January 09, 2006
How To
SUMMARY:
Whether you're in a highly competitive search marketing niche with fierce bidding for your best keywords, or you'd just like to improve PPC campaign results, this story's for you. Features six landing page a/b test lessons, plus two ways to improve results on online registration forms:
|
|
Paid search marketing ads have become almost insanely competitive in the home mortgage industry.
"We have the highest CPC, higher than retail, higher than online banking, higher than other mainstream verticals. We feel the impact on ROI very quickly," explains Imran Kahn, E-LOAN Customer Acquisition Marketing Director.
How does a company in his sector compete? Kahn says relying on improved bid management won't cut it anymore. These days the search engine wars have moved to landing page optimization. Here's what he's learned from three testing tactics:
Tactic #1. Landing page testing -- six lessons
"How do you go about building good landing pages? Everyone has their own idea," he says. To hone the best landing page, his team begins with the mentality that "your idea could be as good as my idea."
With that in mind, they began doing A/B testing, trying out one variable at a time.
Kahn tested landing pages on keywords that saw a high number of searches but that had conversion problems. By testing and optimizing elements such as button color, image (happy couples vs. a picture of a house), layout, and headline, Kahn's team saw conversions jump as high as 165%.
"So we totally fell in love with that, and started doing more of it."
A/B testing has its limits. "We were dependent upon IT, which is challenging. Also, it was not really a scaleable solution because it was one variable at a time."
Now, Kahn's team has begun using multivariate testing, which allows him to test multiple variables at once and learn, with high accuracy, which elements and combinations of elements are most successful at converting visitors.
Six lessons learned from landing page tests:
o Fewer options is better
"The more options and the more decisions the user has to make, the higher the probability that the user will drop out," Kahn says.
People coming from a search engine are very impatient, and they know they have other options. "They can click the back button and go to the next search result." Don't make them make too many decisions right on the landing page.
o Grab their attention in first 30 seconds
"Draw them in," Kahn advises. Make it clear what you're saying in the top few lines, especially the headline. "That's probably 80 percent of the value," he says. "The majority of users don't get beyond that."
o Headline should reflect ad copy
o Don't be clever
Catchy headlines don't work. State it clearly. If you want a user to open a checking account, say, "Open a checking account."
o Offline media doesn't work very well Don't try to use your slogan or other offline approaches.
o Reassure them The landing page should help with the transition from ad -- where they know exactly what they're getting because they asked for it via keyword -- to site, where they're not sure you're the best place to be. Assure them they're at the right place and make it clear what they should do next.
Tactic #2. Application form page testing -- two lessons
Traffic from E-LOAN search ads is sent to landing pages and then to internal pages which are generally application forms for loans -- never the home page.
To ease the transition from landing page to a page deep within the site, "we do some hand-holding," Kahn says. "If they're talking about a loan in California, we tell them even on the next page that yes, this is the place to apply for a loan in California."
The content on internal pages is dynamic, with areas that can have rotating content depending on the source or keyword from the search. That way, "you don't have to build a specific page for everyone," Kahn says.
Two key lessons learned:
o Start application forms with easy questions
"When we asked them tough questions at the beginning we [saw] that using that form on our landing page was not working well for us." Kahn tested an easier, more friendly form. It wasn't necessarily shorter, but the questions drew users into the process with simple questions, such as:
--Where are you looking to purchase a home?
--What is your credit score like (good, excellent, etc.)?
o Allow users to save information on a form
E-LOAN uses a long application form, and it can be "very daunting. We understand that the application is long and they might want to finish the rest of it at some other time."
Now, the form offers the option, at the end of every page, to go on to the next page or to save it. Not many companies offer the option to save a form like that. "In our industry, there are very few players who actually use the longer applications," Kahn explains. "They use short lead forms and try to use the phone to get more info."
By allowing users to save the form they have increased conversions significantly.
Tactic #3. Use focus groups to discover and prioritize test ideas
To decide what to test, Kahn has used focus groups of 15 or so people who give their feedback. "Based on that, we take some of their ideas and implement them, see which one does better in terms of numbers."
Two lessons learned:
o Sometimes the idea that looks like it won't work is the one that actually works
o "Analyst" marketers are hard to get
It's difficult to find marketers with a background in data analysis. "These are people who have a task of not just analyzing but managing campaigns."
Kahn finds it more important to find people with a background in analytics. "I made an offer to someone who has a Ph.D. in economics," he says. "Every consumer is like a data point. When you have enough data points you can decide what do."
Useful links related to this article:
Offermatica, the multivariate testing and optimization company eLoan uses: http://offermatica.com/
E-LOAN: http://www.eloan.com