Get Free Quote
« The Personalized... | Blog home | Will The Social Graph... »
February 27, 2008

Landing Pages & Multivariate Testing

Print Friendly

Good morning, happy people! My tummy hurts. I think its cause SMX only offered sugar-covered carbs for breakfast this morning. Who does a girl have to kill to get a bagel? Any maybe some apple juice.

After a quick schwag run, I’m seated with my conference buddy Ross Dunn at the Landing Pages & Multivariate Testing session. Gord Hotchkiss (hee, I just typed God) from Enquiro is here moderating with speakers Jonathan Mendez (RAMP Digital), Sandra Niehaus (Closed Loop Marketing), and Seth Rosenblatt (Interwoven). Gregg Makuch (Widemile) and Jon Diorio (Google) will pitch in for the Q&A. Okay, then!

Up first is Jonathan Mendez to tell us he’s launching his new company RAMP Digital today. He probably doesn’t mean today. More like recently. Right?

Jonathan says that everything starts with understanding your users. What are their goals and intentions? Search is great because it forces users to put their goal into a query field and let us know what it is. If our ads are working properly we’re messaging to that goal and getting people interested. We’re setting expectations.

Every page is a landing page because every page has a specific goal and intention. A great example of this would be the Google home page. He calls it the greatest landing page in history.

There are two kinds of landing pages: Reference Landing pages (looking for information) and Transactional Landing pages (buying). They can be either static (HTML), dynamic (JavaScript) or application-based (API).

Landing Page Optimization

  • A/B Testing: Incredibly important. It’s the first step in the optimization process. It’s the easiest to get going, offers big lifts and quick results.
  • Multivariate Testing: Provides incredible intelligence to you. Allows you to understand the different factors of influence and how they related to your conversion rate. Get incremental lifts and allows you to fail faster. Test lots of things at one time and then move on.
  • Targeting Content Delivery: The world is not one size fits all.

MVT Experimental Designs: There are two kinds – Fractional Factorial and Full Factorial.

Fractional Factorial: You are taking a fraction of all the different things on the page and testing them. Good part is that you need less data to get results. Cons are that you are not going to look at every interaction on the page since you’re only looking at part of the page.

Full Factorial: You’re looking at every single factor. Advantage is that there are no interactions that are missed. The negative is you need a lot of data to have confidence in the results.

A lot of people say not to put navigation on the landing page. He’s done tests to prove that that’s not true. You want to make the page more usable for a searcher; sometimes that includes adding navigation.

From here Jonathan runs through some case studies at lightening fast speed.

Case Study: Magazines.com

Tested left hand navigation being present or not present, as well as headlines – static image or dynamic rollover.

As they created the test they created profile segments. Putting visitors into buckets based on how often they visit, what they were looking at, etc. Showed them the most important factors of influence for each bucket segment.

You can go to his blog to read 7 Rules for Landing Page Optimization on his blog.

Next up is Sandra Niehaus.

Forty-nine percent of marketers plan to run multivariate tests in 2008. This shows a growing public awareness and enthusiasm. Multivariate testing has so many possibilities in how it can be used to improve your bottom line. However, a lot of us labor in time restraints.

Start with a plan

It’s tempting to just jump in and try everything, but a considered plan will give you better results. Know your business goals for the test. Look at the elements on your page and what you should test.

Have a Goal

Every company h as a unique combination of audience, product/service, approach, etc. Focusing on the business goal is sometimes more important that focusing on the conversion rate itself. Conversion rates are how we measure success, but they can get distracting. Don’t use that as your sole guide for whether or not you’re being successful. Just because you’re performing at the industry standard or better, doesn’t mean you can retire. And you’re not failing just because your conversion rate is lower than average.

Business Goal #1: Increase the number of highly qualified leads.

That might mean adding some new field to your landing page. This may lower your conversion rate, but you’re getting a higher percentage of qualified leads. In that sense, you’ve achieved your business goal.

Business Goal #2: Increase Revenue

Look at conversion rates and how that tracks across all types of your audience. You don’t want to optimize for just one segment. You want to hit everybody.

Page Concerns:

Ask yourself if the page is worth testing. Can it be saved? Is it just unmotivated? Are there code errors? If so, why not start from a better place? Use a revamped template that will increase conversions without sacrificing quality.

Gather your inputs: Who are your visitors? What are your demographics? Where is your traffic coming from? Understand how traffic source affects the conversion rate. What’s your baseline? Where are you starting from?

Once you have your data, apply best practices. Remove unneeded items, customize button design and test, improve relevance to visitors, summarize benefits, include credibility builders, use clear and motivating calls to actions, etc. You want to validate your designs early on and identify areas that may need more clarity. Work with designers who "get" it.

Elements

Does every element matter? Well, how much time do you have?
Elements that are more impactful than others are the headline, calls to action, buttons, "hero" image, bulleted benefits list, copy, tables, charts, forms, etc. These are all more important than things like dividing lines or background colors.

GIGO: Garbage In = Garbage Out

You want to make sure the variations that you create are of high quality and are conceptually different. When testing, test one difference at a time. You’ll get much clearer results.

Seth Rosenblatt is next on the list. He gives a lot of background on his company Interwoven.

Multivariate optimization IS search engine optimization. It’s the same thing. They’re connected because at the end of the day you shouldn’t care about driving relationships or engagement or traffic. You just care about the money. Everything else is just a technique. You’re putting money into an effort and sales dollars are coming out the other end.

Customer Value Formula: Marketers become heroes by maximizing lifetime customer value for the lowest marketing investment.

Traffic x Conversion x Lifetime value

Traffic = How many visitors you get compared to the money you’ve spent
Conversions = The number of customers you get for every customer you get
Lifetime value = What’s the value of each one of those customers?
Marketing = The value per dollar.

There are multiple forms of conversions. You can’t think of it singularly. If you’re a retail company you want to maximize revenue. That may mean increasing average cart size or profit or a million other things. You have to look at the goal of each page and then optimize it accordingly.

Best Practices and Pitfalls

The best practice is testing. For all the examples he’s about to give, there’s an exception to every one. Your business is different from everyone else’s.

  • Copy: Amount of copy. Are benefits really benefits? Headline/Header Construction. To link or not to link.
  • Layout: Ordering of copy.
  • Forms: Why is each field there? Why is there room for a fax number? How many people have you ever sent a fax?

Seth is over time and the mean podium timer is beeping at him. We’re all afraid he’s in trouble.

Jon Diorio

Optimizing your landing pages lowers your cost per sale. Anything that you do to improve the user experience on your landing pages will also have positive affects on your PPC campaigns. It’s not just about sales, there’s also the health of your site and your PPC campaigns to think about.

The Process

  • Identify impactful page sections/elements: Things like the headline, picture, call to action, etc. Explain the benefits, not the features. Links receive more clicks than buttons. "Free" is better than "a trial offer".
  • Develop simple hypotheses for each statement
  • Test a few variations
  • Repeat as necessary

Greg Makuch isn’t going to present, just comment on what everyone else has said. Sweet. Lunch is in 12 minutes and I’m hungry.

You can be a hero with your client if you’re optimizing their landing pages. Not many people are doing it and the results are dramatic. Technology is only part of it. The more important page is the methodology and expertise you apply to the testing process. There is a method to the madness. If a test is poorly designed, you’re not going to get the type of results you could from a more well-thought out approach.

And with that it’s time for lunch! Huzzah for food!

Print Friendly




Comments are closed.



Learn SEO
Content Marketing Book
Free Executives Guide To SEO
By continuing to use the site, you agree to the use of cookies. AcceptDo Not Accept
css.php

Curated By Logo