Posted by Anne Gillespie Mitchell on August 4, 2008 in Website

I’ve gone through and reread the comments that members have left on my initial search post, and I wanted to try and summarize what I am hearing, just to make sure that I understand what you are saying.

I believe that the biggest problem that has been discovered is that the detailed searching on some datasets is not available for the new search interface. For example, on a marriage record, you can’t specify a groom’s surname. This ability to use your knowledge of the data that is stored in any given dataset is hampering your ability to pull out records in a manner useful to you. I think this is a problem. I don’t have a solution for it today, but I believe a solution can be found.

I know some of you would simply like the new search interface thrown out and the old search interface become the only search interface. I don’t think that is the correct answer.
From looking at search results and search patterns, and from comments (and there were hundreds of them) from the webinar, that were largely in favor of the new search interface as well as surveys we have conducted, there is a lot of enthusiasm for the new search.

I suspect that the answer is in adding in a way to filter on fields specific to the database as well as some other tweaks/improvements to get the new search interface working properly.

Some of those improvements might include solutions to these problems that surfaced in your comments:

  • Limit the number of results that you see at one time
  • You want a way to be able to search census records and not see records before the birth date and after the death date.
  • You would like an option to limit your searches to not include certain data sets, for example family trees.
  • Some of you believe, that a document should only be excluded if the contents of that document explicitly contradict what you have entered. (This applies mainly to death dates on census records, mainly.)
  • There are times when you want every thing ancestry can find on a person, and there are times when you want to be able to filter those result by specific data sets or general categories of data sets.
  • Some of you would like the ability to turn off type ahead for names and locations.

Now, I’m sure that I’ve missed a few things, or not stated things in the clearest manner.

I need to have some conversations with engineers, designers and other product people and try and figure out solutions to some of these issues and if we can get solve them in a timely manner. Not all problems are easily solvable. And I don’t want to make promises I can’t keep. 🙂 But I do want make sure I understand the problems before I try and figure out solutions.

Anne Gillespie Mitchell

Anne Gillespie Mitchell is a Senior Product Manager at She is an active blogger on and writes the Ancestry Anne column. She has been chasing her ancestors through Virginia, North Carolina and South Carolina for many years. Anne holds a certificate from Boston University's Online Genealogical Research Program. You can also find her on Twitter, Facebook and Finding Forgotten Stories.


  1. Connie

    There are very few times when I want to search for everything Ancestry has on a person (I would do that only if it is a very rare name).

    Usually, what I, as an experienced and serious researcher, want first and foremost is an easy way to find what records you have for a specific location. So, for example, I want to know what records you have that cover, say, Holmes Co., Ohio. (I am using the USA as an example, but the same principle would apply to each country).

    Then, I want to be able to search each of those records individually.

    If the record was kept at a federal or state level, then I want to be able to quickly access, as an example, the federal land records, or the federal military records, or the federal census records. Then I want to narrow those records down by state, or time period (or however they are organized) before I start searching for a name.

    This is not rocket science. It is a basic genealogical research principle: you do research by locality, time period, and record type, not by shotgunning a name all over the world.

    For the life of me, I can’t figure out why Ancestry is so invested in the bells and whistles on the new search interface: it ain’t about looking pretty and cluttering up the screen; it is about getting accurate results easily and quickly.

    It is this misdirected insistence on focusing on pretty graphics and names (instead of names within a locality and time period) that is surely at the core, ultimately, of all the complaints.

    It is also, I’m sure, why two of my great-grandparents (whose parents and siblings have clearly been identified and documented for decades) have suddenly sprung new siblings on Ancestry trees. (I haven’t had the heart to check out the other 6 great-grandparents yet).

  2. Mike

    “There are times when you want every thing ancestry can find on a person, and there are times when you want to be able to filter those result by specific data sets or general categories of data sets.”


    The above is close to what I have talked about but not quite there. In order to make good research logs/calendars, one must list *every* individual source/reference checked and not just “all the stuff Ancestry has” or even “all the stuff Ancestry has on this general topic (like military)”.

    Thus what I and others too I think would like, is some attention paid to revamping the database specific search forms for individual databases, and in the future, making sure that new databases have the following ability, i.e. the ability to make an exact field by field search on any database specific search form. Instead we have to go through the site-wide search form whether old or new to get field by field exact searches, and mostly only have global exact boxes on individual databases.

    Speaking for myself, if I had that (and all individual databases actually had a search field for all fields that exist), then I would not care much what you did with old or new search for site-wide searching, and thus would not be complaining now or in the future.

    Give us that fine-tuned ability to make exact searches on every field in each and every database.


  3. Tyler

    Very nice post, Anne! Of course, I’m not sure whether or not you have missed a few things either (others have more experience and needs that I’m not aware of), but I like the way you have clearly proposed a list of improvement candidates as points for further discussion/action.

    Personally, I would like to see improvements in the area of your 4th item: no inadvertent (to the searcher) exclusion of records caused by their (often unknown to the searcher) lack of specific fields/attributes within their respective database.

  4. Linda


    I appreciate your efforts in trying to thoroughly understand the issues surrounding the new search. You are dealing with a variety of users who search in different ways so it is not an easy task.

    I agree with you in most of your observations except that there is a lot of enthusiam out there for the new search.

    Personally, I conduct a variety of searches from very general name searches to very specific searches. Sometimes, it would be nice to search in a category like census records or birth records. With all due respect for Connie and what search strategy works for her, there are a number of valid search methods that are used by serious genealogists such as myself. People do not stay in the same place all the time so a search by locality, time period, and record type might result in losing the trail and overlooking a lot of great leads. Reviewing a set of records can help prove or disprove genealogical theories or provide valuable leads. The convenience of a site wide search is a reason for subscribing to a paid service vs. using some of the compariable sources available for free online. But, as Mike said, there is also a demand by users to be able to search by database specific criteria. Sorting by criteria would also be helpful. For example in HeritageQuest, I can sort the results by one data field (surnames, first name, county, age, etc.) which can be extremely helpful. To keep ahead of the competition of free and other paid services, must provide stronger search and sorting capabilities.

    I don’t like the type ahead. One example is that I conduct a lot of census records for Bureau County, IL. Since there is only one county in the US named Bureau, under the old search, I would quickly type Bureau in the County box. With the exact search on, I got only Bureau County results. In the new search with “Match All Terms Exactly” on, I type in Bureau (don’t wait for type ahead) and and get results for Cook County and Knox County (for example with search of James Hensel). Hmmmm….

    A good search interface allows the user to make a variety of searches from very general to very specific and also to be able to sort the results in a variety of ways. It also doesn’t surprise the user with unexpected or missing results. The “old” search fall short of these goals. The “new” search simply fails. Keep trying.

  5. Connie

    Linda makes a very good point: it was not my intent to imply that I never use the site wide search (or expand to other localities). But I almost always do so only after I can’t find people where I think they are supposed to be, or if I’m dealing with a rare name. (And I suspect that is how most serious/experienced researchers approach the task).

    And the fact remains that it is a basic principle of genealogy that you do research primarily by locality, record type, and time period, not primarily by name: that is the point that Ancestry seems to have missed.

