Ancestry.com

Let’s talk about search!

Hi! My name is Anne Mitchell and I’ve just joined the team here at ancestry.com as the product manager for search. My previous job experience has been both in product management and engineering at companies including webshots.com and excite.com. And in case you’re thinking, “well that’s nice, but does she actually do genealogy, does she actually use search?” the answer is a resounding yes. I drive family and friends a little crazy in my quest to find my ancestors, and ancestry.com has been a very important part of that quest. So you see, I have a vested interest in making search the best it can possibly be.

I’ve been reading your comments over the past few months and I have an idea of what you think is working and not working. I’m hoping you will continue to share your thoughts on what works and doesn’t work with us. This is my first week on the job and I’ve spent time talking to our search engineers and the rest of the team trying to get an understanding of how the system works, and what is being worked on. There is a lot going on here, and it will take me a little while to get completely up to speed so I’m hoping you can be a tiny bit patient while that happens.

The new search interface is the hot topic of conversation. It’s not a new search engine; it’s just a new way to interact with the search engine. And yes, I know, it’s not always behaving in a way that you expect. But it’s a never-ending process of improvement for us. There are many plans in the works to improve the interface and the backend engine to bring you better results.

I’ve been reading the posts and some of you have done a good job of documenting searches and unexpected behaviors in the search results with new search interface and the old search interface. Now, I can’t address all of them in one posting, but I am going to starting pulling a few at random and try and dissect them here so that we can make some progress.

The case of John Cousins

So let’s begin with a posting by Tony C about one of his ancestors John Cousins. Here’s the snippet to get us started:

Name – John Cousins – exact
Born 1816 +/- 2 years – exact in Suffolk, England – again

and
Now add the fact that he died somewhere around 1877, again in Suffolk, England – exact all you get are family trees.

OK, so this is what I tried in the new search interface (click on the image to see the results):

John Cousins - search #1

And sure enough, I get family trees. Not there is anything wrong with family trees, but we would like to get some other documents. The issue lies with “exact”, specifically, the location, and how the data is stored.

So let’s back off on some of this “exactness” and see what happens. Unclick the exact on the locations:
John Cousins - search #2

Ahhhh…..that’s better.

Now I’m seeing census links and BMD records. So what happened?

Back to the new search interface. Why is exact location wrong? Then it hits me, a census record doesn’t know the exact location of death. How can it? John Cousins is still alive when the census was recorded. What if I just try “unexacting” the death location:

John Cousins - search #3

I see the census records. Interesting. But no BMD records.

So I try the opposite:

John Cousins - search #4

So what does this tell me. Exact death location will prohibit me from getting a census record. Does that make sense? I think it does. A census record does not contain death information so it shouldn’t match.

In the old search interface, your “exactness” was much more general, and it was sort of hit or miss in how it worked. In the new search interface, you can be very specific, and that ultimately gives you a lot more power in refining your searches. Which can a very useful tool when you are trying to wade through a bunch of records.

Now I know some of you are sitting there thinking, “Um, Anne, I really don’t want to think that hard, I just want to see the records I need. Old interface search gave me what I wanted.” I think maybe some of you are luckier than I am, because the old search interface was NOT always giving me what I wanted. But you probably don’t want to hear my long journey through the western Virginia census records. I ended up having to read through them to find what I wanted (which is sort of fun anyway). I just couldn’t be specific enough to find what I needed. I am finding some better results with the new search interface.

Over the next few weeks, I will try and pull up more of the examples you all have had trouble with and see if we can’t figure exactly why the new search interfaces produces what it does and if it is reasonable.

And I’ll show you some of the successes I have had as well, and hopefully some of you can share those with me as well.

Hang in there, we are going to work on making search on ancestry the best it can be. Stay tuned.

About Anne Gillespie Mitchell
Anne Gillespie Mitchell is a Senior Product Manager at Ancestry.com. She is an active blogger on Ancestry.com 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, and is currently on the clock working towards certification from the Board for Certification of Genealogists. You can also find her on Twitter, Facebook and Finding Forgotten Stories.

49 comments

Comments
1 B. PriceAugust 1, 2008 at 12:49 pm

I generally agree with you, but an “exact” should never disqualify a source that doesn’t have the information. That is silly. The more exact the data I’m entering, the more exact the results should be.

Again, the exact field should be ignored in the algorithm of documents that don’t contain it. Otherwise, since all docs contain a wide variety of different information, I would need to try almost every combination of the data that I have with “exact” and not. That is unacceptable.

Beyond that, I’m loving the new search!

-B

2 JadeAugust 1, 2008 at 1:22 pm

Anne,

Nice try.

I most srongly agree with B. Price: for those using New Fuzzy for general searches, they should be able to specify a date that is a *cutoff* for census, but does not exclude database items not containing that date.

You said, “And sure enough, I get family trees. Not there is anything wrong with family trees, but we would like to get some other documents.”

Yes there most certainly is something wrong with ‘family trees’, mainly that nearly all of them are not evidence-based. There are many trees on Ancestry and also in WorldConnect that contain purely specious accounts of various of my ancestors. Some are taken from books whose authors did no records research. Some are just lifted from something whomped together by someone who very selectively believed in ‘same name = same person’.

These constructs are not records except that they record someone’s opinion, or the opinion of someone who long ago created a gedcom file that is now circulating in 8th- and 9th-generation copy. They are stories. They are narly all fiction. By and large they have little or nothing to do with accurate genealogy.

If someone is doing research they should not have ‘Trees’ as search results.

Persons chiefly interested in building trees can search the trees, if all they want is to add some names to a tree. This is recreation, not research, and I really object to Ancestry’s mixing Trees in with records as if they are equivalent. Particularly with the OneWorldTree horrorshow. . . .

3 David WatkinAugust 1, 2008 at 2:00 pm

Welcome aboard.

A document should only be excluded if the contents of that document explicitly contradict what you have entered.

A census entry in 1870 does not contradict a death date of 1877, so “matches” even though the census doesn’t indicate a death date or location.

A census entry in 1880 would contradict a death date of 1877, even though the census does not contain an explicit death date. So, an 1880 census record should not be returned as a match.

Otherwise, an “exact” search becomes impossible.

4 FranAugust 1, 2008 at 3:51 pm

I totally agree with the previous 3 comments. To me what a birth and death date does is put a bracket around a time period for a search. Even with a plus or minus 10 year interval, one should expect to not to get records about someone with the same name who was born after the person in question died. This happens all to often.

What I want to know is why when you are looking for someone in a specific census, in a specific town, in a specific census area, you have to do a page by page, line by line search. Why with couldn’t the search engine search this small data set by the first or perhaps the first two letters of the last name or by where born. I have recently located my great grandfather in the 1870 census in Grand Island, Hall County, Nebraska using the a page by page search after spending months searching using regular techniques. His name Cleary was spelled Clary. I would think that should have been fairly easy search. I also finally found his wife in the 1930 census using the same method. This I could understand because her name Cleary became Clacky when transcribed. It was due to a combination of poor hand writing and the long loop of a y in the name above. That I understand, but I do not understand why I could not find a Cleary, spelled Clary when I knew the age, place of birth and residence location.

It would really be helpful to be able to use a * with only 2 letter in front. I have lots of Clearys in my family tree. It is spelled many ways. Being able to enter Cl*ry, would really be helpful. Often it is the vowels that are the problem. a becomes o and even sometimes an e or u.

5 KPAugust 1, 2008 at 3:52 pm

Not to pile on. But I agree with the above. I wouldn’t exclude records that don’t have death date, just because you put a death date/location. I would only exclude records that have a different death date/location.

