if what you’re testing just doesn’t seem to have a candidate that is pulling ahead, we advise stopping the test and moving on. After all, the point of a test is to improve or not, not just to stand still.  If you keep running a test that doesn’t seem to be moving the needle, it’s better to just go on to the next test. In that sense, you’d almost prefer a test that “fails” (that is, your original set-up does better than the proposed changes) over one that doesn’t move the needle, because the point is to learn what leads to improvement generally.

Is testing something most companies can do on their own or should they look for outside help?

Yes companies can do this on their own, of course, if there is what we call a “culture of testing” at the company. Often though, it’s good to have an outside ally, someone who can give you ideas for testing, someone to bounce good (and bad) ideas off of, and someone to make sure you’re not drinking your own kool-aid.  That’s something we do a lot of at FutureNow.

Tell us why you wrote your new book “Always Be Testing: The Complete Guide to Google Website Optimizer.”

My co-author John and I found most marketers weren’t testing very effectively.  Most companies (about 77% by a recent report) do absolutely NO testing.   Amazing, huh?  But those that do test, often they start off well but then get side-lined several months in because they need help coming up with a plan for continuous testing.   We wanted to show with this book how that can be done

Who should be reading the book and what will they gain from it?

We positioned this as a weekend read for marketers and business people who want to get into testing without getting caught up in a technology barrier. Let’s face it: Google is a brilliant company, but not known for writing well.

A reader will finish the book ready to start testing, for free with GWO if that is what she wants to use, and can immediately start testing ideas that have been mulling around in her head for weeks or months. And don’t let your tech team talk down to you about testing — it’s likely after reading this book that you’ll know more about testing conceptually than anyone else at your company.

Tell us about Google Website Optimizer.  Why did you focus on this testing platform from Google rather than others?

As I mentioned earlier, it was the price!  GWO is a great leveler, because it means you don’t have to be spending $10k/month just to get started testing. That’s really it, the point is to get started and “free” is a great motivator in that regard.  We like stuff for free, we think you should too.  In fact, with the $25 AdWords coupon inside each of our books, our book might be thought of as “free” too, although of course we at FutureNow would say everything in there is invaluable.

How can people get it?

At your local bookstore or at any of the online merchants.  Amazon.com comes to mind.

Is it hard to implement and what costs and other things can they expect when installing it?

NO, there’s nothing even to install.  GWO is a hosted service to you, just sign up on the Google site and you can start testing right away. If you have access to your web site’s pages you can make the extremely minor changes right away with some simple copy and pasting, or the online tool will send the changes to make directly to your tech team.

What are some of your favorite features?

What I like best is that instead of waiting for the test to “finish” before reporting results, GWO reports the data as of the date you run the report — so you can get an idea of what confidence you can have in the results if you stopped the test right at that moment, and therefore you know what action you can should (or should not) take at that moment.

What other things will people learn from your new book “Always Be Testing: The Complete Guide to Google Website Optimizer?”

We also spend some time talking about testing in general. What it means to test, how it’s important not to start with the conclusion you want but instead to watch what the data is telling you, and to always be thinking of the results not as an isolated event but rather as one in a series of tests all marching towards improvement for your company. Most importantly they’ll learn the 30 factors we identify as impacting conversion and what they can do improve their results.

Can the principles discussed in the book be applied to any testing situation outside the Google Optimizer platform?

Yes absolutely. We use several examples in the book that can be easily implemented with other platforms — you just have to be familiar with your other platform particulars to get it done. But just as valuable are the business examples in the book which illustrate the more important concepts of testing and creating that culture of testing I mentioned earlier — those all go much beyond any simple “platform” argument.

Some companies take the perspective that better initial planning could minimize or eliminate the need for testing.  What are your thoughts?

We’re all for initial planning. We’d just like to see more of it. Our experiment has been that a company that gets hung up on initial planning never gets around to actual testing. They are so concerned with a theoretical “perfect” that they never get a simple test done “good enough”.    Instead we find companies that spend some amount of time planning ahead for testing but then jumping into actual testing sooner rather than later, they end up learning tons of new information that leads them to more insightful testing sooner.  The adage, “no war plan survives first contact with the enemy” comes to mind.  We want to get to testing  – even simple testing — ASAP exactly because we don’t know what new roads those initial tests will open up for us.

In addition to writing and speaking can companies hire you to help them improve conversions and profits?

Yes, in fact that’s the majority of what we do. We help companies optimize their marketing efforts online and we do so through continuous and dynamic testing.  We put a lot of good ideas in the book so companies could get started but obviously we’ve also got a few smart ideas always up our sleeves.

How do they do reach you and stay current with your latest happenings?

The web is always the best way:  www.futurenowinc.com, if you want to contact us to talk about doing some work together, we have a contact form on every page.  By the way, we are also the authors of one of the top marketing optimization blogs in the world, www.GrokDotCom.com  so readers can always hear new and interesting ideas for us there, too.

You’re one of the co-founders of The Web Analytics Association with Jim Sterne.  Would Jim look better with a toupee and a complete shave?

Well that should be tested, no?  Maybe he can have the rug on one side and shaved smooth on the other.  I would want to hear him with an English accent saying “Engage, Number One!”.  We’ll have to see if we can convince him to try it at the upcoming eMetrics Marketing Optimization Summit in Washington, DC in October.

Bryan Eisenberg is recognized worldwide as a leading expert on online marketing and marketing analytics. One of the co-founders and Executive VP of FutureNow, Inc., Bryan has been helping companies realize that to maximize results it is essential to incorporate expert persuasion techniques and a deep understanding of customer behavior into all marketing efforts.