September 27, 2004
Case Study

Ancestry.com Reveals Results of Landing Pages Tests That Raised Revenues by $500,000

SUMMARY: How can you turn visitors into trial subscription-takers? The marketing team at Ancestry.com has been testing new landing pages for their affiliate program for the past eight months -- and today they revealed the results exclusively to us. Although this relates to affiliate campaigns, we feel the winning landing page style might work extremely well for search campaigns as well. Find out why... Yes, of course we included creative samples for you from the tests so you can see what worked and share it with your Web design team:
CHALLENGE

Thanks to four years of aggressive recruitment efforts, Ancestry.com has more than 30,000 affiliates (marketing partners who are paid commission on sales) on its rolls.

As other subscription sites and eretailers have found, only about 5% of these affiliates focus hard enough on their efforts to bring in hundreds or at least tens of thousands of subscription dollars each. However, the rest bring in enough onesy-twosy sales over the course of each month that it's worth it to keep them on the rolls.

Drew Izzo, Director of Online Marketing, explains, "Typically they're a higher qualified subscriber because of the nature of affiliate marketing. In order to end up on our landing page, they've gone through an experience already, perhaps starting with search, and drilling down on the affiliate's site before ultimately hitting Ancestry."

Thanks to this high conversion rate, Ancestry.com's affiliate program pulls in tens of millions in sales each year.

In the never-ending quest to grow these sales, Izzo's team changed their campaign technology last fall. "We started using our ad server to serve the entire landing page as an ad, so we could rotate and optimize the landing page."

This meant they didn't have to get affiliates to manually change their landing page links every time there was a new page. In effect, the landing page was syndicated through the serving software and changeable in a blink of an eye.

But getting every single affiliate link switched to the new system proved impossible.

"We realized there was a substantial chunk of affiliate pages that had the old hard-coded ads just sitting there on their sites. Affiliates didn't even know where all the links were that were still driving impressions for us. It was a huge volume of traffic, but it was hard to optimize because of the hard-coded link."

CAMPAIGN

Izzo's team reviewed the 'classic' landing page that most old links directed to. It was averaging a 2.47% click conversion rate with a conversion defined as a prospect entering their contact data and credit card info to start a free trial. (Link to sample below.)

Next, they created a series of about a half dozen tests based on what was working in other campaigns (link to samples below)...

Test A: Who Are Your Ancestors?

This page enticed viewers with carefully written copy (lots of "you") in fairly large type on the main benefit of subscribing -- discovering more about their family.

They added a 468x60 Flash banner, with the same message, that had performed well in separate tests to the top of the page to lend a little graphic excitement. And there was a free trial offer at the bottom, part of which was below the fold on some visitors' monitors so you had to scroll to submit.

On the first page visitors were asked to fill in nothing more than firstname, lastname, and email. (Note: If the visitor didn't convert into also entering the credit card required for the trial on subsequent pages, he or she was then sent an automated series of conversion-enticement emails.)

To encourage conversions, the team put a paragraph of reassuring language about privacy just under the email box.

Test B: Explore Your Family Tree

Are words or images more powerful for converting visitors to trial takers? In this test, the team lowered the point size of the copy and bumped up the graphics to a larger Flash box.

Again the free trial offer only requested firstname, lastname, and email on this first page, reserving credit card questions for a page further in the process. Again there was reassuring language about email privacy. Again it was partly below the fold on some viewers' monitors.

The team was pretty sure this landing page would be a clear winner because similar creative had worked very well for them in other online ad campaigns.

Test C: Find Your Ancestors at Ancestry.com!

This was the down-and-dirty page. It had no Flash and no old-time photos of ancestors. There wasn't even any language about a free trial.

Plus, breaking the rules of landing page best practices, there wasn't one clear main click path. Instead, visitors had their choice of three different main buttons (each of ultimately led them into a trial offer on subsequent pages):

- Begin Today: "start your family tree today!" - Quick search: look up names in Ancestry's library - Search old US Census records

None of the three click options asked for an email address up front. The page was designed in two columns so everything was above the fold and required no scrolling for most visitors

RESULTS

The winning landing page is currently averaging a 3.37 conversion rate - almost an entire percentage point higher than the old standby control. "That increase equals about a half a million bucks a year in incremental sales from those old outdated affiliate links," notes Izzo. "They are still performing!"

Test C was the winner. Why? Izzo thinks the secret may lie in the fact that most affiliate link visitors are in the process of researching when they click. Interrupting that search mentality with promotional copy for a free trial may be the wrong thing to do.

In fact, Test A pulled a miserable 1.13% conversion rate, and Test B pulled a 1.78%. "This was a shock," notes Izzo.

We also believe that asking for emails too early in the process can slow results in these please-god-don't-add-me-to-more-lists days. You have to involve visitors more in your brand now before sicc-ing an email request on them, no matter how reassuring your privacy language is.

And of course, anything you can do to move your landing page in its entirety above the fold generally helps responses.

Next, Izzo's team has to test whether the winning landing page would work as well with all Ancestry.com standard site navigation links stripped off. We look forward to results.

Useful links related to this story:

Samples of the original landing page as well as the tests mentioned in this Case Study: http://www.marketingsherpa.com/ancestry/ad.html

Atlas DMT, the ad serving software Ancestry.com uses to serve landing pages: http://www.atlasdmt.com

Ancestry.com's affiliate program info: http://www.ancestry.com/home/partner/main.htm

MyFamily.com Inc, Ancestry.com's parent company http://www.myfamilyinc.com


Improve Your Marketing

Join our thousands of weekly case study readers.

Enter your email below to receive MarketingSherpa news, updates, and promotions:

Note: Already a subscriber? Want to add a subscription?
Click Here to Manage Subscriptions