On a positive note, I love being able to see the image previews on Newspapers and several other records. I would love to see that eventually expanded to other records (but I’m guessing its only available on records that search using OCR).

6 amitchellAugust 1, 2008 at 4:37 pm

Good comments.

As for trees, that’s probably a long discussion, so I’ll leave it for another time.

And this is really splitting hairs, I know, but you can get the census included, by excluding exact on the death date and death location and leaving everything else exact. That is currently the way the search engine works and because the exact was on for death, that was the result.

When I do a search, I do not expect, or even want to get every document about that particular ancestor in one fell swoop. Too much for me to process. I tend to look for specific pieces at a time. But that may be me. Is it better to have the ability to push a “magic button” and get everything or methodically go through certain types of documents? I’m very curious to hear everyone’s thoughts on that.

BTW, creating the ability to cutoff documents by death date (no one wants to see 1930 census pages for someone born in 1820) is on the list of things to do. I’ll address that specifically in a later post on why that is happening and what the solutions are.

Fran, I’d like to dig into your Clary/Cleary problem a little deeper, because Soundex should catch that. What are first names/birth years you are looking for? And I understand your frustration with spelling. My family name is Gillespie and I have at least 10 different variations to run through.

KP, I also love the previews. :-)

Also, Jerry Bryan posted a very interesting problem with a search for marriage records a week or so ago that I am trying to dig into and figure out. Look for a post on that as well next week.

Keep posting and commenting and I will keep reading. All of these comments and examples will help shape where we go and how we build search.

7 B. PriceAugust 1, 2008 at 5:45 pm

Thanks for the great discussion!

Yes, I do want the “magic button” many times, as I don’t always know what records to search from random-third-cousin-twice-removed that I haven’t spent tons of time on. Give me the bulk, let me filter out…the other way around is a heckuva lotta more work.

8 Carol A. H.August 1, 2008 at 5:51 pm

Anne,

Perhaps I misunderstood, but I tried to follow what you did by clicking on the images you posted and got 0 results for all 4. I really wanted to see what happened, so I got on the new search and duplicated what you did, step by step.

The first search yeilded one personal public tree.

The second search yeilded 14,952 records.

The third search yeilded 146 records.

The fourth search yeilded 27 records.

I really can’t say how much of value these results were because John Cousins in not my ancestor. I know nothing of his decendants and how Tony C. got there. In most cases, not all, when you are searching for a person, you have some idea of who, or where or what you are searching. The new search does not display enough records on one screen to do a good mental evaluation of what you get. My brain can juggle other facts to determaine if a record may possibly be of use. I don’t think your interface can do that unless you give it more to work with simultaneously. And it seems the more information I give for a new search the fewer results I get and most times zero results. (Maybe Ancestry just doesn’t have ENOUGH records and ones that are contiguous.) So I guess I finally figured out that I do many “fuzzy” searches; look at what I get and compare with what I have and consider the place and time in history. I know what was happening where and when, which will help determine if I think a record is pertinent. Also I consider the possibility of transcription problems, spelling errors, the human element, and on and on and on… All this at once. Every good genealogist does this. Computers don’t unless the program is very expensive.

So….I find more of value faster with the old search. I’m the one who does the evaulation, not the computer.

I have 5 trees and one is in West Virginia, and western Virginia, so I know how difficult that is. And yes, I have done many a search in a census, page by page due to poor indexing. Also, I never take anyone’s tree for gospel. I never merge any tree into mine!

Ancestry has taken some of MY information and merged it with trees that are WRONG and undocumentated! I might look for a tree as a curious effort, but I repeat, I never merge any other tree with mine. I don’t want a search that includes trees! Jade in #2 hit it right on! Keep the tree search separate.

Both B.Price #1, and David Watkin #3, had excellent comments, too.

I hope you don’t feel your fingers got burned. Good luck in your new position. WE need someone to listen to US.

9 Carol A. H.August 1, 2008 at 6:07 pm

Well while I was typing several other people posted. I’m a slow terrible typist and not too great at spelling.

Yes, I “want to get every document about that particular ancestor in one fell swoop.” I agree with B. Price on that…totally. My brain works fine.

10 B. PriceAugust 1, 2008 at 6:07 pm

One more comment, I do merge trees, so I like seeing the results (now that I can be selective about what I merge). I’m going through and doing large “casual” sweeps for data (and trees help here), then I go back and do the tough, detailed work..but I have found using other’s work to get a leg-up can be a life-saver (and a bother when they did it wrong, too).

11 Carol A. H.August 1, 2008 at 6:25 pm

Anne,

You posted: “Exact death location will prohibit me from getting a census record. Does that make sense? I think it does. A census record does not contain death information so it shouldn’t match.”

But there is some birth informaion in most censuses. The person DID LIVE at some time in some place. Of course they should not be in a census after the death date.

I do get census results that make no sense time wise. The name may be the same, but other things are out of range of time and location.

12 JadeAugust 1, 2008 at 7:56 pm

Ann,

In your reply to some comments, you touched on the fundamentally different sorts of search:

“When I do a search, I do not expect, or even want to get every document about that particular ancestor in one fell swoop. Too much for me to process. I tend to look for specific pieces at a time. But that may be me. Is it better to have the ability to push a “magic button” and get everything or methodically go through certain types of documents? I’m very curious to hear everyone’s thoughts on that.”

It appears to me and to others that the very point of New Fuzzy Skin is indeed to retrieve every pertinent result in Ancestry databases concerning a person for whom one gives birth and death dates and places, parents, spouse, siblings.

This is a valuable idea for researchers who have not organized a search routine, who may just be starting out with a particular family. It is particularly a useful approach for post-1870 US families where the searcher already has fundamental data, and is seeking documentation of readily surmisable events.

New Fuzzy’s interface with user Trees supports this approach.

The problem is that New Fuzzy retrieves either thousands of completely irrelevant results (such as British and Welsh vitals for persons who never were there), or presents the search results in completely unretrievable ways – such as ordered by number of ‘hits’ in a particular database in a list of thousands, or requiring one to navigate through pages and pages of the irrelevant to get to possibly one relevant one at the very end. The results cannot be ordered in any sensible way by the user.

For those of us searching for a particular puzzle piece, New Fuzzy is not suited to searching specific types of databases or a specific database. Why should the searcher have to do all that clicking to get the World War I Draft Registration database when one can immediately go to it by using their own browser bookmark and using Old Search?

Further, New Fuzzy often does not retrieve results that researchers know are there because they have found them before. Examples have been posted on these boards, with the presenters’ having been careful to duplicate the spelling given in the records.

There is a bias in New Fuzzy toward retrieving what Ancestry has, rather than *not* retrieving what is irrelevant. Since the Ancestry database is deficient in such items as 18th- and early 19th-century tax lists for any location, lacks 18th-century vitals for most locations, and indexing of pre-1850 Census enumerations does not include age-range of head of household, New Fuzzy does not help much with pre-1850 households, even with bracketing dates. It much prefers to return 1930 US Census results with many ‘same-name’ hits, and post-1890 miitary results, despite the target person’s having died in 1875.

Furthermore, New Fuzzy tends to disregard the given dates unless they are made ‘exact’; and you have touched on the counter-intuitive problem with how the computer interprets ‘exact’ dates.

I don’t understand the reasoning behind creating New Fuzzy. A global search already exists that is much easier to use than New Fuzzy, although its results-ordering scheme is no better.

I personally would prefer it if Old Search were enhanced by allowing the user to make specific fields in a search page to be ‘exact’, and by adding boolean exclusionary capability.