    While “old search” is not perfect, it is “clean.” When one does choose to use the site wide search, the results need to be presented “cleanly” so that you can quickly zero in on your guy. You can do usually do that with “old search;” you can’t do that with “new search”: the results are a visual mess that take forever to sort through.

    And I concur: the type ahead feature is not an improvement. It wastes tons of my time because it guesses wrong. The old search drop down menu for narrowing by locality is much easier/quicker to use (and less of a visual mess).

  6. Karen


    I appreciate your summary of the concerns many of us have about both and old and new search. I admit that I’m one of those wondering who those people are who prefer the new search. Since your survey’s are rigged to elicit only positive comments, they are really useless as far as determining what it is users really want. I assume/hope that you have other ways to study real world users.

    In my opinion, the single most helpful improvement that Ancestry could make would be to enable sorting by field (column heading) on the hit list. That is such a common feature of most web applications that many users can’t believe it’s not possible. They really don’t care about Ancesty’s rankings, they want to sort things for themselves.

    I don’t the type ahead fields either, I think there should be some way to turn them off. What is the point of logging into a profile that does not actually store user preferences for these things? We can’t control what appears on the home page, we can’t eliminate the unintelligent “smart” tags that appear in messages on the message boards, and there’s no way at all to customize the search page. The web team needs to focus more on actually letting users customize their own “experience”.

  7. W. David Samuelsen

    I do NOT like the new home page!

    I am STILL trapped in somebody else’s family tree page with NO way out.

    Sought out all ways – NOTHING!

    I want OUT of somebody else’s Family Tree!

    It is NOT nice being a hostage!

  8. Anne Mitchell

    W David, I’m sorry you don’t like the new homepage, but I think this is probably not the best blog posting to comment on that. It helps if everyone stays on topic.

    Connie, Mike, what I am hearing is that for you, and I think many others, the best search is a specific search with the ability to search on all the fields in any given data set. I have been discussing with my engineers and others on the best solution for this and hopefully I’ll be able to give you an idea of what we hope to do and when we hope to do it. And I also agree with Linda, that being able to go back and forth between general and specific is important; being able to do it easily is even more so.

    Tyler, the exact search idea, not excluding specific records is easier to implement in the more specific searches. More to come on this…it can be simple on one dataset and not so much on another.

    Sorting. Karen, I agree, sorting would be lovely. I know I use it on other sites, but they usually have fewer records in a query and they store their information in ways that search indices usually are not stored. If you do a search that has 100,000 possible results and then you do a sort, it is very cpu intensive, we simply aren’t built for that currently. It’s the large numbers of results that make it prohibitive at this time. That of course doesn’t mean there aren’t solutions, there are. We can start limiting the number of results (which has been requested), and we are thinking about this as we constantly look to improve our search engine. Are there are plans in place for implementing a variety of search engine improvements in the future.

    I’m also a big fan of personalization and hopefully we shall be able to work more of that into the site in the near future.

  9. kkandtc

    The newly loaded 1900 census is already a problem. In the old set, there was an Elen and Hattie Haddleston living in the household of Henry Mitchell in Roane County (Harriman), TN. They are now correctly indexed as Huddleston. They do not show up in the old or the new search unless you search on Henry Mitchell.

  10. Andy Hatchett

    Hi Anne,
    I find one sentence in your post that just struck me the wrong way…

    You said” From looking at search results and search patterns, and from comments (and there were hundreds of them) from the webinar, that were largely in favor of the new search interface as well as surveys we have conducted, there is a lot of enthusiasm for the new search.”

    To be quite frank about it, we who find problems with the new search do *NOT* want to keep hearing from *ANYBODY* associated with Ancestry about how great others think the new search is- WE DO NOT CARE WHAT THEY THINK!

    The answer to the problem has been presents on at least more than one occasion, i.e. Database specific search screen or ignoring search fields when a database is not indexed on those fields.

    A simple spreadsheet will show what is needed for each database. Use that to build the search screen.

    I am glad that someone is finally listening! Your appearence and blogs decided me to give Ancestry one more year (my old subscription expired August 2). If, however, I see no BIG improvement in the comming year then next August I will, with regret, be gone.

  11. Jade


    In response to Karen you said:

    “I agree, sorting would be lovely. I know I use it on other sites, but they usually have fewer records in a query and they store their information in ways that search indices usually are not stored. If you do a search that has 100,000 possible results and then you do a sort, it is very cpu intensive, we simply aren’t built for that currently. It’s the large numbers of results that make it prohibitive at this time. That of course doesn’t mean there aren’t solutions, there are. We can start limiting the number of results (which has been requested), and we are thinking about this as we constantly look to improve our search engine. Are there are plans in place for implementing a variety of search engine improvements in the future.”

    Since presently your New Fuzzy search results of erroneous thousands or even more than a million results are presented (in ‘category’ view) by database in order of number of hits, it should be a simple matter to allow the user to sort the listing in reverse: the database with one or two hits is actually much more likely to have a valid result concerning the search target than is the database with 2800 hits.

  12. Jade


    Your initial post here said in part:

    “Some of those improvements might include solutions to these problems that surfaced in your comments:
    *Limit the number of results that you see at one time
    *You want a way to be able to search census records and not see records before the birth date and after the death date. . . .”

    The problem with results often is raw numbers, but these numbers are related to **irrelevance** as to time and place. If New Fuzzy were to direct the search engine according to name, date and location given in the search fields, the numbers would not be a problem. The global search of Old Search has this problem too, with bias toward 1930 US Census regardless of death date. New Fuzzy does not interfere with this bias much, and appears to add one toward BMD vitals in the England and Wales databases.

    Once again I conducted a New Fuzzy search on a person for whom I know exactly what entries exist in the Ancestry databases.

    Search A:
    First name Daniel (left not exact since one database entry is for ‘Dan’);
    Surname Davis (exact since the surname is common enough without adding all the Davies’ and soundex equilavents)
    Birth date: 1760 plus or minus 2, not exact
    Birth place: Sussex County, Delaware (exact)
    Death date: 1816 plus or minus 2, not exact
    Death place: Monongalia County, West Virginia (exact)
    Lived in: Monongalia County, West Virginia (not exact)
    Spouse: Lurana Hudson

    All results were to be sorted by Category.
    1 Public member story
    70 Public Member Trees
    17 Private Member Trees
    4 One World Tree

    Search B: Same as A but added a ‘Lived in’ place, Sussex County Delaware (not exact)
    Results: Identical to those for Search A.

    Search C: Made first name exact
    Results: 1 Public Member Tree

    Search D: Made first and surname Not exact
    Results: same as Search A results except rivate Member Trees now 18 instead of 17, and One World Tree entries 5 instead of 4.

