Add3.com

SMX West – Landing Pages & Multivariate Testing – Amplify Overview

I had the opportunity to go to the Search Marketing Expo (SMX West) in Santa Clara, California last week. Here are some notes I jotted down during the session: Landing Pages & Multivariate Testing. For further reading, I’ve found another resource covering this topic here.

SMX West: Landing Pages & Multivariate Testing

Speakers:
Jon Diorio, Group Product Manager, Google AdWords
Jonathan Mendez, Chief Strategy Officer, OTTO Digital
Sandra Niehaus, VP User Experience & Creative Director, Closed Loop Marketing
Seth Rosenblatt, Vice President, Business Development & Marketing, Optimost solutions at Interwoven
Gregg Makuch, Vice President for Wide Mile

Jonathan was the first speaker. “Every user has a goal.” You need to understand the visitors coming to your site. “Every page is a landing page” and each page should have a goal. Two types of landing pages:

1) Reference pages – seeking information
2) Transactional pages – seeking transaction

Three different testing styles:

1) A/B testing. It’s easy, big lifts & quick results
2) Multivariate testing. Incremental lift, lots of intelligence from it. And you can “fail faster”, which is a good thing!
3) Targeted content. Recognition, reinforcement & relevance.

Launched into some case studies. Tested five different “welcome” pages. %142 conversion increase in initial test. This was essentially A/B testing. Then did a more in-depth test and did a multivariate test.

Test navigation. Sometimes having some navigation options can lead to good things. So give it a shot.

Sandra was the next speaker. She’s talking about how to “get more” out of MVT (multivariate testing). Her powerpoint had a lot of images & pictures and such – you could tell that she spent some time on setting it up. Conversion rates shouldn’t be the end goal – it should be qualified leads. For instance, you may have a result in a decreased conversion rate by adding more to the form fillout for more qualified leads but you get more qualified leads.

Landing pages shouldn’t be an existing page shoveled in if it’s a problematic page; why not start from scratch and redesign it to something more attractive?

Look at who is inputting your forms! If they’re just not the right audience, create your form so only the qualified individuals will want to fill it out.

Elements of the page that are important to test:

1) Headline
2) Call to action
3) Buttons
4) Hero image
5) Bulleted benefits list
6) Copy
7) Tables, charts, etc.
8) Form

Remember to test entirely different elements when doing a MVT test. Images & text changes that are very different are important. Also, only test one difference at a time. Set up a test, test one thing, find the clear result, then test your other idea.

Seth was the next speaker on the panel. “Multivariate testing is the SAME as search engine optimization.” Customer value formula: traffic x conversions x lifetime value. If you maximize that formula, the more leads you’ll get. At the end of the day, it’s value over $$$s spent. He was pretty fired up about how search marketing is a great industry and all but lading page optimization is just as much as a conversion rate increaser as SEO is.

There isn’t a “magic formula” or way to do landing page optimization; each test will produce different results by testing different things.

“Sometimes little changes can make quite a difference.” Make sure to test a “Satisfaction Guarantee” message in a MVT test. Also look at the page layout – moving around the order of the copy blocks can change rates dramatically. Forms are very important. Do you REALLY want a fax number? If it helps when getting a qualified lead, then go for it. But think if you really need that required field.

Jon was the last speaker on the panel. Being the Product Manager for Google AdWords, he went into how PPC can help landing page optimization. Improvement of landing page optimization & conversions = improved performance for your PPC campaigns.

Simple process:

1) Identity impactful page, sections, elements
2) Develop simple hypothesis for each element
3) Create 1 – 2 different versions of the page
4) Repeat as necessary

Case study into homepage for Picasa and how they changed three different things to improve downloads by 30%.

A quick QnA followed.

About Christian Bullock