Your attention to these matters is very much appreciated.

13 Jerry BryanAugust 1, 2008 at 10:11 pm

Anne, I’m so pleased you are on board and that you are a real researcher.

Despite my welcome, I couldn’t possibly agree more with the others who have already posted about how exact matches should work. The example of an 1877 death not disqualifying an 1870 census is a great example.

I think if you will read the blogs of the last few weeks carefully, most of the people who have posted do understand the problem very well. So it’s not that they don’t understand how New Search works with respect to exact searches. Rather, they do understand and they think that the way it works is very dysfunctional.

Switching to fuzzy values does not help a search very much. In fact, switching to fuzzy values usually makes things worse. For example, if you make the 1877 death date fuzzy, you simply open yourself up to hits on 1880 censuses, 1900 censuses, etc. which can’t possibly have any real hits.

Despite your characterization of the Old Search way of doing exact searches as “hit or miss” (a characterization with which I would probably agree), it was extremely functional in practice. And despite how logical and “correct” the New Search implementation of exact search is, it’s extremely dysfunctional in practice.

Let me give sort of a generic example of the value of giving lots of exact values with Old Search, where the same search is hopeless with New Search. Suppose you know somebody was born in 1925 and died in 1998. For example, the data might have come from a tombstone. But suppose you don’t have a month and day for either date, and suppose you are pretty sure of the birth place but not of the death place. With Old Search, you are likely to get hits on the 1930 census, possibly on a few marriage indexes, likely on a couple of death indexes, possibly on an obituary index, and possibly on a cemetery index. With New Search, you are likely to get nothing, and the best you can hope for is a death index entry if the death index includes the birth date. Including the date of death will eliminate the census and marriage index entries. Including the date of birth likely will eliminate most of the rest of the possibilities except for a death index entry. Notice that I’m not looking for trees.

How’s this for a compromise. Suppose instead of two kinds of searches, there were three. It’s hard to know what to call them, but for now I will call them type #A, type #B, and type #C.

A. Works more or less the way exact searches presently work with New Search. Let’s call this one EXACT SEARCH.

B. Works more or less the way exact searches presently work with Old Search. Let’s call this one NORMAL SEARCH. (I would have a number of improvements to suggest, but let’s leave such improvements to another time.)

C. Works more or less the way fuzzy searches presently with with New Search. Let’s all this one FUZZY SEARCH.

I would have to think a little about about how best to set up the search screens to support these three options, especially if you wanted to mix exact and fuzzy fields within the same search. But if these three options were available, I would bet that EXACT SEARCH would be used about 5% of the time, NORMAL SEARCH would be used about 90% of the time, and FUZZY SEARCH would be used about 5% of the time. At least, that’s probably how I would use them.

14 ValerieAugust 1, 2008 at 10:12 pm

Thank you for responding directly to some of the issues we’re having. To respond,

1. “So what does this tell me. Exact death location will prohibit me from getting a census record. Does that make sense? I think it does. A census record does not contain death information so it shouldn’t match.”

To me, this does not make sense. The death date should work as a cut off. As in, John Smith died in exactly 1900, I do not want to see a record that references a John Smith who died in 1901. However, the fact that I *know for 100% fact* that he died in 1900, should not stop me from receiving census results from 1880.

Once again, the death date should work as a time cut off, not a records blocker. There is no reference to death in a federal census, so the search should understand this and disregard the dod field when pulling up results. It should not matter whether you have ‘exact’ on or not – it should disregard this field if it does not apply to the record.

2. “When I do a search, I do not expect, or even want to get every document about that particular ancestor in one fell swoop. Too much for me to process. I tend to look for specific pieces at a time. But that may be me. Is it better to have the ability to push a “magic button” and get everything or methodically go through certain types of documents? I’m very curious to hear everyone’s thoughts on that.”

I like to see everything. Then I decide where I want to go, but at least I know that I’m looking at all of the possibilities.

3. Here’s a example of an ‘exact’ issue. When using the new search, I start typing a name (Leverett Edley Waters) and the search engine fills in the results from my family tree, keeping everything ‘exact.’ Well, I have zero results. Now, if I unclick “match all terms exactly,’ I now have 1,469,794 results. I can narrow this down by picking & choosing my ‘exact’ fields, but depending what I’m looking for, I keep having to change my criteria (this goes back to wanting a “magic button,” but that’s another tangent)

So I’m wondering, why *every single one* of these results excluded before. Based on what you’ve said, I understand why I don’t receive a lot of the records that match. But, I don’t even return the photos that I’ve uploaded – that are attached to an individual in a public tree who matches *every single field* exactly.

15 Ron LankshearAugust 1, 2008 at 10:35 pm

I think a search in UK census with place for birth and living is now working. I previously posted that this was a problem.

However a search with spouse or any family member with a wildcard in the name is not finding people that you then find when wildcard is removed.

searching UK census for
lanksh*
living oxf*
finds 1901 census with 3 lanksheer
in Oxfordshire
PS image is Lankshear

now try a spouse search with rob* and no results – change name to robert and sarah is found

I still do not like the results change that does not have Image direct from hit list – old version allowed that.

And when I am searching and changing parameters it is annoying to have to keep clicking open the parameter box for item I just changed.

16 MikeAugust 1, 2008 at 10:35 pm

Hi Anne,

I want to repeat a point I have made on other blogs about “new search”. Many of us have issues with this search “experience” and mostly relating to how exact searches are handled. But I also share the view of Jade in that I am not interested in “everything in one fell swoop”, because except in a preliminary survey phase, good research plans are crafted to look for specific facts which are to be evaluated and which forms part of the basis for the next research plan.

So what I want is the ability to have exact boxes on every field of the *database specific* search forms. Thus if I go to the 1830 census, then I want an exact box for every field (and a field in the form for all actual fields in the database!), and not instead be forced to go through either old or new search to get that exact feature.

An important part of genealogical methodology is the research calendar/log with both positive and negative searches. And it is very specific for the sources that are searched. Listing “searched all Ancestry’s stuff on such and such a date for John Doe” just doesn’t cut the mustard.

So give us more experienced genealogists the exact search on every field by specific database ability we need, and do whatever you want with some site-wide search function for the less experienced which returns tons of hits that they don’t have the experience to wade through and evaluate, but which makes your marketing team happy.

Mike

17 cgAugust 1, 2008 at 11:25 pm

Why can’t the users have a CHOICE?

18 cgAugust 1, 2008 at 11:27 pm

I just signed on to find everything is changed AGAIN. The search boxes, everything is changed again.

Don’t be fooled by the sweetie pie niceyness of the woman posting. she still works for ancestry and is still trying to shove this horrible new search down all of our throats whether we want it or not.

If I can’t find the documents I need with this new crappy search, why should I continue my subscription to ancestry?

I am outa here.

19 FranAugust 2, 2008 at 5:33 am

Anne,

After I posted my last comment, I noted that I forgot to thank you for your interest. It is so nice to find somebody who is listening.

I began trying to figure out how I had done my search to find James Cleary/clary in the 1870 census and write it up. I have written a dissertation. I think what I have may be useful to you, but it is way too much to insert here. If you will email me (I assume that you have access to my email that I enter when I make a comment)I would be happy to send it to you. I think my comments may be helpful because I am relatively new to genealogy and had and have some of the logic of a new person, but with enough experience to perhaps explain it at least partially. One of the problems may be distinguishing between a fuzzy search and a specific search. I fuzzy a search is helpful.

Again thank you for listening.

20 TonyCAugust 2, 2008 at 7:43 am

Morning Anne and welcome.

