Monday, February 22, 2010

Ancestry.com Bloggers Day: Search

Last year I intended to do stupendously rich articles about Ancestry.com Bloggers Day presentations. Since I never got around to it, this year you’re getting my stupidously poor notes.

Tony Macklin is the Ancestry.com search product manager.

  • How many user-submitted corrections are being made?
    • There were 80,000 a month before the new record page that combines the record view, the image view, and the correction page all into one.
    • There are 900,000 a month now
  • The new combined page was not immediately used for all databases
    • It was launched on a few census databases, then expanded to all census years
    • It was on 19 databases for the latter part of 2009
    • It will be used on between 4 and 5,000 databases by the end of January
    • It will be used on all structured (fielded) databases by the end of the year
  • Ancestry.com and Google.com share some common characteristics. They both index content and provide a way to search the content. They both present search results, and both allow users to click a result to view the content .
  • But Ancestry.com and Google have several striking and important differences.
    • Ancestry.com’s search is about individuals. Ancestry.com users are not looking for the most popular person, but a particular person.
    • Ancestry.com’s search is about understanding places. Names and boundaries change over the years.
    • It’s about understanding names. Names have various spellings, nicknames, abbreviations, and translations. For example, Catherine has over 800 variations.
    • It’s about understanding records. Google creates indexes solely for electronic documents, while Ancestry.com creates indexes for records of many different types, formats, characteristics, and appearances.
      • Ancestry.com works with a large range of sources
      • Handwriting is hard to read.
      • Searching needs to handle inaccuracies in names, dates, and places.
      • Searching needs to blend together results from newspapers, photographs, and structured records.
  • What are the snippets of information you know?
  • New search is currently used for 87% of searches and Old Search is used 13%. However, Old Search users are twice as likely as New Search users to successfully find what they are looking for.
  • The issues are:
    • Relevance and control
    • Browse (which is very important and also instructional)
    • Speed and efficiency

Next week I’ll talk about the newer than new search prototype that Tony shared with us.

Tony Macklin of Ancestry.com Tony Macklin is senior director of search for Ancestry.com. This includes the search experience, the search algorithms, and the technology behind hinting. Tony has spent most of his career in the UK and joins Ancestry.com with more than 10 years experience developing products for US based online companies such as Intuit, Ask.com and eBay. Over the last 5 years he has concentrated on how to make search deliver better, more relevant results. He is based in the Ancestry.com San Francisco office.

1 comment:

  1. Congratulations on being a part of the top 40 in the Family Tree Magazine!

    ReplyDelete