    Search E: Like D but made birthplace Not exact
    Results: 27,254 total !
    64 Census Mortality Schedule Index
    3956 Social Security Death Index
    170 Death Records from AL, IN, MT, NC
    9 Birth Records from CT (Barber Collection)
    7 Military – Civil War Soldiers
    1578 Member Photos and Stories
    22,008 Family Trees

    Search F: Like E but added
    Marriage date 1789 (not exact)
    Marriage place: Sussex County, Delaware (not exact)
    Results: identical to those for Search D


    The entries for this person in the Ancestry database are:
    1) Many silly Tree entries, some largely lifted from posts of mine, some combined with creative use of ‘same name = same person’ giving unproven or erroneous parental identifications.
    2) A useless American Genealogical and Biographical Index entry that derives from a publication of a misinterpretation of an old DAR Magazine item.
    3) An entry in Ancestry’s highly muddled Delaware Marriages database for a 1789 Marriage Bond concerning impending marriage between the above Daniel and Lurana, spelled exactly that way (the database gives it as a Marriage Record, which it is not).
    4) An entry in Ross B. Johnston’s mis-titled _West Virginia Estate Settlements . . . to 1850_ for Daniel Davis listed in Monongalia County. There is also an entry for Lurana here, but of course as ‘Davis’ so would not be retrieved by your computer.
    5) An 1810 US Census enumeration in Monongalia County, WV for ‘Dan Davis’.

    Despite my hints to the computer, it retrieved not one of the database entries other than the tiresome Trees.

    The computer response to Search E, adding thousands of 1850+ *death* items once I made the *birthplace* inexact, really needs to be fixed. Given the death date of before 1818, how could SSDI entries of 150+ years later be valid results? Why would Civil War soldier listings be more relevant with an inexact birthplace, given that no one born about 1760 was a Civil War soldier, regardless of birthplace?

    Does New Fuzzy really disregard all dates and places unless made ‘exact’?

    Why isn’t the option to give more than one place of residence on the initial search form?

    Why isn’t the option to give date and place of marriage on the initial search form?

    With what combination of exact and Not exact data can any of the actual database entries be retrieved (please, disregarding Trees and Member Stories), using the above facts? If some combination does retrieve them, why is such a combination necessary – why are they not retrieved with a straightforward account of the facts in the search form?

    I must say that either it is a fluke or something good has happened to New Fuzzy. I fully expected to get 29,000 ‘hits’ in England and Wales databases for 1830s+, but these were missing from this particular search regimen, along with the immigration entries and photos of Port of New York, New York in the Revolution, and so forth, which were among the totally non-relevant restults returned in some other attempts to use New Fuzzy.

  13. Fran

    I, for one, am very pleased with your list. I think you captured my major concerns. I really appreciate your listening. I understand members concerns about getting long lists of Family Trees, when they are not interested. Many of these results are very bogus. However, on several occasions, I have been able to contact other researchers who have information on possible targets of interest that have been very helpful. Thanks to a tree that listed an aunt, I was able to reconnect with my Father’s mother’s family. I now know the accurate maiden name and lots about the person in the picture that I choose when we cleaned out my parents’ home after my mother’s death. Let the inclusion of trees be a users option.

    The second important item is number four. On occasion, indexes have a blank where one would logically expect information. Race is an example. A blank should not automatically eliminate that person from a search if all other facts match the search criteria.

    One idea might be to allow the user to specify a search range in data sets such a census etc. If you believe that the person was born in 1850 +/- 5 years and died in 1872 +/- 5 years, you should be able to limit the data sets explored to those years. You could enter a search range – 1845 to 1877. In addition, it would be helpful if you want to use a maiden name in a search. The search could be limited to the years before you think she was married.

    Finally, it appears that often searches where you have enter the city county and state for a search, the actual search is done only on the state. I often search for Nebraska location and I have to go through information from all counties (alphabetized) until I get to Hall or Holt County information. Why have us enter the County if it is not used in the search?

    Thank you for listening.

  14. Donna Dallman

    Many times, I may not know exactly where someone was born or the exact date, etc., until I get a clue here or there. However, there are TWO things I almost ALWAYS know = and that is the Century and the Continent!

    And when I know FOR SURE that one ancestor was born in Mass in 1822 = then I don’t want to look at records from 2004 or 1987 or 1998 = and 99% of the results are from ENGLAND!!

    It is even worse when I am looking for someone born in the early 1700’s in Kentucky, etc – and your lists give me tens of thousands of records = all in the 1900’s or 1800’s or even in the early 2000’s – and most records are from Italy or France!

    So = end result = I would just be delirious, if you would only give me records from the RIGHT CENTURY and the RIGHT CONTINENT!!!

  15. Re-type from the relay: “Naturalization details: State or Federal? There are no footnotes here of WHERE the poster got their information from WHAT source! Instead I’ve got to choose which fork in the road. Example: George Z. Singal of Portland, Maine, see my with Record #193408910 that I first thought was an external # to an outside document somewhere, but merely the run-around here of an internal # leading nowhere, and so I give up! Account closed. Please contact me by e-mail: Thank you, JosephsHaas at hotmail dot com , P.O. Box 3842, Concord, N.H. 03302, Tel. 603: 848-6059 (cell phone).”

  16. Steve

    Well, according to the wildly-popular “new search experience” my father (born 1924 in Minnesota, died 1987 in Minnesota) never existed.

    This is a mild disappointment to my mother, not to mention my siblings. I am not at all sure I wish to do a search on my own self, lest I disappear from the face of the earth.

    When I revert to the hoary old search page, results still come up in the “new search experience” format, which is nearly as annoying as the “new search experience”.

  17. Jerry Bryan

    This is my first night on the Internet in a couple of days, so I’m just now getting around to Anne’s message of August 4th. I’m not quite sure what I think about her list.

    On the one hand, her list seems accurate, balanced, and complete. On the other hand, it organizes the information very differently than they way I usually think about it. So it’s a little hard for me to really get my arms around her list. In any case, I really, really appreciate her listening and making the list.

    I’ve decided that this stuff is really hard to talk about just with narrative text. E-mail, blogs, wiki, etc. — it doesn’t matter, it’s hard. I would love to sit down with Anne and watch her use ancestry for an hour or two while we talked about it, and I would love to then have her watch me use ancestry for an hour or two while we talked about it. I think we would make much more progress that way.

    To that end, I have decided to try an experiment. I have created what I’m calling some mini-webinars. Of course, they are not real webinars because they are not interactive. They are just some screen captures of me working on with a voice over. I have created three such files so far, each about five minutes long. They may be found at

    This particular technology is not something I work with on a regular basis, so let’s treat it as an experiment. Sometimes experiments fail. Also, be aware that the little mini-webinars that I have completed so far deal with very general topics. I have not yet even gotten to the point of doing a real search.

    The files are shockwave files, and they play as streaming files through your Web browser so you don’t really have to download them to play them. They work fine for me on both IE and Firefox. If you can’t play a Shockwave file, then they won’t work for you.

    I captured the files with a tool called Jing, but I’m storing the files on my own web site rather than on Jing’s web site. We’ll see how it goes.

  18. Val

    I’ve been reading the comments on this blog lately, and recently I decided to have another go at New Fuzzy (I have used it, VERY briefly before), to try to come to a more educated conclusion about what I think of it.

    Well, frankly I can’t stand to use it long enough for that. It’s really nothing specific, it’s more of that “search experience” they keep talking about. It’s slow and hard to navigate. When I try to refine my search, the search box that comes up only stays up for a moment then briefly disappears and comes back again, I have no idea why. The way it organizes search results is not clear and not very logical, but mostly I have been able to find what I’m looking for anyway. Also, it is extremely difficult to find a specific collection with New Fuzzy. Absolutely infuriating–too, too many clicks. Supposedly this thing was designed to be more user-friendly among other promises but I’d say that’s where it falls the shortest. I have lots of issues with Old Search, but it’s still highly preferable to New Search.