Now that was a surprise when I started reading your post, someone is listening – and responding.

I think you may be hearing what you didn’t want to hear. From the majority of the 18 responses I think you may get the message, the new search is not working, well not the way we would like.

I’ve heard a few times, and please correct me if I’m wrong, that this really isn’t a new search engine – it’s the same as the old search engine with more options. If that’s the case use the John Cousins information on the old search and put the dates in the year range – it does not exclude the census records – so what is going on?

If Ancestry intends to continue excluding certain databases from searches based on data entered in the search then it may be an idea to publish, for all to see, exactly what we can, or can not, expect.

As for “When I do a search, I do not expect, or even want to get every document about that particular ancestor in one fell swoop. Too much for me to process. I tend to look for specific pieces at a time.”.

Maybe that’s the way you search but I like others on Ancestry prefer to see the bigger picture, but not 1.5 million hits :)

We all know that Ancestry is trying to improve the access to resources for all of us, but please – please listen to all of these people, don’t make assumptions and exclude records because it doesn’t make sense to you (Ancestry).

I look forward to more responses from you – TonyC

21 Jerry BryanAugust 2, 2008 at 7:57 am

It seems to me that the complaints about New Search tend to fall into one of three categories:

A. The merger of trees and social networking into the search experience, and especially the way in which trees and social networking have taken a position of primacy with respect to searching for data. Even though Kendall’s Myth #2 is that the New Search and the New Home page are related, I think they actually are very closely related from the point of view of vision. They may have been totally separate projects accomplished by totally different teams. That doesn’t change how they fit together hand in glove when you are using the system. The trees get in the way of searching for data in any number of different ways.

B. The difficulty of using the new search interface. Interface problems include: the way the search boxes work in general, pre-population of the initial search box from your tree, type-ahead in place fields and how the place fields work, how clumsy it is to refine a search, huge amounts of excessive clicking and scrolling, first names and last names being combined into one search field in many cases, the way dates are specified (e.g., exact dates in the SSDI), what dates mean after they are specified and how they are used to control the search process (e.g., death dates preventing census entries from showing up), search box fields that no longer exist and that have to be specified as keywords, etc.

C. Search results frequently are not useful. It’s only the search interface that has changed and not the search engine. The behavior of the search engine itself has not changed. We all get that. Nevertheless, the changes in the interface have changed the behavior of the search experience rather radically, and it is now much more difficult to get useful search results than it used to be. I truly don’t think that ancestry gets the fact that the results from New Search are usually so much less useful than the results from Old Search. I know that ancestry wants to start with New Seach and move forward, but you really have to look back at what used to work great and that now works not at all.

I’m very suspicious of such lists of categories, even if I’m the one who has made the list (especially if I’m the one who has made the list of categories!). Such lists are subject to at least two fallacies. One fallacy is that there may well be a secret category #D (or #E and #F!) that is being left out. For example, I thought about including a category #D that would have said that fuzzy search has taken a position of primacy over exact search, and that the primacy of fuzzy search has made it more difficult to do good quality exact searches. I think that’s true, but also I think the primacy of fuzzy search really manifests itself as a problem either in category #B or category #C. So I didn’t want to list overlapping categories. The second fallacy is that despite my best efforts to create disjoint categories, the categories do overlap. A lot of the problems in category #C are really caused by problems in categories #A and #B (especially by problems in category #B).

ancestry has acknowledged problems in category #C but not in category #A or #B. Anne’s first message concerns a problem in category #C, for example. But ancestry continues to believe and to insist to us that the new interface is actually easier to use than the old interface. And they seem to be most proud of the new features that are the very worst of the new features. That message came through loud and clear in Kendall’s webinar.

I’m tickled to death that Anne is willing to engage with us on problems in category #C. I hope that also she will engage with us concerning problems in categories #A and #B. That’s because I think that despite my attempt to create some sensible categories of problems, the problems have to be solved holistically. I don’t think you can solve problems in category #C without also solving the problems in categories #A and #B.

22 William WestallAugust 2, 2008 at 10:50 am

The new interface creates something new to learn and master in order find the information. It is always frustrating not to be able to find things where you left them last. The “exact” search has always been a frustration because it is always an “everything or nothing proposition”. What would be wonderful would be able to chose by field which you want to be exact and which fuzzy. If a person doesn’t have a very common first name you can very easily find it spelled four different ways not to count nicknames. If you only want to look in one county or township or for a fixed span of years for date of birth it would be nice to select what is to be exact and what isn’t from a menu on a single screen.
By putting in as much information as available into the search parameters for exact searches should not exclude records that don’t have a field for a bit of information provided. If the record matches all other search parameters except the fields that are not included on the record then it should be included in the results.
As for family trees posted, they can be extremely helpful and can provide information that can point someone in the right direction. The trouble is that many are the first efforts by people just beginning and solicited by the software to publish after a couple of hundred names have been added. If possible check every entry. The entries you can’t check always remain question marks.

23 JUDY ADAMSAugust 2, 2008 at 2:09 pm

WHAT THE F***

I HAVE JUST LOGGED ON AND WHAT DO I FIND YOUR ******* NEW SEARCH. I DID NOT WANT TO USE IT I DID NOT ASK TO USE IT SO WHY THE **** DO I HAVE TO USE IT

NONE OF YOU HAD THE GUTS IT SEEMS TO SAY WELL AS OF AUGUST THE 1ST YOU WILL HAVE TO USE IT

THERE AGAIN WHAT MORE CAN I EXPECT FROM THE POWERS THAT BE WHO HAVE BURRIED THEIR HEADS IN THE SAND AND WILL NOT LISTEN TO US YOUR LONG TEWRM PAYING CUSTMER (NOT FOR MUCH LONGER AS THE DAYS GO ON I AM MORE SURE I WILL NOT RENEW) FIRST YOU FALSE THE AMT WITH ALL THOSE UNWANTED BELLS AND WHISTLES ON US MY REACTION I NO LONGER CONSTRUCT OR EVEN USE MY FAMILY TREE ON LINE … I USE TO IN OLT EXCLUSAVILY. THIS MEANS THAT MY TREE IN AMT IS ALWAYS AT LEAST 3 TO 6 MONTHS OUT OF DATE I HAVE EVEN REMOVED THE PHOTOS

NOW WE HAVE THE NEW PAGE AND IT SEEMS WE HAVE TO USE THE NEW SEARCH IT IS FULL OF BOXES MANY OF WHICH WE HAVE TO WAIT FOR A PORTIONED CIRCLE THING TO STOP SPIRALING BEFORE WE CAN DO ANYTHING . PROGRESS MY FOOT

ITS ABOUT TIME ANCESTRY HAD A CRISES MEETING AND THE POWERS THAT BE SACKED ALL THE JOBS WORTHS WHO DO NOT HAVE AN INTREST IN GENEALOGY BECAUSE RTHE ONLY WAY THEY WILL EVER GET IT RIGHT WHERE WE THE CUSTMERS ARE CONCERNED IS TO EMPLY PEOPLE WHO ARE INTERESTED IN GENEALOGY AND WHO WOULD THEREFORE UNDERSTAND WHAT WE YOU CUSTMERS WANT WITH IN A SITE DEDICATED TO GENEALOGY

