Landing Page Testing & Tuning - SES London 2009
How to increase the conversion rate on your landing pages? Panellists Andrew Goodman, Anders Hjorth, John Marshall and Jon Myers discussed this question in the session ‘Landing page testing & tuning’ on day 3 of SES London.
Andrew Goodman, Principal of Page Zero Media moderated this session.
Jon Myers, Head of Search for MediaVest, is recognised for his unparalleled experience and knowledge within search engine marketing in the U.K. Myers started his session with a short introduction about the importance of a good landing page. According to several studies, we have 13 seconds to convince visitors there're on the right site.
Myers explained there are 2 key components:
Advertising relevancy: A good landing page improves the quality scores.
2. Sales optimised
Basic optimisation principles must be followed:
- Deep linking
- Offer driven; Make the reason to stay compelling.
- Easy navigation; Can I get to where I need to be?
- Sales focus; Where do I buy?
According to Myers, we certainly may not forget to test outside of the tool:
- Perform usability testing
- Try hotspot testing (like crazyegg.com)
- Speak to the user
At the end of his session, Myers took a leap into the the Future. What's the new Landing Page? GOOGLE! Using Google as your landing page, can really increase the traffic to your site.
The second speaker was Anders Hjorth. He was responsible for the service offering of Relevant Traffic and managing a team of webmarketing experts across Europe. Recently he joined GroupM as the CEO of Outrider EMEA. Hjorth shared some interesting opinons and tips.
- Match your landing page to your site.
- Consistency and repetition is important.
Design principles applied to paid search:
- The best scenario: 1 landing page for 1 keyword.
Costs and benefits:
- The key to succes is Enough volume and Homegeneous traffic sources.
Detailed landing page design steps:
1. Define measureable and actionable objectives
2. Define target persona (one target one page)
3. Graphic elements, lay-out and action-mode
5. Build-outs and site integration
6. Test scenarios
To conclude the session, Hjorth gave us 6 interesting tips:
1. Structure: remove most on navigation
2. Personalize: repeat elements from refering page
3. Contents: simplify and keep only essential information
4. Image: work on the hero image
5. Usability: try to avoid scroll
6. Testing: A/B split testing or multivariate
Last speaker at this session was John Marshall, CTO for Market Motive. Marchall is a Netscape alumnus and a seasoned technologist, with roots in programming high-speed algorithms and statistical analysis. He shared us some general insigths about testing:
AB Testing (2 or more versions of a page)
Marshall started with a brief introduction how to perform AB testing with Google website optimizer:
- Create addition versions of your page at distinct URL's
> randomises users and sets cookie
> redirects to page B, C, D, etc.
Multivariate testing (MVT, substitute elements within the same page)
The more combinations you test, the more information you will recieve. This could help to make better decisions.
How does MVT works?
- Java executes within the page on the client (no server coding needed)
- Each page has unique script that identifies the elements to be placed
According to Marshall, we must be aware of some common MVT problems:
- MVT is NOT necessarily better than AB. In most cases it's always better to start with AB testing.
- The code can be fragile: test your tests.
- Beware testing products who talk about Taguchi, secret sauce, etc.
- Knowing what to test is the key!
All our stories on Search Engine Strategies London: