Hulet noticed that experienced users came to Ancestry.com day after day and reentered the same searches time after time. Plus, they wouldn’t enter enough information for Ancestry.com’s search engine to find many of the matching records. At the same time, Penrod noticed that new users came to Ancestry.com and typed in their own names, then left when they found nothing.
Commiserating together, they wondered how to solve these problems. Then they had an idea. Respond to visitor’s natural instinct to enter their own name by having them build their own tree. If they allowed users to create a user tree, they would work past themselves, and they would provide enough data about ancestor to do a good search. That’s how we came up with the idea of member trees and shaky leaves, said Hulet.
They now have
- 60 million trees created
- 7 billion profiles created
- 5 billion hints accepted
- 3 billion records attached
The three presenters emphasized that there needs to be a common vision among their three teams: engineering, product management, and UX (user interface) design. “If you only remember one thing from this presentation, getting to a shared vision is the most important thing.” The shared vision needs to include what they are solving and who they are solving it for, said Hulet.
To better communicate who they are solving it for, the team creates personas—market types—that describe the archetypical customer. They give the persona a name so they feel a greater connection to that customer segment. An audience member asked if they could share more. They declined; they felt that information confidential. But I’ve seen this done enough times, let me sketch one out:
Nelly Newcomer is a forty-something married female. She’s well educated and has a career. She has expendable income, but little free time. Her parents are both alive, but aging. She has one sibling and no children. She’s not the scrapbooking type, but enjoys looking through old photo albums with her parents and reminiscing. She’s seen part of an episode of Who Do You Think You Are and was intrigued with the idea of learning more about her ancestors.
It takes a lot of interaction with customers to understand their needs. You have to get your product ideas out in front of your customers. Ancestry.com product managers have been known to get in line at Starbucks, strike up a conversation with the person next to them, and offer to buy their coffee if they will sit down and review an idea. (They said in Salt Lake City they might frequent cookie shops instead of Starbucks. That’s not to say we don’t have them in Utah. I walked two blocks to McDonalds this morning and passed three Starbucks. I heard Jay Leno say once that Starbucks was proliferating so much, they had started putting Starbucks in the parking lots of other Starbucks. But I digress…)
The rest of the presentation consisted of different ways and products that Ancestry.com uses to watch and learn from their customers. They fly all over the place. They use software to remotely watch the computer screens of website users. They have popups that ask users if they would be willing to participate in some testing. If they are willing, the survey gathers some demographic information that Ancestry.com can use in the future to determine who to contact.
They have “listening labs” without prescribed agendas. A notice on the site invites a user to participate. If willing, Ancestry.com calls them immediately and asks them what they’re hoping to do on the site that day. Then Ancestry.com just watches them work, perhaps asking a question now and then. “We do these every Friday,” Penrod said. Sometimes the CEO sits in. If he sees an obvious problem, you know you’re going to get a “we should fix that” assignment. These sessions are also great bonding exercises for teams when the team watches together and sees failures in the features they are responsible for.
In the question and answer segment, someone asked that if Ancestry.com was so interested in learning from their customers, why wasn’t there a feedback link somewhere on their site. They explained that Ancestry.com has millions of customer interactions and that if they had a feedback link, they would not be able to review all the submissions. It’s a bad idea to ask for input and that you don’t actually read.
They admitted that they don’t always get it right. But from all they said, it is apparent that they try.