24 JUDY ADAMSAugust 2, 2008 at 3:41 pm

thank god the switch back suddenly reapeared

but why i was stuck in new search i did a test search

my 2x ggfather was born in 1840 in hampshire (england) and died in 1908 an ideal candidate because he apears in all census and is on the free bmds and a number of trees including my own

i did an exact search the first thing i hated was when i put in james a box then apeared showing all my james in my tree with 3000 people onit and james being a popular name there were many entries …. if i needed to decided who it was i was searching for i would have consolted my tree prior to starting the search a list of all james in my tree is not required even if i can click on it my conclution an unwanted bell and whistle feature and seems some what condecending to me and very aoying.

i then put in westbrook, birth 1840 +/- 5 years (gave it a bit of lee way) then death 1908 +/- 1 year and hampshire all exact search in the old search i would have had the uk census along with the us ones and the freebmd and asorted others even with out specifing a country but still using the old exact box ok i could not put in death date the result i got on new search NOTHING MATCHING not even one tree

SO I TOOK OUT DEATH DATE still i did not get the results from britain result was 1 tree from the usa defently not my relative

SO I TOOK OUT BIRTH DATE now i had 7 records 1 usa birth no good to me and six trees all belonging to a single tree owned by a relative of mine but the target james was not among them

the next thing i did was to untick the hampshire and re aply the 1840 +/- 5 now the search finally gave me the record i wanted

i can only concluded that when using exact in the location a single countie be it texas (usa) or hampshire (england) is not enough to produce a result. putting in the couny hampshire should in my mind have yeilded both usa and english results firstly hampshire is a location in england (from which more than 1 james westbrook came) and as there is a hampshire in the usa albe it having a prefix of new there should have been results shown from england at least yes i know the location box has a drop down asking which one i was after but people being people all dont add in the full locatio bits and i was testing. anther detail the word hampshire apears in my tree attached to james seperated by a commer from the rest of the location thus ‘wield, hampshire, england’ so the fact that i just put hampshire into the location search should have caught my tree at the first search along with all those i have seared my tree with (i am responsible for the entire reseach on the line that others now have encoperated in their trees having merged from my file) the only conclution i can make is that the exact searc needs some more work

yes the page does not look to bad but the home page takes for ever to load because of those circle things. GIVE ME OLD SEARCH EVERY DAY

i just thought i would just mention why i use dot com while i am researching i log on to both dot com and co dot uk then if i find a record on one i can search the other for anything which matches. i normally search the trees in dot com and the records in dot co dot uk so theres a tip for you all you can have both running at once it sure helps while you search!

25 C. J. SmithAugust 2, 2008 at 6:35 pm

Welcome Anne!

I am finding that the new search can be helpful although I am finding some specific problems. For instance:

I searched the California Birth Records for a child with the last name Blevins who had a mother whose maiden name was Lavadore. (I already have two children that fit that criteria and was looking for a third.)

First I checked both boxes and come up with nothing but expected at least two children would be listed.

Next I unchecked the exact box by Lavadore. A multitude of names came up. I found one of the known children far down the page and did not find the second after going through the first two pages. Keep in mind the the listing I found did have the mother’s name spelled Lavadore.

I then deleted the name Blevins and conducted the search for any child with a mother whose maiden name was Lavadore. Nothing…

Then I decided to switch back to the old search. All I did was type in the name Lavadore in the section “Mother’s Maiden Name. Three children were listed up top. Two who carried the last name Blevins.

The new search needs to be more responsive to all the criteria.

Good luck.
CJ

26 KPAugust 2, 2008 at 7:16 pm

On the search here’s another example of something that doesn’t work as expected. When I’m lookinga the the “Historic Land Ownership and Reference Atlases 1507-2000″. If I search by location “Ray County, Missouri, USA” I get no results when using exact location. Now I know there are clearly Ray County maps. Now if I go in and search by Keyword “Ray Missouri” bingo all my maps pop up. I’ve had similar problems with other maps using the exact location search. But if I didn’t know those maps were there I would have just assumed they weren’t and quit looking.

27 Marie Cardoza (Stella)August 2, 2008 at 8:25 pm

Although I bought the 2008 version I have not used it because I could not use reference number IDs that were automatically assigned in pevious versions of FamilyTree Maker. I use these ID numbers as the basis of my filing system. I have many names in my family tree that are identical, so the unique numbering system sets them apart for easy access to paper documents that are filed numerically. Please help! Thank you.

28 Ron LankshearAugust 2, 2008 at 10:50 pm

I DO NOT LIKE NEW SEARCH.

I have persevered as I assume you intend to implement. But if you do implement as it is I will probably NOT renew my subscription.

Many UK databases have special search screens to cater for the special format.
Example BMD – BUT new search appears to have one basic search panel. So that year of registration in not there for BMD

Similarly Pallots marriage index does not have the search for two lots of forenames and surnames

The Results presentation is absolutely ridiculous. I posted yesterday that the ability to open census image from first hit list had been dropped. Today I see that the box/frame around the results will actually hide data if you click increase font etc. This is in Firefox and with a census the “residence” column disappears as I click larger font. The data is lost “under” the frame.
In IE8 it looks like only the first part of final field is lost. of course perhaps the image click is still there but I cannot see

if you are trying to get a search panel that covers searching family tree then leave it for that purpose but provide something else for source data

29 SteveAugust 3, 2008 at 5:57 am

I recently paid several hundred dollars to Ancestry for an advertised product. Now I’m being told I’ll be getting a different product instead.

This is called “bait and switch” and is illegal, at least in North America.

Will you be offering a refund to those who purchased something other than what’s being delivered?

Perhaps you could take a look at the 500-pound gorilla in the room — Google. Nice clean interface and resounding success. Microsoft, all glitz and glamour, barely registers on the search-engine radar.

I don’t want a “search experience” — I wish to do serious research.

30 Jerry BryanAugust 3, 2008 at 11:22 am

I have reported various problems in the vein of old searches converting themselves unbidden into new searches, exact searches turning themselves unbidden into fuzzy searches, etc. I think I have a clue as to why this may be happening.

I regularly have multiple windows open into ancestry.com at the same time. I’m not sure whether you are really supposed to use ancestry.com this way or not, and I could certainly picture it causing problems with searches. But it is a very useful way to use ancestry.com, and I never had any problems without before new search came along. For example, I could be looking at a census entry for a couple, and before finishing the census entry I could open a new window to look for another census year or to look for a marriage record. It’s very useful to be able to compare results from multiple windows at the same time.

The following seems repeatable the create some of the problems I’ve been talking about. Open a window into Old Search in the Tennessee Marriages Database with the exact option set. Open a second window, go into New Search with the fuzzy option set (or really with the exact option not set, there not really being a fuzzy option) and conduct a search (any search, I think). Return to the first window. It looks like it’s in Old Search with the exact option set. But the next search will be a fuzzy search with New Search instead of an exact search with Old Search as expected.

Are search options stored in cookies on my local machine, or are they stored on the server? And if the former, are they session cookies?

31 Carol A. H.August 3, 2008 at 1:54 pm

Ah, Jerry Bryan, a man after my own heart! I do the same thing. More than one window of ancestry open at the same time. I too need to compare information frequently, before I proceed to examine the next hit. And I have other windows open, too. My database (not FTM), Google and occasionaly something else. How else can you evaluate what you have? Isn’t that the purpose of windows?

I have not done this while in the new search, however. I have NEVER had any problems while in the OLD search. My computer is a happy camper and so am I.

Old search….Yeah!!!

32 Jerry BryanAugust 3, 2008 at 2:51 pm

Here’s an example of a search problem from today. I’ve been looking for the courthouse marriage record for Martin Cole and Elizabeth for years. They were married about 1840; Martin was from Hawkins County, Tennessee; the couple lived in Hawkins County after they were married; and the courthouse marriage records for the time period are extant for Hawkins County. But there is no courthouse marriage record for them in Hawkins County.