  19. Carol A. H.

    Jerry, I went to your web addresses and watched your mini-webinars, all of them twice. Very good by the way. I agree, the names should always be separated. First name/middlename and then surname. I had noticed how new search doesn’t always keep them separate but it slipped my mind. Most of the details slip by me and I’m left with a bad “experience” but find it hard to define exactly why it was bad except that I got lots of hits but no good results. Sort of like a headache with no known cause. I know I mentioned in a blog that I have had several different problems and can’t seem to find a pattern. I don’t have the time or expertise to go hunting them down. I guess I should keep a log of problems: what, when and where, but when I have the time to do reasearch, I just go with the old search so I can get some value for my money.

    I previouly reported using Ancestry on two different computers this past weekend, tying old search and new search on both. Some strange differences in results. These computers had different versions of Windows. One had XP and the other 2000 Professional. I have XP at home and there are additional differences. Some are slight, but I think it’s a matter of hardware difference. I’ll try to do better on reporting problems…if I have the time. As it is, I’m spending too much time on these blogs but I don’t want to miss anything. Lots of folks are putting forth a good effort to track these bugs.

  20. Jim Davidson

    I really don’t see the need to change the search process from that now in effect. It certainly does nothing to make my searching any easier.

  21. Dan

    Sorting, sorting, sorting. Absolutely crucial. It would be perfectly reasonable to force a limit on the number of results when sorting; as a web programmer myself, I know that sorting 500 or so results by a specific field is trivially inexpensive as far as CPU is concerned.

    Yes, 100,000 results sorted by a criterion would be rough on the servers, but I’m sure most users would be more than happy to accept the limitation that only the 500 most relevant results be included once you decide to sort by one column or another.

  22. Dan

    Two more quick comments:

    First — I find it really frustrating that the new search interface has the actual search form surrounded on both sides by other content. Much much nicer to have a single, main content area with the search fields, and then a single sidebar to the right with the extra content.

    And second — I agree with Connie about the silliness of defaulting to a worldwide, all-databases simple name search. It makes sense to have something like that on the home page for non-subscribers, maybe. But I doubt many folks who are spending hundreds of dollars a year to use this site have any interest in such searches. Keep the simple search on the non-subscriber homepage if you’d like, but it doesn’t make sense for paying users.

  23. Mark

    After reading a number of comments to this searching issue, I see I don’t need to state my own. They seem to cover most of them…dates and names not related to what I entered in the search box and so on.
    I’m not noticing anything different in the search page content…did you improve on it or change it? I do notice that when I type a first and last name, birth date and “all countries” selected and “exact match” is unchecked, I get the exact name but the wrong birth date and when I scroll down past some incorrect names, I get the exact name and exact birth date-very frustrating.
    When I select exact match, I get a different layout of the historical matches with no persons name. BTW, these are both doing a “historical” search.
    I also want to add that I found people on and (both free) that do not show up anywhere on, including “people finder”
    I believe there is some kind of glitch in the search engines.

  24. cris

    I agree with all of the above. The membership fees are way to high for such an inept exasperating search experience. Maddening that Ancestry can’t get a handle on a good search engine. Very annoying that it isn’t possible to do a partial name search. All the clicking that one must do to perform the most simple basic actions is plain irritating. This website stresses me out and the new search has made it unbearable.

  25. Jerry Bryan
  26. Mark

    All I can say is a search engine is useless if you have to scroll thru 14 gazillion results before you find the name you’re looking for. Why are the results in order of type of record and not the name of the person. The person’s name should be first and foremost whether I check exact or not and then the birth date should come into play.
    I also got a yearbook result that matched the exact name-first, middle and last- after, of course, scrolling thru some incorrect names, and when I clicked the yearbook image, the names were highlighted, which is good, but they only matched the first name here, middle name over there and last name somewhere else…none of the names were combined to form the original name in the search. That’s a useless search result. It should state that it found a first name match, a middle name match and a last name match…not necessarily link as one name.

  27. Jerry Bryan

    Here’s an error for the programmers to look at. In New Search, go to the Tennessee State Marriages, 1780-2002, and do any exact search on Sullivan County, Tennessee, USA. The search will turn itself into one of those weird fuzzy searches that I talk about, even though you specify exact. The error happens even if you select Sullivan County from the type-ahead drop down box rather than typing it in by hand.

    I research Sullivan County rather a lot. The courthouse was destroyed by a cannon ball in the Civil War. Marriage records are extant from 1863 to the present. But the ancestry database does not include any of the Sullivan County records, even those after 1863. I’m sure it must be the absence of any records at all that is causing the exact search to go haywire.

    In Old Search, Sullivan County is not in the drop down list of counties at all. Therefore, you can’t even start a search for Sullivan County, and the search can’t go bad.

    It would be really nice to have the post-1863 Sullivan County records added to the collection.

  28. Frances Allison

    Unless Ancestry can come with a SEARCH ENGINE which pays attention to the information typed into the form, it doesn’t matter which version is used. When I enter the name of someone born in 1820 in Pennsylvania I don’t want to sift thru the 1930 Census or a marriage in England in 1920. SEARCH ENGINE is the biggest problem!

  29. Jerry Bryan

    Here’s a particularly egregious example where New Search gives gazillions of hits for an exact search, but Old Search only gives one hit: Howard Kingsbury Cross.

    Cross can be a particularly difficult surname to search with automatically indexed articles because “cross” appears often as a regular word. For example, an account of a military campaign might mention a unit that will cross a river or a creek. But you would think that the inclusion of “Howard” would eliminate many of the false positives and that the inclusion of “Kingsbury” would eliminate the rest. But apparently not.

    I think it would be a reasonable goal for that an exact search for Howard Kingsbury Cross should provide only one hit. That would be a great benchmark for the new software to try to achieve.

  30. Jerry Bryan

    I repeated my search of Howard Kingsbury Cross, adding birth and death dates of 1895 +-2 and 1980, thinking that this might make things better. It didn’t. If anything it made it worse. Remember that this was an exact search. It was acting a lot like fuzzy search. I promise I really had the search marked exact, and I had only one window open. I didn’t do anything that I can think of that would have persuaded the search to go into fuzzy mode.

    In Summarized by Category, there were lots of databases with hits. Just for example, I chose WWI, WWII, and Korean War Casualty Listings because there were only 54 hits.

    The hits included all manner of first names that were neither Howard nor Kingsbury, such as Albert B. Cross, August Barton Cross, etc. No amount of refining the search with different first names would change the 54 hits. It would change the rankings (and remember that this was an exact search), but not the total number of hits. Removing the birth date didn’t change anything. Finally, removing the death date reduced the number of hits to zero. So it’s like the presence of the death date seemed to throw the whole search into fuzzy mode.

  31. Stella Nemeth

    I tried out the new search feature yesterday. I tried it on my father. I used to be able to locate documents for him, even though he had a common name. Now I can’t find anything on him.

    Why isn’t there any way to use the EXACT name? When I looked up information on a cousin, I found information of a different cousin. The only thing the two of them had in common was an unusual first name and last initial. I doubt if they are even related to each other.

