Thursday, July 10, 2008

Kendall Hulet Interview, Part 2

Last Monday I got to speak frankly about Ancestry's New Search with Kendall Hulet, director of product management. (Kendall and I go way back. Someday I'll have to tell you about the argument we had about trees and—it's still painful to think about—how I turned out to be wrong. Now that I think about, I wasn't wrong so much as Kendall was right. But I digress...)

We sat down together to hash over my concerns with New Search. I wrote about a couple of the worst problems earlier this week. See Inside the problems of Ancestry's New Search. As of Wednesday night, the promised fix for the new home page Quick Links hasn't been released. I assume this will be released as soon as possible.

Here's my entire laundry list of problems with New Search. Following the list, I'll tell you what Hulet had to say about them.

  1. New Search returns too many false positives, as was discussed in my last article.
  2. New Search requires more clicks than Old Search.
  3. New Search lacks the database specific search forms that Old Search has.
  4. Old Search has bugs.
  5. New Search has bugs.

New Search requires more clicks

Hulet and I discussed this at length and I provided a variety of examples contributing to the rise in the number of clicks needed to perform everyday tasks. Hulet explained that they are aware of most, if not all, of these issues and are exploring different ways of addressing them. But, he said, fixing known bugs is a higher priority.

I think New Search presupposes use of trees and ranked searching. So users who depend almost exclusively on exact searching are running into bugs and design flaws that Ancestry didn't find or didn't think about. New Search is proving to be painful and extremely less productive for exact searchers.

While trees and ranked searching are very useful and important new tools for searching, I'm shocked and surprised that many long-time users are not using them. I consider myself a power user's power user. I think there are very few people on the planet who understand and use the search system for real-life searching more than I do. I think you'd be hard pressed to find anyone that could generate successful searches faster and more efficiently than I can. And I consider tree-based searching and ranked searching to be essential tools for effective searches in some situations.

Consequently, I'm surprised that so many people are not using these tools alongside the primary and essential tool of exact searching. I wonder if Ancestry knows this? It's a real failure on Ancestry's part to not communicate and teach users how to utilize these tools.

I guess I should try and show how these tools can be effectively used. If I forget to address this in future articles, keep reminding me to teach you when and how to use tree-based searches, ranked searches and exact searches to add value to your Ancestry experience.

My wife says it's time for bed. (I'm writing this last night and scheduling it for morning publication.) So what follows is the remainder of my outline, with some reminders of points to cover. I'll cover these topics in a future post.

New Search lacks database specific search forms

Old Search has bugs

Include scary big numbers and Results by Category here.

New Search has bugs

I'm aware that nearly simultaneously with the release of New Search, at least one database got broken.


  1. Requiring MORE clicks is a problem for those with disabilities and arthritis. In another setting, it would be a violation of the Americans with Disabilities Act. I appeal to Mr. Hulet's better part to consider the morality of making things harder for those physically challenged.

  2. Insider,

    First to be clear about a term. When you say "tree/s" you mean family trees submitted to Ancestry or associated sites don't you? That is, you do not mean "tree" as a data structure correct? This is the way I am going to interpret it here, i.e. family trees.

    When you talk about including family trees and using ranked searches, you are basically making the equivalent of a preliminary survey on Ancestry's total set of data. But the reason we use exact searches on specific databases is that we have already done that and gotten way too many irrelevant results.

    You extol ranked searches, i.e. ranked by by relevance. But that presumes ranked according to some set of criteria. But who specified those criteria? That's right, largely it is Ancestry due to their
    constraints imposed by their search methodology. But we are interested in *OUR* criteria. We want searches ranked by what is relevant to *US*.

    If we had the ability, which we should, to have a versatile search form that can be used to make field by field exact search specifications for an individual database, then since we are not going to check every field exact box, we are in fact getting a ranked search, but to *OUR* criteria.

    We really should not even be having this discussion to the degree that we are with Ancestry. They can do whatever they want with site-wide global searches, both old and new. But if they refuse to upgrade the search forms for individual databases from census to military or whatever, so that we can specify exact search on a field by field basis, then they really have insulted us "type 1" customers and basically told us they don't care all that much about our business.

    Again, Ancestry can do whatever they want in an "old" or "new" site-wide search for the masses, but give the most experienced customers the option to do otherwise, i.e. meaningful specific database field by field exact searches.


  3. I agree that more user education is needed. However, I am afraid that the level of animosity shown by posters on the Ancestry blog indicates that many would only see it as more "marketing" of a bad idea.

    It will take more independent input to get through the din of discontent. I think it is great that you are covering this in your blog. I hope more people read it.

    I use the new search exclusively for both fuzzy and exact searches. I like it because I know how to get results. Several times I went back to the old search and just did not find it as productive (for me) as the new search.

  4. I just tried the new search again looking for a specific Missouri Marriage and was stymied by the drop down boxes and had to hunt to find where to enter the "spouse" name which was a drop down under "family". It was most discouraging. I regularly use the old advanced search and if they take that away I'm not at all sure what I'll do. One item that would help would be the ability to resort the search results by different fields. Now THAT would be an improvement. Sorry for the long post.

  5. Dear ADA Advocate,

    You've brought up an issue that I hadn't even considered. Many people have complained that "it takes too long to use the new search." That's not helpful. I gave Kendall a detailed click-count analysis of at least one search scenario (suggested by the good Randy Seaver). That level of detail is required to make complaints actionable.

    But you've added an important flavor to click-count analyses. I'm very glad you responded.

    -- The Ancestry Insider

  6. Dear Jim,

    Thanks for the contrasting point of view. I like it too. Plus, I think we we end up with it or something very much like it after the dust settles. That makes it very important to rationally figure out--for those of us that don't like particular aspects of it--exact what it is that we don't like about it, quantify it, describe it in detail, and get it to Kendall.

    Thanks for reading and for commenting.

    -- The Ancestry Insider

  7. Your comments re Tree-based searching are interesting. I wish you had explained whether most of your searching was in 19th century USA folks so well covered in census enumerations, or what.

    The truth is that Ancestry's databases have huge gaps for USA (ME Town Records? WV Tax assessment rolls? Church records from PA or the Midwest?). I suspect that this is one reason for not fixing the 'false positive' results in New Fuzzy Search more thoroughly - a marketing tool since they do not want us to realize how limited the resources really are. I use a tree on another site to catalog research done. Ancestry's limited database will not help me with ancestry since I have already done most of the work in *records* over the years, and need site-specific searches in local records not available on Ancestry (estate, land and tax records) to make progress. It does not help that Ancestry does not fix broken database elements that they were informed about years ago.

    Another reason for not putting a reference tree on Ancestry is that, as a paying subscriber, I simply will not post my hard-won data for Ancestry, in effect, to require me to continue to subscribe to view or modify, or require that others pay to view. Nor do I want silly Ancestry Hints pointing to the many erroneous trees containing numerous of my ancestors and cousins, often from publications that began what I call 'widely held mistaken beliefs'. Enough is enough.


Note: Only a member of this blog may post a comment.