Ancestry.com is an online only business that depends heavily on visitors finding and staying on its website long enough to buy into its services. Getting people to sign up – either for a 14-day trial or a monthly subscription – is the ultimate goal for the site’s conversion team, and this past year, it deployed several successful testing strategies to optimize its efforts. As a result, Ancestry.com posted a 20% lift in site conversions and significantly reduced the bounce rates on many of its pages.
We caught up with Ancestry.com’s senior manager of product marketing Emily Titcomb who talked us through Ancestry’s successful site testing strategies and the tools it used to achieve them.
How Ancestry.com works
People searching for records about their ancestors, or simply looking to build a family tree can use Ancestry.com’s built-in tools to create the records or reports they want. In exchange for access to its tools and database, Ancestry asks users to sign up for monthly or long-term subscriptions. However, before users pay for anything, they can test drive the service with a free 14-day trial. Since the majority of people that engage with the site’s services sign up using the free trial, there’s a lot of focus on how to convert them into paying subscription customers.
Ancestry.com collects all customer data and stores it in a homegrown database system, which integrates the data, and if possible, matches different sources to come up with a single, unified customer identity, which can be used for retargeting purposes. “We take in data sources coming from all our analytics, and look at where there are gaps or big drop off points in user activity,” says Titcomb. “We then come up with A/B testing strategies to fill those gaps.”
Optimizing the offer page
One of the conversion team’s biggest wins last year was the optimization of Ancestry’s “offer” page. This is the page that displays pricing options for the site’s subscription packages to non-logged in visitors. Ancestry used to display four pricing options to all site visitors, no matter where they came from. These options were “Choose Monthly,” “Choose Fixed (long term subscription)”, “US package”, and “World package.”
Titcomb says the team started tracking where all the site visitors were coming from and, based on their arrival paths, were able to group them into two categories. The people who arrived at the offer page by clicking on a ‘free trial’ or ‘subscribe’ button were grouped as “self-selectors,” visitors who were actively looking to see what Ancestry.com had to offer in terms of subscription pricing. The second group of visitors was classified as “interrupted browsers.” These were people who were trying to build a tree, connect to someone or access a record through independent web activity, but kept running into Ancestry’s paywall. In other words, they weren’t actively looking for an Ancestry offer, but kept stumbling onto one anyway.
“We decided to test whether both those groups of people needed to see all four offers, or maybe each group would be better off seeing only one or two,” Titcomb says.
Using Adobe Target to run the test, and Adobe Analytics to track visitor traffic, the team determined that the self-selectors should be able to see all four offers, since they were actively looking to purchase subscriptions. Interrupted browsers on the other hand, were better off seeing only one offer, the one that had the lowest barrier to give them access. In this case, that option was the US monthly package. “We saw a 20% lift when we did the test,” says Titcomb. “Those offer pages are a major driver of our revenue generating traffic, and this helped move the needle in a big way for us.”
Optimizing the home page
Much like the offer pages, testing of Ancestry’s home page began with grouping the page visitors according to their registration status. The Ancestry team identified four main groups of logged-out page visitors:
1) Non-registered user (nothing is known about the visitor)
2) Registered user (visitor’s email address is known)
3) Trialer (visitor is still inside the 14-day trial period)
4) Subscriber (visitor is paying for complete access to Ancestry’s service)
“We used to serve one home page to every logged out visitor,” says Titcomb. “We then started to test which elements on the page should be shown to which type of visitor.”
Typically, the home page has several items placed to cater to every type of visitor. These include elements such as a login widget, a “getting started” button, a “subscribe” button and more. By isolating each element on the page and testing its effectiveness in getting visitors to move to the next level, the team was able to come up with recommendations for each group.
The most prominent element for non-registered users would still be the “getting started” button. Titcomb says the team cleared everything else out of the way on the page and made this the focus so that people could click on it and quickly get into Ancestry’s tree building tool.
For registered users, subscribers and trialers the site prominently displays the login widget, with the goal of getting users quickly and efficiently into the main site. Once they get into the site, it becomes easier to tailor the content to each user, since they used a login, and Ancestry now knows who they are. “Now we have four flavors of a home page for logged out users,” says Titcomb. “By clearing out barriers and distractions, we were able to cut down our bounce rates, especially for registered users, in a very effective way.”