I’ve long been suspicious that Martin and Elizabeth were married in a nearby county where there has been a courthouse fire or where the courthouse was destroyed in the Civil War. So it’s not entirely fair to ask New Search to look for such a record. Nevertheless, I gave it a try. The results are instructive.

I set up the search as an exact search for first name Martin, a fuzzy search for last name Cole, and an exact search for spouse Elizabeth. Cole has spelling variations such as Coles, Coal, Coale, Coals, and Coales (not to mention Kohl, etc., but I’ve never run into any of the K spellings in my family). Also, the spelling of “Cole” is frequently indexed incorrectly as Cale. So it’s essentially impossible to set up a single exact search for the spelling variations of Cole, given the restrictions on wildcards. So when I’m researching the Cole family, I usually have to do a series of exact searches on the spelling variations to be sure to find the record I’m after.

Kendall mentioned that fuzzy search uses a name authority to look for spelling variations, so I thought I would give it a try with a fuzzy search for Cole. I purposely chose to leave the date and place blank. It does not seem very straightforward with New Search to setup places and dates in such a way that they apply to marriage dates and places, although Old Search handles the same problem just fine.

Needless to say, New Search didn’t find the record I was looking for, a record that probably doesn’t exist anyway. But what caught my eye was 3 hits in the “Missouri Marriages to 1850″ database. Many of my Tennessee families had branches that moved to Missouri, and sometimes the people that had moved to Missouri would return to Tennessee. So I peaked at the 3 hits.

They were for Elizabeth Hunter to Martin Powell, Elizabeth Usher to Martin Tailor, and Elizabeth Walser to Martin Shelby. Powell, Tailor, and Shelby are ridiculous as hits for Cole, even with fuzzy search. But here’s the good part. All 3 hits were in Cole County, Missouri.

Fewer members of my Tennessee family moved to Illinois than to Missouri, but the same research results happened with Illinois. There were 3 hits for the “Illinois Marriages to 1850″ database. The surnames of the men didn’t match Cole at all, but all 3 hits were in Coles County, Illinois.

It appears that New Search is treating the search values as keywords (or at least some of the time it is treating some of the search values as keywords), rather than tying the search values to specific data fields. That’s ok if you are searching on keywords, but it can’t be right if you are searching on specific data fields.

33 GloriaAugust 3, 2008 at 3:31 pm

Someone developed this new software and got paid big bucks, and it is a disaster!

I agree with other comments completely, so will not repeat them. Rather, I will say that my ancestors come from eastern Europe, and had difficult names that were fairly slaughtered by census takers and transcribers alike.

Often, I have done a global search using a first name or just a portion of the last name with a wildcard. And I was able to isolate the database I wanted (I don’t need to explore Hawaii Landings) I have had success with that method.

NOW, global search either brings up a zillion names to wade through (in some cases, it doesn’t even show how many – you just plod ahead.) OR, it brings up “nothing found”. Now, if I’m searching the NY Immigrations Database for specific years for a name like “Mary” (as I must examine the supposed last name) — and I’m told “nothing found”. REALLY? Not one Mary?? Not one John??

In other cases, entering a portion of the last name with a wildcard comes up with “nothing found”; but if I enter the entire name, it’s there. The whole idea is to look for abberations of the name! If I knew all the details, all the data, I wouldn’t need Ancestry!

Many of these records I already have, but I would like to examine again on the screen. “Not found”!

At some times, I can be looking at it on screen, and try to go back to it again a minute later – “not found”.

Tech Support tells me it’s my settings. I have adjusted my settings as instructed, along with other suggestions. WILL NOT WORK!

It works; then, it won’t work.

Ancestry charges a LOT of money for this access. Now, I cannot trust the responses of Ancestry. So why should I continue to pay for this mess?

You can imagine how FRUSTRATING this is!!

Also, I have tried posting on the Message Boards (and, yes, I know how to do this), to find messages not posted or only posted in part. With Tech Support’s assistance, messages were noted as received, and re-posted. In other instances, I requested to be notified of answers to my posts – but NOT.

Last (but NOT least), the weekends are when many have the time to work on their search. Yet, Ancestry.com and the Generations Network as a whole are CLOSED. There is NO ONE to talk to, except venting on a blog!

This software is SOOO screwed up!!

I have paid a lot of money for this service and it’s become an aggravating waste of time, with untrustworthy results.

I can go to the public library or LDS locations to access this site. But I will not continue to PAY.

34 AthenaAugust 3, 2008 at 5:02 pm

“So what does this tell me. Exact death location will prohibit me from getting a census record. Does that make sense? I think it does. A census record does not contain death information so it shouldn’t match.”

It makes no sense whatsoever to me. When I input “exact” (which can in fact be quite inexact) death information, it should tell the search engine that there is no point in including census years after that date in the search and that hits that match the death location should be ranked higher than those that do not. In other words, I if I say someone was born in NC and died in MS in 1860; there should be *no* census hits after 1860 and any that are in MS should be ranked above those from MI or elsewhere.

That’s what makes sense to me.

35 Anne MitchellAugust 3, 2008 at 8:19 pm

I guess I should have known better than to post on a Friday when I couldn’t do a good job of checking in more regularly over the weekend. But you all have been busy!

A couple of ground rules please: No yelling (CAPS are yelling), no asking for people to be fired, and no swearing, even if you have part of the word ***’ed out. We know what you mean, and it’s not nice.

Now, if you want to express frustration, unhappiness in a more or less civil manner, I’m all ears. I want to know what makes you unhappy, and what features you would like. You may not be a big fan of the new search interface, but I don’t think the old search interface was perfect. I want the best possible experience for all types of searchers on ancestry.com, whether they are casual searchers, serious researchers or something in between.

I’m going to go through all the posts, and try and summarize what I think you all are saying, and I’ll post again sometime on Monday.

And thanks for all of your examples and opinions. One thing is for sure, I’ve found a very passionate member base here, and that’s always the most fun group to build products for.

36 Jerry BryanAugust 3, 2008 at 9:05 pm

Here’s a simple marriage search that works fine on Old Search and not so fine on New Search.

Thomas J. Cloyd to Elizabeth Ann Miller, Tennessee State Marriages, 1780-2002. I’m starting from within the database, not with a generic search, and I’m using exact search.

You never know if you are going to get T. J. Cloyd to E. A. Miller or what, so I usually start with last names only.

With Old Search, an exact search for Cloyd to Miller gets one hit (the correct one), and an exact search for Miller to Cloyd gets one hit (the correct one).

With New Search, an exact search for Cloyd to Miller gets one hit (the correct one), and an exact search for Miller to Cloyd gets no hits.

I’m sure that the problem is related somehow or other to the fact that the spouse’s name has the first name and last name combined into one search field in New Search.

If you then try fuzzy search on New Search, Cloyd to Miller works fine and Miller to Cloyd does not get a hit on the first search page.

Finally, if you go back to Old Search and try a fuzzy search, it works fine whether you look for Cloyd and Miller or Miller and Cloyd as the last names.

By “works fine” on fuzzy search, I mean that there are a gazillion hits and that the marriage I’m looking for is the first one in the list.

37 Jerry BryanAugust 3, 2008 at 9:17 pm

I’m glad Anne is back. I look forward to any messages she may have for us on Monday.

She said (in part): You may not be a big fan of the new search interface, but I don’t think the old search interface was perfect.

I would give the old search interface an A, and I would give the old search experience about a B-. The reason for a B- vs. an A is because of problems in the search engine itself and because of problems in the indexes. I think the old search interface itself is really quite good.

