Posted by on June 3, 2013 in Analytics, Big Data, Inside our Offices, Machine Learning

I joined Ancestry.com as a developer in June 2002 – exactly 11 years ago – just as Ancestry.com was inventing the family tree technology that differentiates us today. The last 11 years have flown by as our team innovates against the constant challenge of improving our unique and dedicated community’s experience. Looking back over that decade-plus-one, I’m amazed at how much has changed in the technology industry, and pleased we’ve been able to adopt and leverage those changes to accelerate our core mission.

When I joined the company in 2002, the first iPhone wasn’t even in development, “android” was a science-fiction term for a form of humanoid robot, Google was working on its first-generation of search tools and Mark Zuckerberg had just graduated from high school. Hadoop was simply the name of Doug Cutting’s son’s stuffed elephant, and clouds were cirrus, nimbus, and cumulonimbus.

Eleven years ago, Ancestry.com had 200,000 subscribers, historical records numbering in the thousands and 40 developers. Today we have 2.7 million subscribers, 11 billion records and images (4 petabytes of data), and our developer corps has grown 10 fold, to about 400. In 2002, our search capability we were using was AltaVista. Now, thanks to our innovative vertical search technology customized for our business, our servers handle 40 million searches every day, and our technology-driven ‘hints’ are generating an impressive 60 percent of user discoveries. Back then, all customers used our software on desktops.

Change in the tech world has fueled change at Ancestry.com. This has always been a technology company, first and foremost. Last year we launched a new direct-to-consumer DNA test, AncestryDNA, that’s caught on fast and is helping subscribers build out and understand their family trees. AncestryDNA is generating exponential amounts of data that we’re managing with Hadoop, MapReduce and Hive.  We’re also exploring narrative technologies to give users another way to turn their family history into compelling stories, and photo recognition software to identify similar facial characteristics of ancestors and thus help users find more family matches. We’re always on the lookout for great engineering talent, and have hired gifted data scientists who are using Natural Language Processing, machine learning predictive analytics to understand future user behavior and enable us to make smarter business decisions. Our teams also innovate in the way we scale and manage our systems to be more robust.

What hasn’t changed since I joined Ancestry.com is our commitment to put the latest technology advances into the hands of the world’s best engineers to improve our customers’ total experience. Eleven years have gone by in a blink, since I’ve been fortunate to experience some amazing challenges and solutions we’ve developed in that time. And I can’t wait to see what interesting problems we’ll be handling in the next 11 years, in the technology industry as a whole and right here at Ancestry.com as we grow our exceptional team to serve the expanding global audience for family history research. It’s been a great ride so far, and it feels like we’re just getting started.

About Scott Sorensen

Scott Sorensen has served as Chief Technology Officer of Ancestry.com since April 2013. Scott has been at the company since June 2002 and has held multiple positions including Senior Vice President of Engineering, Vice President of Search and Vice President of Commerce. Prior to joining Ancestry.com, Scott was co-founder and Vice President of Engineering and then President at Coresoft Technologies. Scott was an engineering manager at WordPerfect / Novell and a software engineer at IBM. He holds a B.S in Computer Science from Brigham Young University.

1 Comment

Michael Mealling 

I found this blog yesterday after posting an enhancement request on the message boards and the Facebook page. The company I run, Pipefish.com, is building a taste graph based social network so I’m familiar with many of the graph related systems you guys are probably using. I even saw several ancestry.com people at the MLConf Workshop for GraphLAB on July 1. Does Ancestry.com use GraphLAB at all?

Anyway, my question is when/if Ancestry.com is planning on giving users direct access to the graph? Looking at several of the threads on the Message Boards there are requests such as tree audits, relationship confidence, typed edges other than familial (co-worker, army buddy, etc), etc. Some of these could easily be done by third parties who could build add-on apps that could help build out where Ancestry.com doesn’t have the resources. I definitely realize how much of a headache creating an API is but its something I kept wanting to reach for as I used the system.

I’ve been using Ancestry.com for about a year. I have two trees of about a thousand people. And I’m sending in my DNA kit tomorrow.

Thanks!

July 30, 2013 at 2:26 pm

We really do appreciate your feedback, and ask that you please be respectful to other commenters and authors. Any abusive comments may be moderated.

Commenting is open until Monday, 17 June 2013