by Allison Banko
When attempting to unearth the roots woven by their family trees, subscribers look to Ancestry.com for the tools to do it.
The website is the world's largest family history resource with more than 13 billion records and images to help its 2.7 million subscribers "discover, preserve and share their family history," explained Emily Titcomb, Senior Manager of Product Marketing, Ancestry.com.
Ancestry.com uses information from users' family trees to search its vast collection of historical records. While making a family tree is free, users must sign up and become a subscriber through an offer page to access any of the records.
This offer page can be accessed through a variety of paths on the Ancestry.com site. However, despite the fact different users along different journeys were hitting this page, it was the same for everyone, designed as a one-size-fits-all solution.
Because the offer pages are the highest trafficked areas on Ancestry.com where visitors convert to members, the team felt these pages were worth testing. With the overall goal of increasing conversion rates, the team's first motion was to optimize the current page
in an effort to differentiate subscribers.
But once the in-house testing and optimization team started to dig, they realized there was more to be done. While users were more targeted through the optimization of the page, now the design of these offer pages wasn't consistent. Because of this, the team sought to test a consistent look and feel across all of its offer pages.
To accomplish these goals, Ancestry.com developed a three-phase plan:
- Optimize the current page
- Create a template
- Customize and optimize
"It was a group effort across many teams to get this whole process out the door and tested," Titcomb explained.
The in-house teams involved in this project included:
- Product marketing — provided high-level strategic direction
- Creative — offered UI and UX expertise and vision
- Testing and optimization — ensured the test integrity was sound and that the team would get the data and answers to needed questions
- Analytics — provided sound and detailed analysis used to make the business decisions for the tests
"All these teams were involved, and the shared goal across all of the teams is to improve the user experience and conversion on our offer pages," Titcomb said.
Phase #1. Optimize the current page
Ancestry.com's starting point was to optimize the one-size-fits-all offer page hitting all of its users, regardless of the path they took to get there. To optimize the package options page, the team first began to map out various ways people could reach it. From this, they observed two distinct groups of users: "self-selectors" and "interrupted browsers."
Self-selectors were those that went directly to the offer page by clicking on the "free trial" or "subscribe" buttons located in various places throughout the Ancestry.com site. Interrupted browsers were those who were trying to access a particular piece of site content, but were denied.
"They're being denied access to view the content because they aren't actually a subscriber on the site," Titcomb explained. "They can't view the record or they can't view the picture or whatever type of content they're trying to access."
The team then examined how large the population was of those interrupted browsers who were being denied the content, noticing this group comprised 60% of offer page visits. The conversion rate for the interrupted browsers was five times lower than that of self-selectors.
Due to these two metrics, the team decided to test the interrupted browsers on the offer page.
Titcomb said the goal of this test was to improve conversion and lifetime revenue by lowering the decision-making hurdle on Ancestry.com's highest-trafficked offer page.
"Our hypothesis for doing this was, let's lower the decision making hurdle, so [we] don't show them all four options and make them choose," she added. "[We'll] just default them into the lowest package and see if that helps with the decision hurdle and getting them in the door and converting."
Within the test, the team looked at three factors:
- Number of packages
- Price point
- New creative
This was achieved through a four-cell test.
The test showed the highest conversion and the highest lifetime revenue winner stemmed when interrupted browsers were only shown one package option at the $19.95 price point.
This resulted in a 20% increase in conversion. The team concluded that reducing friction of choice
on the offer page led to a significant increase in conversion, but only at the right price point.
Phase #2. Create a template
While Ancestry.com now had a new, high-performing page for interrupted browsers, its creative was much different from that of the other offer page for self-selectors
. Titcomb explained that people throughout the company wanted the look and feel of the offer pages to be more consistent.
Now that Ancestry.com knew what it should show to users, the next step would be getting all of the offer pages on a standardized template where branding was uniform. However, according to Julia Babiarz, Senior Interactive Art Director, Ancestry.com, there was more than just a desire to unify the design and branding in this phase.