  32. Reed

    Dear Anne,

    First of all, a tip of the hat to you. You are the first employee/blogger to actually respond to (at least some of) the issues posted by blog readers. Your several responses in less than 2 weeks makes you the most responsive Ancestry employee—by far—in my 2+ years experience with the company. Keep in touch, we appreciate it.

    Hats off as well to the many fellow-researchers/bloggers who have invested hours in investigating, documenting and posting problems with New Search (and Old Search). It’s bad enough to experience these problems in the course of one’s research. Documenting them in comprehensible, readable form takes a lot of time and effort. (Bonus points to Jerry Bryan; I don’t know how you find the time—or patience!—Jerry.)

    Since the previous New Search discussion thread is now closed, I want to reiterate here my main issues with Ancestry’s search process. Anne, please ask yourself if, compared to Old Search (with all ITS many problems), when you use New Search do you:

    (1) get more (and better-quality) hits?
    (2) get fewer meaningless hits?
    (3) find New Search’s interface more logical, more ergonomic, faster, and easier to use (and can be customized to turn off patronizing pop-ups and distracting advertising—especially animations)?
    (4) find it easier to sort and resort search results?
    (5) find that the wildcard functions (*) and (?) work with less than three initial characters?
    (6) find that Exact searches really give exact answers with easily set (and altered? and Boolean?) parameters?
    (7) find, in general, the time you spend on Ancestry is more productive and more user-friendly than before?

    If you answer YES to these long-standing customer requests, then please show us how you get such results, because New Search is NOT meeting these criteria in our daily experience. As amateurs and professionals with a passion for genealogical research and high-quality data, your users enjoy well-designed software improvements (that actually WORK) and we pay a lot of money to Ancestry for access to well-indexed, easy-to-search primary and secondary sources.

    New Search is a step in the WRONG direction—PHILOSOPHICALLY, TECHNICALLY and ERGONOMICALLY—and most of your users and blog-commenters agree.

    Furthermore—for a variety of reasons—I have NO interest in creating, saving, or posting Family Trees on If New Search is inextricably linked to creating and using trees on the Ancestry site, then New Search represents a huge step backward in functionality for me and other researchers like me.

    Finally, I agree—emphatically—with the other commentators that have discussed the wrong-headedness of the HYPER-EXACT literalness of New Search.

    For example, if I’m searching for “John Smith” and put in a birth date of 1820 (exact or not) and a death date of 1875 (exact or not), then I SHOULD get every John Smith on every US Census from 1820 to 1870. And I should NOT get census results from 1790, 1800, 1810 or results from the 1880 or later censuses. It should not matter that census returns do not have death date information; by giving a death date to the Search engine I am giving it a useful search parameter (the “terminus ante quem”) that should not be ignored just because the census databases do not have fields labeled “Death Date.”

    And similarly, the Auto-Fill function is over-exact to a mind-numbing degree. Here’s an example from my own research, looking for Levi S. Baker (1827-1910), Lived In (Residence): Chicago. Keep in mind that Chicago is a city within Cook County in the state of Illinois, USA. Chicago was founded as a town in the Illinois Territory in 1833 (although the name appears on a French map or two as early as the late 17th century).

    Now, using New Search, if you just search for “Levi Baker” Lived In (Residence): “Chicago” you get 108,526 hits in dozens of databases, most of which are useless (Please explain the relevance of “English Settlers in Barbados, 1637-1800” ???).

    Change the “Lived In (Residence)” from “Chicago” to “Chicago, Illinois” and you get 76,538 hits (most still useless).

    Change the “Lived In (Residence)” from “Chicago” to “Chicago, Cook, Illinois” and you get what I assume are still the same 76,538 hits.

    Change the “Lived In (Residence)” from “Chicago” to “Chicago, Illinois, USA” and you get 89, 836 hits (most still useless).

    Just for fun, try searching for Levi Baker, Lived In (Residence): “Cook County, Illinois, USA” and you get 22,056 hits (most unbelievably useless). Just using “Census and Voter Lists” results as an example: NO US Federal Censuses are included, but these (and ONLY these) ARE:

    •Minnesota Territorial and State Censuses, 1849-1905
    •1828 New South Wales, Australia Census (TNA Copy)
    •Kentucky Census, Reconstructed, 1790
    •Families of Cabarrus County North Carolina 1792-1815
    •”Second Census” of Kentucky 1800

    Now, go back and try this Exact search using Old Search:
    “Levi” “Baker”, dates 1820-1910. place: USA, Illinois.
    And look at that! 83 hits, almost all of them plausible, and including the vast majority of sources that I have found regarding “my” Levi Baker, and hardly any “junk hits.”

    End result? I don’t care what your “customer surveys” and other marketing research show, New Search is a stinker, and needs to be re-thought from the ground up.

    I’ll be offline for a week or so, but I look forward to reading the responses from you and the other reader/blogger/users.


  33. Carol A. H.

    Three cheers for Jerry Bryan, maybe more! He has taken the time to document so much for us. Be sure to check out all 9 of his mini-webinars. Straight talk-no hype.

    I went to Jerry’s home page today, via Rootsweb, and found so many “goodies” that I spent a long time there. He has an incredible list of good places to go. I was able to get into some Eastern US places AND, I got into the page for Harvard University Alumi and FINALLY found the exact year my father-in-law graduated. I was so stoked by my good luck that I got bold and got into the MIT archives for my grandfather-in-law! I have emailed them requesting further information.

    So with all these good places to go on Jerry’s web page, I have bookmarked it so I won’t have to try to remember so much. I’m not stupid, I just have these senior moments and can’t always think of where to go next until I let things jell a bit.

    I did have some good luck with the old search today. Reed has a good point, old search is not perfect but it sure put the new search to shame. Why should I use (and pay) for somethng that makes me crazy because it doesn’t work and it’s too mouse intensive. I don’t need any more “experiences.” At my age, life is just one big experience!

  34. John

    Another question about search and everything else: When I copy something andp aste it into FTM, all formatting is destroyed. Then I have to go back and try to put in all the spacing, punctuations, etc.

    Why can’t this be as adaptable as Version 16?

  35. Donna Aho

    My ancestry records are from Jamaica and Mayaguana the Great Bahama Island as known back in late 1800s. Can’t find my grandfather Cumberland or Robert Saunders and grandmother Mary Thompson who were born and raised in these places and were of British descent. Their son, Robert Joseph immigrated to MA, USA around the turn of the century and was an oral surgeon as most of his ancestors were doctors. I wish I could find these records.

  36. Donna Aho

    Does ancestry plan on listing cemeteries or burial places possibly churches or temples besides the LDS sites? I am trying to find burial sites of ancestors and all I have is country last resided.

  37. JMaule

    I think I know why everything has crashed 2 evenings in a row. In an attempt to “jazz up” search, someone has messed up the image databases. Why the need to change search? It was fine the way it was. The answer, of course, is that someone sits at a desk thinking, “What can I do to justify having a job? Oh, ok, I’ll change how search works.” It’s no coincidence that the problem with the image databases being down (from what I’ve read on the blogs they’ve been crashing throughout June and July while I was away) matches up with the tweaking and bedazzling work on the site. Here’s a suggestion: ditch the new “improved” search (which isn’t improved after all) and restore what was there a few months ago.