I would give the new search interface an F, and I would give the new search experience about a D+. The new search interface really is that bad. I still can’t get through a night’s worth of research without having to give up on the new search interface and going back to the old one. I give the new search experience a slightly higher grade than the new search interface because I can get some work done through the new interface. But I get work done with new search despite the new search interface, not because of the new search interface.

38 JadeAugust 4, 2008 at 5:39 am

Anne,

Your fresh set of eyes on the search engine problem is very welcome.

I also greatly appreciate the views of serious researchers Jerry Bryan and Tony C on this.

I have been heading back to basics.

One is that the existing Search Engine is very limited. Some have suggested the engines by G**gle and eB** as having very useful ability to search up to billions of pages, with boolean exclusionary capability and very easy-to-use search-term forms.

Another is that the Ancetry search engine is searching a very odd hodge-podge of indexes, perhaps some installed in some type of SQL server.

In addition, Marketing wants to integrate searching Trees with searching databases.

On the user side, opinions are divided as to whether the odd assortment of Trees on Ancestry have any relevance (on one hand) or (on another) whether the records databases are of any interest.

On the New Fuzzy introductory page,

http://search.ancestry.com/search/preview.aspx

a text area on the right states:

Ideas for the future
Based on the feedback we’ve already received, we are also considering many of the following additional features:

*Combined record page and image viewer, allowing you to see all of the information at once

*The capability to submit corrections on transcription errors of dates, locations and more

*Search engine enhancements that will return more relevant results and reduce inaccuracies

*Results about the locations and time periods you are searching

*Results that indicate which content you’ve already viewed or saved to your tree

*And more

.
I find it rather shocking that “Results about the locations and time periods you are searching” is the fourth item in the list, as an “added feature”. For me and many others, this is the crux.

A more or less sidebar issue is that the Ancestry databases are very limited in *records* for numerous areas of the US. Persons with research problems in most Southern states, Delaware, NJ, and in the Midwest prior to 1850 do not have many resources here. Oh, there are some mostly awful books and poorly executed ‘extracted records’ databases, but ‘paucity’ best describes breadth and depth of coverage.

39 kkandtcAugust 4, 2008 at 8:10 am

I spent most of my weekend comparing the old and new searches. I completely agree with Jerry. I was a member of Ancestry for a few yrs when they first came online. I cancelled my membership due to the cost of membership vs. the time I had available to do research. Four months ago I re-joined as new information in my line surfaced. So I am not a newbie to Ancestry.com.

Yes, there are problems with the old search, but still very usable if you know how to use search engines. What it presently lacks could have been upgraded/fixed without a complete overhaul. It is much easier to do specific database searchs with the old search. A nightmare with the new search.

1. Rather than a “sort by stars” other sort options would improve the search tremendously.

For marriages, you can already narrow down a county in your original search. Further sorting by year, sort by groom/bride.
For census records, sort by birth year, sort by surname, given name, state, age, county, sex, age, birthplace, location.
For all other databases and searches, make the sorts database specific.

2. Rename the “view record” link to “quick view”. Put that column next to “view image” column, either both on the left or the right.

3. When refining a search, get rid of the code that puts past requested search options into the keyword box. If you are changing a search word/option you don’t want the old info throwing the new search off.

4. If we are going to be forced to use cookies, re-do them to make them user friendly, if they are creating the following problem. If not cookies creating the problem, fix your coding. Many of use work with multiple windows open ie: multiple census, census/marriage etc. If you click on the “back to the record” link, many times it switches to another record (if you use the back button most of the time you can get around that glitch).

5. The ideal search engine will not find information of a database is indexed incorrectly. I appreciate how difficult it is to read many of these documents. I’m seeing a lot of human error/sloppy translations. One of the latest additions, the Tennessee State Marriages, 1780-2002 looks as if you outsourced the indexing to a foreign country.

Yesterday I spent over an hour just submitting spelling corrections. One of my family members, Beshears was transcribed as Besbeass. Although not my family member I did correct this one, with common sense and doing a bit of quick research.

B D S Royston was transcribed as B S S ??Oyster. Sarah Ann Frisby was ??Rah Ann Frisby.

Other names on just this one “hit page” William Bagby was William Bagbygs. On the same search page other names were Reza Baghaevaji, Francis Beeshbosed, Curtis Roy Besbaes, Bosbyshell, Aline Buckves.

Many of the TN marriages have 2 entries. One with more normal, if not correct, spellings. Those were not compared to help keep the errors to a minimum.

The NC death certificates have many of the same type of mistakes. I didn’t save a list of my corrections, but a recent post in the last few days by someone else mentioned. I can’t find it to quote correctly, but the gist was compare the spellings. Typos are being passed through to the indexes.

On a recent census record the person was indexed as “One without a name” Huddleston. Do a census search for all years using “no name” as the first name and leave the surname blank. If you leave the box unchecked for exact, click on “see more” to the left.

6. The soundex search is not working/coded correctly. Recent fuzzy searches for Huddleston H342 have yielded Hildago H432, “Mary L.” (a given name highlighted) if all together it would be M-640. If we were able to sort census records given the options I mentioned above, we would get much better results.

Specifically a 1900 census search for Huddleston yielded 15 hits. No soundex/spelling variations of the surname listed. Manually you have to input all other spellings or use wild card searches. I was doing a Beta test search this weekend and 2 Haddlestons came up also. When I searched for Haddleston directly on Ancestry, yes, they were there. This just proves that the soundex coding can be done to include variants.

http://www.microsoft.com/communities/newsgroups/en-us/default.aspx?dg=microsoft.public.sqlserver.programming&tid=7ca9e33b-6ddb-4084-9aaa-06c7ba01fc9a&p=1
http://microsoft.apress.com/feature/72/sql-server-soundexing
http://blogs.techrepublic.com.com/programming-and-development/?p=656

Did anyone take into account the number of your customers who are using older machines and dial up connections? Yours is one web site that I do not want to see all of the Java bells and whistles. I do like the ability to change the font size and the “Quick Link” feature.

The Learning Center page (link top right hand corner of this page) is a mess. As customers, we don’t need snazzy new graphics. On my laptop, even with my browser opened all the way up, the video and surrounding graphics are covering up the links on the top left. If you are going to insist on having videos, put then in a total sleep mode, and make them smaller. Give us an option to turn off all of the little pop up hints, or get rid of them. I don’t think any of us need an explanation box for what are clearly worded links.

40 ReedAugust 4, 2008 at 10:22 am

Dear Anne,

Welcome to the hot seat. We know you are in a difficult position and we do appreciate your apparent interest in our experiences with the New Search.

RE your comment no. 35: Your etiquette requests, while reasonable in a general-good-internet-manners sense are—I’m searching for the right word here—a bit out of touch with our experiences with Ancestry and its approach to “customer service.” We are yelling (IN ALL-CAPS!), asking for people to be fired, and even hinting at the occasional expletive because, month after month, year after year, Ancestry has shown an almost complete disregard for the comments, questions and suggestions of its user/members.

In the words of the blogger and former Ancestry employee known as “The Ancestry Insider,” your company has—and continues to maintain—a “culture of unresponsiveness” in dealing with its paying customers. (Perhaps this will begin to change through your efforts?)

