Friday, January 09, 2009

Touring TGN/Ancestry

See also: TGN/TNG: Or was that a trekie reference?

DearREADERS,
Several have written to inquire why Ol' Myrt here has not blogged yet today. "You are uncharacteristically quiet," Peg wrote.

Well, truth be told, I spent the entire day and early evening touring The Generations Network [TGN] facilities and talking with upper management. TGN is Ancestry.com's parent company. I was one of seven that comprised a visiting focus group of genea-bloggers, authors & librarians. I'll let most of the other six guests reveal their identity as they speak of our tour of the TNG data center and corporate headquarters.

But Ol' Myrt here can't resist telling you that she sat across the table from the Ancestry Insider, whose face was shrouded in the shadows. It was as though Mother Nature herself conspired against us to meet the man, by providing a stunningly bright, sunshiny glare from the snow-covered mountains outside the window behind AI. Or maybe he, himself, chose that exact spot to further thwart our efforts to see him clearly.

Shall we never know AI's true identity? He was wearing his trademark suspenders, and managed to blog while we were taking notes and learning more about the process of preparing a record group for scanning, indexing, optimization and final presentation on the Ancestry.com website.

The fact that the Ancestry Insider was in attendance let me know that TGN is open to conversation about improving their website. For my newbie readers, the Ancestry Insider is a former TGN employee now working at FamilySearch, who is very bloggal (blogger +vocal) about the challenges users face at the Ancestry.com website.

Randy Seaver reported some of the stats we learned on our tour of the massive TGN data center. We actually walked past the hard drives that house our online trees, ancestral photos and those tens of thousands of precious US, Canada and UK census records. The back-up systems for data storage, machine repair, fire protection, and power are mind-boggling. This is a multi-million dollar server center -- all for our benefit.

Perhaps the most impressive expressed commitment is TGN's significant challenge to refine the Ancestry.com search interface. A good search engine makes or breaks a website for genealogists seeking to find ancestors through indexed entries and related scanned images of documents.

And Ol' Myrt here had a little fun pushing forth her personal belief that family history isn't so much a name search, as a locality search.

Why Myrt's emphasis on locality?
Finding proof of ancestral relationships rests on my ability to canvas all surviving record groups (at Ancestry, FamilySearch, WVR, GenealgyBank, the local and state archives, libraries, etc.)

But I need to know what those records are generally, and which specifically are to be found in one of Ancestry.com's massive computer hard drive arrays. With a name like SMITH, it would really be better if I could search Ancestry.com FIRST for the LOCALITY and secondarily by name.

For instance, this means the search algorithm working in the background, would query all Ancestry.com databases to only include events in Scioto, Ohio, and thereafter search for a fuzzy set of name matches but only within those Ohio entries.

I am particularly frustrated when I've painstakingly completed the search form at Ancestry.com, clearly specifying the ancestor was born and died in Scioto, Ohio, but what I get at the top of the hit list are entries that are a close match to the name I specified but include oddly irrelevant localities like Florida or California, clearly not even adjacent to Ohio.

Why would I want a good name match, if the locality for those matches are for people who lived and died in Florida or California?

Beginning genealogists look for an ancestor strictly by name, and then give up quickly.

Experienced researchers know that spelling of names doesn't really count.

Even when I've proved an ancestor lived in Scioto County, Ohio, if I don't know where to go for earlier generations, a thorough review of all surviving town, county and state records in Ohio must be searched for a clue about a previous residence.

But how can one possibly discover what records groups have survived for Scioto County, Ohio and more specifically which of those are on the Ancestry.com website?

How about a hit list of related database titles for the specified locality?
It would be ridiculous to expect Ancestry.com to tell me about every available record group in the world. But failing a good match on the name sorted hit list, I'd like a column that lists all Ancestry.com content for Ohio. Browsing that content might turn up info on a name that was written so illegibly as to defie indexing.

Why not populate each Ancestry.com search results page with more than just name matches?

The short story, bottom line is that TGN is missing the opportunity to direct researchers to Ancestry.com relevant locality content because the current search prefers to rank name matches higher than locality searches and related locality databases are not immediately available. I like my beginning students to print out a list of the first 20 hits, and then check them on the print out as each is eliminated, so they don't get lost in the search process.

All in all
Unequivocally, today's meetings receive high marks in my book. TNG's team managers described their missions and provided new insights to the challenges of data acquisition, document preservation, digitization, indexing, family trees, Family Tree Maker, DNA, MyCanvas and such. Several clearly care about their personal family history research.

Interesting fact: It took 6 months to flatten the rumpled Gretna Green, Scotland marriage records, so they could be scanned. The scanning itself took a mere 3 days.

In a discussion of allocation of resources, I learned that 85% of indexing/correction effort is spent on new content, and 15% goes toward fixing mistakes in existing content. Even though there are flaws in some databases, due to some no-longer used indexing systems, if I know the fix is on the list, I can wait. I really want new content, and think that TGN ratio is a good mix. Eventually, as new content goes up with less errors, the old content errors will get cycled out of the picture. Prior to meeting with TGN employees, I was under the impression that Ancestry.com had turned a deaf ear to the indexing problems Ancestry.com users are reporting.

I walked away from today's meetings with a healthy respect for this company's commitment to providing a quality experience for Ancestry.com visitors. Remember, Ol' Myrt here has always recommended my readers include Ancestry.com on their "A list" of "must search frequently"genealogy websites.

Summary
How do you grow a website to support an expanding collection? Certainly, TGN needs the marketing people to bring in the revenue to provide the content and content presentation to meet their client's needs.

Taking into account there are at least two types of "genealogists" out there -- the novice and the professional, consider the vast difference between what a marketing expert thinks is important at a genealogy website, and what genealogists of either type actually value in that same website.

If acted upon, today's conversations have the potential to go a long way toward bridging that gap.

Happy family tree climbing!
Myrt :)
DearMYRTLE,
Your friend in genealogy.
Myrt@DearMYRTLE.com

Blogs:
http://blog.dearmyrtle.com/
http://blog.teachgenealogy.com/

Podcasts:
DearMYRTLE's Family History Hour
http://podcasts.dearmyrtle.com/

Family History Expos Genealogy Podcasts
http://fhexpos.libsyn.com/

DearMYRTLE's Facebook® Group Message Board
http://www.new.facebook.com/board.php?uid=2960625373
© 2009 Pat Richley All Rights Reserved.
This and previous blog entries are fully searchable by going to
http://blog.DearMYRTLE.com. Myrt welcomes queries and research challenges, but regrets she is unable to answer each personally.