  38. Mark

    I don’t know if any one has touched on this but why does advanced search do away with the middle name entry? And why does everything that was entered before I decided to do an advanced disappear? I have to reenter everything again.

  39. Nancy Rogers

    I just tried to go a specific census year from the short menu I started with census records and then went to the next page and when I did I got the following message “Database id or database name specified incorrectly or the database is no longer valid.” What is going on with this?

  40. Ellie May

    I know this isn;t the best place for this… but I am trying to get on to the site and instead I get a database “runtime error”. It worked fine last night!

  41. Dave

    I have not been able to view any search results – I keep getting the message “Database id or database name specified incorrectly or the database is no longer valid.”

    Can someone tell me what the problem is?

  42. Jerry Murdock

    Anne, you mention the problems with sorting 100,000 records in a search result, but if you listen carefully to what your users are saying, the result will be much smaller. If my ancestor lived in “Texas” (exact) from 1834 to 1885, DON’T give me records from Yorkshire, UK in 1941. If I give you the “exact” given name Elam F., don’t give me records for some guy with the initials F.W. (especially as #1 MOST relevant ABOVE the “exact” name I gave you). ESPECIALLY limit the locations and dates to what I give you. LET ME worry about mistakes in what I KNOW to be fact. If I have any doubts, I can tweak my input MUCH better than your fuzzy search can do, because I know what I may have wrong (HINT – NEITHER the continent NOR the century).

    As an afterthought, I almost wonder if you may be using dates as “text” rather than “numeric” fields in some cases.

    Jerry Murdock

    records from

  43. Jerry Bryan

    I’ve still been thinking about Anne’s list of problems. I think I want to make one amendment to my comment that I think the list is complete. Anne’s list includes the following:

    I believe that the biggest problem that has been discovered is that the detailed searching on some datasets is not available for the new search interface. For example, on a marriage record, you can’t specify a groom’s surname. This ability to use your knowledge of the data that is stored in any given dataset is hampering your ability to pull out records in a manner useful to you. I think this is a problem.

    I’m not sure I understand which problem is being identified. The marriage databases I most frequently search allow both spouses to be specified. Is this maybe he problem where the search box for the second spouse in a marriage combines the first name and last name into a single entry area? That’s the only thing I can think of. But it’s not just marriages. There are other places where first and last names get combined into a single entry area. In my opinion, first and last names should never be combined. However, I may be misunderstanding the problem to which Anne is referring.

    I think the biggest problem I have identified is not on Anne’s list at all. I’m not 100% sure because Anne and I tend to categorize some of the same problems in a little different way. I would actually list two “biggest problems”, but I suspect that the two problems that I would identify are really two sides of the same coin.

    The first “biggest problem” is that exact search does not return enough hits (many fewer than did exact search in Old Search) and that fuzzy search returns far too many hits (pretty much the same as Old Search). This way of describing the problem actually encompasses several items from Anne’s list, such as death dates eliminating census entries from an exact search. But the flip side of the coin is that paradoxically, exact search frequently seems to return far too many hits because it frequently behaves as if it has converted itself into a fuzzy search.

    Converting exact searches into fuzzy searches may not actually be what’s happening internal to the search, but that’s what the results of the search look like. I have previously posted several examples of how to make this problem happen. But the real underlying cause is not clear to me. Whatever the cause, converting exact searches into fuzzy searches renders New Search completely unusable when it happens. Until this bug is fixed, I regret that I have to give New Search a grade of F. I can’t even think about whether it deserves an A or a B or a C or whatever until it least does whatever it does reliably.

    I hope that Anne will comment on this problem.

  44. Mark

    I’m wondering why no one is responding to any comments. Is this the place to report a technical problem because I get no results from the help section on the main site.
    What kind of advanced search has no place for a middle name but I can enter one in the basic search? And when I hit advance search, everything disappears. This is ridiculous. Good thing I only paid for one month.
    Anybody home?

  45. Jerry Bryan

    With reference to my note #46, I put together a little mini-webinar to illustrate the problem of an exact search morphing itself into a fuzzy search. I’ve previously posted several narratives about the problem but I think this is the first mini-webinar. The mini-webinar may be found at

    It’s a little more than three minutes long.

    I didn’t say so while doing the mini-webinar, but I had freshly rebooted my machine and had only one window open, namely the browser that was logged on to

  46. Carol A. H.

    Just a quick note and then I’ve got to go to bed.

    I watched Jerry’s latest mini-webinar(twice)and noted the numbers. Then I tried everything as he said he did it. I got exactly the same numbers, 14 with the old search, 21,448 with the new search. And I had the same “morphing” situation: 150 results in the voter-census stuff that changed to 997.

    I don’t have a set-up to watch my cpu, but I do have XP.

    Weird! To say the least.

    I have tried some old searches and then new searches with the same people I have in my tree on ancestry and I get different results. I also get different trees right up top and I don’t want trees! And my whole system is sooooo slow with new search, I could go take a break while I wait for it to do it’s gymnastics.

    Got to go to bed, have to go to work tomorrow.

  47. Ron Lankshear

    Just a quick note on Jerry’s
    “believe that the biggest problem that has been discovered is that the detailed searching on some datasets is not available for the new search interface. For example, on a marriage record, you can’t specify a groom’s surname. This ability to use your knowledge of the data that is stored in any given dataset is hampering your ability to pull out records in a manner useful to you. I think this is a problem.

    I’m not sure I understand which problem is being identified. The marriage databases I most frequently search allow both spouses to be specified”

    The problem which I think I originally posted and Anne included in her summary is that Yes some panels for a database have special panels which allow extra fields

    Example UK Pallots marriages which has 1st spouse and 2nd spouse fields.

    presumably New search is replacing these special panels as it appears as a Sidebar fixture

    If that is so then a lot of analysis needs to be done so that special search panels are not lost…

  48. Ron Lankshear

    The type Location feature is a problem

    For UK census the place born field is free form. It is best estimate by the people raising the form.

    Example in 1871 census
    John Lankshear Honor Swyer abt 1844 Hooke (Hook), Dorset, England Head Broadmayne, Dorset

    The type head gives
    Hook, Dorset, England, United Kingdom

    which does not find him

    manually making it Hooke eventually found him

    entering Hook* did not find anything – Please ensure wildcard work.

    Type Ahead implies it covers all records – ie all records will covered by the type ahead list. This is obviously not so and is misleading. I suggest remove. Well at least for place born in UK census. Perhaps it is a complete list for Residence in UK census. These were set by the authorities.

  49. Ron Lankshear

    The result list has a frame around it. When large font icon clicked some of data in last column of results disappears under the frame.

    Example UK census 1871 with EAGLE – only the first two characters of residence can be seen.

    There is plenty of “real estate” on the page to simply increase the size of the frame.

    Also I like the Old search list results as final item was click to see census image. Now I can no longer see this – is it lost under the frame? If the new design expects we to view record then open image you are wasting my time.

    As also is the search sidebar which requires me to click open each item to change place born etc etc

    Old search I can see all the options.
    Oh and new search method is also strange as when I change say Place born the Old one is still there in the box immediately above so that I double take as to what i have entered.

  50. Jerry Bryan

    Here’s a quote from Ron Lankshear’s #52.