"We saw the opportunity to move forward and optimize further through customizing the content and serving imagery and messaging that would be more tailored to the person that was seeing the page," Babiarz explained.
Customization could be achieved through customer segments, how a user reached a page or via other information Ancestry.com knew about the user.
However, the existing designs of the offer pages were in a format that wasn't conducive to changing out content blocks. The team needed to create a template that would not only unify the look and feel of the pages, but would also allow for imagery and messaging to be easily replaced for customization in the future.
The team created a new modular template design and launched a simple A/B test, testing the old design
of the interrupted browsers offer page versus a new template-based design
of the page. The template design resulted in a 7.5% decrease in conversion.
"That is the risk of running a redesign as an A/B test, because there were many things on the page that were completely different," Titcomb explained. "We were hoping that our new design would win, and then we could move forward. Since it didn't, we then were left with the challenge of why. What caused it to perform poorly?"
To answer this question, the team conducted a second stab at the test, this time identifying the different elements of the design that may have influenced conversion:
- Size of the buttons (standard vs. large)
- Placement of the imagery (side vs. top)
- Placement of the price (near CTA vs. far from CTA)
Ancestry.com designed a nine-cell experiment that tested the controls across all of the variations of those three factors on the new template.
Of the three factors tested, the placement of the price had the largest influence on conversion. The test showed that all of the treatments with the price near the CTA had the lowest conversion rates, while those with the highest conversion rates had the price far from CTA.
The winner, with a 1.94% conversion, contained the following elements:
- Standard button size
- Images on the side
- Price far from CTA
The team now had a template to move forward with and rolled out this template on other variations of the offer pages, including:
- Free trial deny page
- Hard offer deny page
- Upgrade deny page
On the upgrade deny page
, there was a 10% lift in conversion that further solidified confidence in the new template.
"I think we ended up being successful because we were able to quickly go back and then launch a follow-up test that did look at individual elements on the page," Titcomb said.
Phase #3. Customize and optimize
With a tested-and-true template, Ancestry.com was now positioned to target, customize and optimize on a more personal level for Ancestry.com visitors, including:
- Interrupted browsers
- Past free trial users
"We are in the midst of testing tailoring the content to different folks based on user journey and customer segment," Titcomb said.
This phase is not yet complete, but the team envisions reaching a point where they can develop items such as targeted sub-headlines containing the names of specific family members a potential subscriber has already begun to build in their tree.
As seen in Phase 1, Ancestry.com discovered its opportunities by mapping out different user journeys and relative conversion rates. Specifically, the team recognized poor performance in its interrupted browser segment on its highest-trafficked offer page.
It was also during this initial phase that the team discovered the curse of too many choices. In Ancestry.com's case, reducing friction on the page by presenting only one package option resulted in a 20% lift in conversion.
Phase 2's findings prove that isolating your testing elements is imperative. The team's first test template test was unsuccessful, but because they performed an A/B test with multiple changes, they couldn't identify why the new design lost by 7.5%. It was when Ancestry.com outlined the three elements (button size, imagery placement and price placement) that the team could clearly identify the true element for success, and failure.
Titcomb said pinpointing different user behaviors and journeys and learning as much as they can about them can result in significant impacts — and it has.
"If you add up all the small wins," Titcomb said, "it actually becomes bigger when you look at it holistically for the whole company," See Babiarz and Titcomb's featured subscription case study presentation "Offer Page Transformation: From one-size-fits-all to customized experience" at Web Optimization Summit 2014, held May 21-23, at The TimesCenter in New York City.
- One-size-fits-all offer page
- Optimized interrupted browsers offer page
- Self-selectors offer page
- Phase 2 control
- Phase 2 treatment
- Upgrade deny page
Julia Babiarz, Senior Interactive Art Director
Emily Titcomb, Senior Manager of Product Marketing
Related ResourcesWeb Optimization: VacationRoost implements 2 testing methodologies to boost total conversion rates by 12%B2C Web Optimization: 23% lift in conversions by optimizing step one of funnelB2B Web Optimization: 140% surge in mobile transactions through responsive design effort