    Also I like the Old search list results as final item was click to see census image. Now I can no longer see this – is it lost under the frame? If the new design expects we to view record then open image you are wasting my time.

    I have seen the same problem, but I have never reported it because I cannot repeat it reliably. That is sometimes with New Search I get a list of hits and the hits do not have the links to the images as described in #52. But I can then at some point in the future repeat the same search in New Search, get the same hits, and the hits do have the links to the images. I can’t figure out what, if anything, I’m doing differently so that the image links are there some of the time and not there some of the time.

    It’s not a fatal flaw, just a very irritating one, because you can always get to the images by first clicking on View Record, and from the record view you can get to the image. But the extra click should not be necessary.

  51. Nancy Rogers

    Are any of these problem going to be addressed? I went back and looked at the last response of Anne Mitchell which was dated August 5th according to this blog. We have herd nothing since then unless people have gotton individual emails from ancestry. I am not sure what ancestry is doing to the system, but when I run into problems getting one of the US census records to come up, it concerns me greatly because in many ways the census records are among the most important on line data bases we have.

  52. Anne Mitchell

    Again, thanks to all of you for your insightful comments.

    I’ve been working on comparing and understanding the differences between the results generated by the old and new search interface, and trying to determine where they are different and why. Tricky business trying to learn a whole system, and unfortunately I am not doing it as quickly as I would like. 🙂

    I’m a little unclear on what you guys are refering to on #52 and #53. OK, I’m a lot unclear. A little help please?

    I’m curious (#41) which template did you see the middle name with a specific box just for that? If you type a first name and a middle name into a name text box, the search engine searches for both, which can be very useful since many people seem to refer to themselves by either. I know my great grandfather never could quite decide whether to use Wyatt or Paul as his given name.

    Oh, and btw, the changes to the search interface had nothing to do with the image database crash (which has been fixed). Totally different systems, totally different servers.

    Over the next few days (it may take a few weeks!), I’ll start posting some discussion on specific queries and what if any differences there are between old and new, and why. I suspect we may have a few interesting conversations.

  53. Scott


    I’d guess that we aren’t going to get a response back from these folks. Surprise, surprise, huh.

    I’ve just been using the old search as it’s much less of a headache.

  54. Scott


    It seems that checking “Exact” still returns soundex search results. As an example:

    1. Use the old search, exact spelling, for “Charles Bonmot” for “All Countries” with no date range specified. You’ll get one hit, the 1860 Mississippi census.

    2. Try that same search on the new engine with “match all terms exactly” and you’ll get 8,799 hits. All of them are soundex variations of the surname. Oddly, the 1860 MS doesn’t even show up in the search results.

    Good luck,

  55. Anne Mitchell

    Scott I would agree with that assessment. And I would agree we need to get an Exact/Soundex option back in the UI.

  56. Jerry Bryan

    Anne, on #52 and #53 if you get a list of hits, most of the time there is a View Record link at the beginning of each line (on the left) and there is a View Image link at the end of the line (on the right).

    Well, you only get the View Image link for databases that have images, and not every database has images. But suppose we are looking at a database that does have images. The View Record link is always there at the beginning of the line. The View Images link at the end of the line sometimes is there and sometimes is not. I have not been able to make the View Images link “be there” or “not be there” on demand. It’s simply not repeatable whether it’s there or not. Or maybe it’s repeatable and I just haven’t figured out how to make it repeat. I don’t know.

    So I can’t send you a good example. The next time it happens that the View Image link is not there, I’ll try to capture the screen and post the result.

  57. Ron Lankshear

    Hi Anne
    Similar to what Jerry said re 52.
    I am looking at UK census. With old search I get a line per person and at right end there is a click to view Image.

    With new search the image on right does not appear. I just wonder if it is hidden under the frame.

    Yes I can click View Record and then find the image BUT why Ancestry remove function?

    The old search made good use of “real estate” on the browser – the new search does not. New reduces the output width to around 80% compared to Old. This means that fields (place born) are wrapped into 2 to 3 lines whereas 1 line in Old. This means more vertical space is lost. And not as easy to read/scan

    I have not noticed 53 – images being lost altogether

  58. Ron Lankshear

    New search large hit lists.

    New has less function to move around large search lists.

    I have hits set at 50

    Example 1871 England census – search for EAGLE. Result 1280.

    OLD search there are clicks for pages 1-10 and a next. This is at top and bottom of results

    NEW has nothing at top and at bottom of results has page 1 2 … 12 then an > (for next)

    as you page through the items change

    I would like an ability to move quickly to middle or end of any result list.

    To overcome limited function of the three letters before wildcard I often do a search with blank for a first or surname and then page to the initial letter. New Search is worse to use for this than old

  59. Ron Lankshear

    Three Letters before Wildcard

    Yesterday I was looking for a name HOUNSLOW. Those initial vowels can get very badly written and are hard to get right in an index. Some versions appear as HANNSLAW.

    It would be good to be able to search for H??NSL?W

    Similarly SKEGGS can come up as SKIGGS SKOGGS SKAGGS etc – being able to search for SK?GGS would be good

    Unticking Exact search on surname as SKEGGS and entering Canada 1863 and ticking exact on Elizabeth does find her as 25 out of 27 all with 2 stars. Surely Elizabth 1863 and Canada should give her 3 stars not 2 and SKIGGS Being so close to SKEGGS should be a half star

  60. Ron Lankshear

    My 63 lost a paragraph when I submitted – the final paragraph is abut my grandmother Elizabeth SKEGGS whose surname is spelt SKIGGS in 1881 census.
    2 star was also given to names like SACE SACH etc

    LDS search picks many variations when searching for SKEGGS. These seem to be very viable alternatives.

  61. Carol A. H.

    Ah, Anne you are there…so glad.

    One of my problems is the extreme, and I really mean extreme, slowness of anything I do in new search. Do we need to upgrade our computers and what should be upgraded? It seems I noticed this more lately. If I try to use FTM 2008, it is even worse, so I don’t use that at all.

    I don’t mean all the extra mouse clicks (that’s another problem), but everything is so slow. Any suggestions would be welcome. I have a high speed cable connection Thanks.

  62. Carol A. H.

    I did want to mention, I watched the July 30 Webinar by Kendall Hulet again, and it looked like it SHOULD work, but….as I once posted it doesn’t, and the problems are not consistent.

  63. Ron Lankshear

    FreeBMD search

    Ancestry has a copy of the FreeBMD index to UK GRO index.

    In FreeBMD web site itself I can search for marriages with the names of both spouses

    Example LANKSHEAR to SHAW in 1838.

    This does not work in Old Search or New.

    I assume Old Search was only programmed to work for marriages after 1911 when the spouse name started to be shown in GRO index.

    So I tried EAGLE and WALSH in OLD search and yes it came up in 1916.

    New Search did not find it.

    The FreeBMD spouse search before 1911 works based on potential spouses having same GRO and Disrict references. This is only available in Ancestry by finding one of the names and then View Record and click page number to View Others on Page. You can also do that in FreeBMD but they also have a search box for spouses name.

    Having the field there is misleading if it will NEVER find anything either because the data is not there or search program will not look.

    I did check a birth search in freeBMD with mothers surname for a 1913 Eagle and Cowell event. But of course it will not work before 1911 so is confusing to people who do not understand the databases.


  64. TarHeelz

    Bravo. The comments on this thread are detailed and dead-on. The current search functionality is so useless that I wonder whether the authors of Ancestry have ever tried to find, umm, anything, with these algorithms.

  65. Mark

    In your comment #55, I found the middle name option…I was using the old search which didn’t have one in “advance options”. I switch back and forth but I believe my other comments still hold true.
    And as #67 states–
    “Bravo. The comments on this thread are detailed and dead-on. The current search functionality is so useless that I wonder whether the authors of Ancestry have ever tried to find, umm, anything, with these algorithms.” ….
    …I say, hey, simple solution-go back to old search (since it’s, umm, still there) and get the bugs out of THAT one since THAT one works better.

  66. Tony Cousins

    #65 – Carol, the configuration of your home PC will not have any impact on the speed of the actual search. It may be slow on displaying the search results or there may be another issue.

    Imagine – all these people trying to use the new search and all of them trying to retrieve ‘gazillions’  of entries, that has to have an impact on the search servers, I know it does in the applications I monitor in my daily toil.

    Anne – Question #1 – has the load increased on the search servers? Or is there another reason for the apparent decrease in performance?

    #68 – Mark, I seem to remember in a previous post that we were told that the search engine is the same in the new and old versions. I think the problem is most probably how the new search front end presents the user inputs to the back end engine. That theory seems to make sense given the number of posts we’ve seen detailing different results given the same input search data to old and new.

    Anne – question #2 – is it the same search engine behind two different front end screens?


  67. Karen

    You want a way to be able to search census records and not see records before the birth date and after the death date.

    Not just census records, but any records. If an exact birth (or) death date/location is specified and the data base doesn’t include specific fields for that information, it should be used the same way that “also lived in” is used.

  68. Leslie

    I would really like the search engine to use the “exact” search options when searching in the U.S. Federal Census Mortality Schedules Index. I specifically picked the ‘exact’ on each part of the name and the death year (1860), I gave the system a specific place and it returned over 2000 hits. I expect the real number to less than 20 if it is working correctly.

  69. Jade

    Leslie #71 has discovered part of the problem that is in Indexing, not in ‘search’.

    Various databases have been indexed in various ways.

    In particular, newspapers, the California Voters Registration Lists and the England and Wales BMD Indexes (to mention a few) have been indexed word-by-word, not by Surname-and-first-name.

    In these databases the ‘exact’ search function, whether Old or New Fuzzy, locate pages that have both words on the same page. A keyword search, not a name search.

    Thus in the California Voters Lists, exact search for ‘Arnold Moore’ will retrieve every page that has ‘Arnold’ and ‘Moore’ on it, such as Arnold Drive and Moore Avenue, or Arnold Schwartzeneggar and Margaret Moore.

    Since indexes of the England and Wales BMD indexes have not been completed, now is a good time to change the indexing protocol to a surname/first name protocol instead of word-by-word.

  70. jessie sorensen


  71. Carol A. H.

    For Tony C. : Yeah, I thought of that, the number of people trying to do searches at the same time. I have noticed that late at night or very early mornings (Pacific time) things go more smoothly. Not many people up at 1 and 2 in the morning here, and too early for Australia. But it would not hurt anything to add more memory, except my purse which is pretty thin right now, so I guess it will have to wait. But I could keep track of the problem times.

    For Jade (who always seems to have a clear idea and can express herself so well) I also noticed the strange indexing in newspapers, and city directories, but could not say why. You made it clear to me because when I use a first name and a surname, my results are all over the place. I get initals for first names and surnames, too. I can’t figure why I get initals!

  72. narleo

    How come I can not call up a name that I have already printed up – Say I got a SS Death Index name before and now you cannot find it.

  73. Jade


    It was mentioned on the next version of this blog that on Wednesday it was discovered that there was a problem with doing exact-name searches in some databases. They did not say what databases were involved, but said feverish work was being done to find and solve the problem.

    When you notice a sudden problem such as this, it might be useful to contact Ancestry at the email address that you can find through the “Help” link.

    Good hunting,

  74. I also want to comment on the search function of Ancestry. To put it simple, it sucks and doesn’t work the way that it should. The returned results are simply terrible.

  75. Margaret Janco

    I agree with all of the previous criticisms of the “New Search”. My impression of the “New Search” is that it assumes that we don’t really know what we’re looking for, and that the computer knows better. But, see, I almost always know exactly what I’m looking for, and I’ve used the site long enough to know how to get around errors in transcription, name variations, etc.

    After trying the “New Search” I did complete the survey, but I didn’t feel that I was able to express exactly how bad I think this “improvement” is.

    The only changes to the “Old Search” that would have been helpful to me would have been to enable the use of “?” in name searches, as Ron showed in #63, and a change back to the previous way of searching the Newspapers. Previously, the initial searchof the newspapers involved specifying a location, which produced publications produced in that location, and a date range, which produced hits in papers PUBLISHED during the specified time period. Currently, one can search using both these criteria, but instead of “date and place of publication” hits, one gets hits within all publications in the database,with the limitations of the optical sacanning of these publications, which include a lot of junk, difficult and time consuming to wade through. My only alternative is to pull up a list of all newpapers availble in a particular locality and search each one individually. The newspaper search is no fun anymore.

    Rather than spend time and money developing the “New Search”, these resources could have been better utilized tweaking the transcriptions, including “municipality” rather than “village” names in the censuses databases; improving the wildcard function; acquiring new info; and fixing things that you promised in the past to correct but have yet to accomplish, such as linking the correct page 2 to page 1 of the WWII Draft Registration for Pennsylvania. Oh yeah, and providing real people with real answers to our individually sent specific inquiries, instead of canned answers to general questions.

  76. Stan

    I have been a Ancestry subscriber for many years but this latest change is pathetic. What happened to the list of familiy and county histories? Why do I neet to scan thur census and birth records to find a death record? I get Heritage Quest for free! Why do I need your service?

  77. Ron Bestrom

    I do not like the new Search screen.

    I do not save “cookies”, so EVERYTIMES I login I get the new Search screen. I am FORCED to enter the name “Jones”, to get to the next screen. I THEN can hit the link to use the old search method.


    Ron Bestrom
    PAID MEMBER for 8-9 years?

  78. Nicole


    You are correct that sorting large datasets is a resource intensive task — either for the database server(s) if the dataset is sorted there and then returned to the web server(s) or for the web server(s) if they are sorting raw data.

    However, the database server(s) constantly processing the enormously large (and mostly garbage) datasets produced by the New Search is sucking up far more of your server resources than an occasional sort query.

    So the complaint about server resources being the impediment to sorting data doesn’t hold water — the behavior of the entire New Search shows an utter disregard for those same resources.

    Additionally, the dataset IS being sorted, it’s just being sorted by predetermined criteria. Permitting the users to specify a sort order at the same time as the initial query — combined with proper database indexing — would have minimal impact on the server resources over sorting in the default manner.

    I do not believe only permitting the sorting of x number of “most relevant” results would be beneficial in many cases. The relevance ratings assigned by Ancestry are often horribly flawed, as any user who habitually gets pages and pages of highly ranked results from other continents and centuries will tell you.

Comments are closed.