Now, this may not be fair to you, as a newcomer, but many of us have been pleading with Ancestry for fixes to the New and Old search engines, the Home Page and the database indexes. We have used the online “Report a Problem” pop-up forms and NEVER gotten a reply. We have contacted “customer service” and rarely gotten a reply. We have complained about indexing problems and missing images on databases, promises are made, but they are NEVER fixed. We volunteer for the online Beta-testing of New Search and make detailed comments and criticisms. We answer Ancestry’s online “surveys,” even though we know that the “survey” is not a well-designed evaluation instrument, but a marketing tool that has been constructed to “push” the respondent to answer in certain (positive) ways. Year after year, promises are made and yet the same problems remain and new ones arise.

And recently? Ancestry dumps New Search (and New Home Page) in our laps, claims it is tested and ready to go and—no surprise here—it’s not. Many of us send emails and write blog posts (often in great detail), describing what’s not working with New Search and the company responds with a PR department webinar, describing “exciting new features” we have already discovered, and failing to address the substantial shortcomings of the New Search experience itself.

Frankly, I am tired of writing and re-writing posts to this and other blogs, outlining and correcting the errors of your software designers. That should be the job of your programming staff, in-house testers and Beta-testers. I suggest you go back and read the Ancestry.com blog comments over the last few months. Try the sample searches outlined by Jerry Bryan, Tony C., Jade, myself and others in this and previous blog posts.

When you do, ask yourself if—compared to Old Search—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)?
(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?

Once you can answer YES to these long-standing customer requests, then please show us, because New Search is NOT meeting these criteria in our daily experience. I’m sorry if this comes off somewhat harsh, but you’ve inherited a real mess. We wish you the best, but we need to see results, not more promises.

All the best,
—Reed

P.S. And FYI, we are not a bunch of technophobic, change-adverse, stick-in-the-muds. We are amateurs and professionals with a passion for genealogical research and high-quality sources. We 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.

41 Tony CousinsAugust 4, 2008 at 12:06 pm

Reed

That was indeed a masterpiece. I think you covered everything.

TonyC

42 MikeAugust 4, 2008 at 1:27 pm

I too would like to commend Reed on his last post, as it sums up very well the background to customer service most of us have experienced. By “most of us”, I mean serious and experienced genealogists (type #1 customers), the kind who expect and complain the most, but also the kind who are more likely to renew year after year.

I also would like to repeat what I have said before, give the unwashed genealogical masses the new search “experience” (what a pathetic marketing word), but give us type 1s (serious/experienced) alternatives, and by which I don’t mean the ability to just move things around the page.

Reed mentioned a very important word and it is (shouting here) CONCEPT. Sure we get to “test” out some of these so-called ideas, but many of them as Reed mentioned are flawed in the conception. What Ancestry needs to do is get our input in the concept/design phase and not when it is too late after the thing is done and the marketing/PR dept is just trying to get us to fix bugs in their poorly designed/conceived idea while it tries to put lipstick on a pig.

If one looks at Mr. Sullivan’s figures in a speech this year, one can see that Ancestry spends (shouting) FOUR TO ONE on marketing versus data acquisition. 4-1. That speaks volumes. My strong suspicion is that it is an attempt to put a band-aid on high customer turnover. Listen to us sincerely and address our concerns, including involving us in the concept phase, and not only would Ancestry likely reduce the turnover rate of us type 1 customers, but also get our help to increase renewals by less experienced customers. Our experience means we know what works and the newbies need that.

Mike

43 gail johnAugust 4, 2008 at 3:59 pm

Thank you Reed, you outlined the concerns of SERIOUS researchers to a tee. Also a heartfelt thanks from me to Jerry Bryan, Jade, Tony C., and the other posters who have taken the time to outline in great detail the virtually useless “new” search.
I have only been a subscriber to Ancestry.com since the beginning of this year, but I am increasingly frustrated and disappointed with the results I have gotten when using the “new” search. I generally avoid using it, and will more than likely decline to renew my subscription when the time comes unless SIGNIFICANT IMPROVEMENTS ARE MADE, AND ANCESTRY.COM SHOWS SOME REAL PROGRESS IN DEALING WITH CUSTOMER CONCERNS.

44 Carol A. H.August 4, 2008 at 10:49 pm

After reading the posts from this past weekend, Jade’s post #38 made me realize that I never read the introductory page to the new search. I just jumped in and used it. So I read it tonight.

One thing stuck in my mind after reading “Ideas for the future”

It was this: *Search engine enhancements that will return more relevant results and reduce inaccuracies.

Now that should have been in the new search before it ever was given to us and I think there would be fewer problems for me and for others.

Unless all this commotion is a marketing strategy to get us to do all the work and give marleting the improvement ideas. If so, it could backfire.

This past weekend I got the chance to try ancestry on another computer, (two other computers to be exact) both old search and new search.

There were some very noticable differences in how ancestry behaved. So maybe hardware could be part of the problems. We all can’t go out and buy new computers and some have slow ISP.

I have cable so that is not one of my problems but I think I may have to upgrade my memory. The advertising which I find inappropriate most of the time is slowing my computer in new search. Old search is beter. Some ads are not worthy of ancestry. Low class stuff. I won’t support your adverisers.

45 JadeAugust 5, 2008 at 12:51 am

Thank you, Reed, Mike and Carol.

Reed and Mike touched on need for customer beta testing. In Mike’s words,

“What Ancestry needs to do is get our input in the concept/design phase and not when it is too late after the thing is done and the marketing/PR dept is just trying to get us to fix bugs in their poorly designed/conceived idea while it tries to put lipstick on a pig.”

Unfortunately, even when Ancestry does beta-testing it disregards the user opinions. Just ask those who beta-tested the year-ago demolition of the Message Boards. Only after very clamorous complaints that Management had destroyed the boards’ functionality did they restore some of the thread structure, widen the field and allow wrap-arounds so that the message subjects could be read (!?!). Duh. The odd thing is that they did not have to destroy the Message Board structure to accomodate the additional ads, which was the true purpose.

Except then early this year they rolled out the revolting marketing tag-links, which was the underlying purpose for the change in Message Boards software. Evidently there was no beta-testing of this phase, which had very amateurish bugs, and despite minor tweaks still has such impressively idiotic features as links to other-counties-by-same-name. And anyone who’s worked on genealogy for more than 20 minutes realizes that the offerings of Ancestry Books on surnames are useless.

Whoever constructed that part of the Message Boards change has evidently had a hand in programming New Fuzzy’s treatment of surnames when not checked off as ‘exact’: one poster here found that New Fuzzy seems to treat such surnames as Keywords, retrieving search results for Counties named the same (or nearly the same) as the surname.

The Ancestry Insider’s view of Ancestry’s non-responsiveness to customers is really an understatement. The destructive pecuniary steamroller will fairly soon make finding anything but Trees nearly impossible.

And we haven’t seen anything yet. This is only the first phase. The additional ads and Marketing ploys have not been rolled out. The Marketing scheme behind the Message Board mess will be repeated. I do not doubt that this is the essential reason for New Fuzzy’s returning so many thousands or millions of irrelevant results.

Remember the days before G**gle searches were available on the web, where the first 50 or 100 results for any search were advertisements for stuff unrelated to your search? Could this time-worn concept be behind refusal of Ancestry to enable a searcher to order search results in any sensible way? Stay tuned.

46 RobertAugust 5, 2008 at 9:36 am

I am a longtime subscriber to Ancestry.com. I have used and appreciated the access to records otherwise impossible to find. I especially appreciate the ability to access the census records.
However, I agree with the comments of Reed regarding the deficiencies in the search engine, especially the requirement of the initial three letters and the limitation of wild card searches.
I am also greatly troubled by the non-responsiveness of the technical staff at Ancestry. For years I have complained about the omission of half of the names in the 1820 census of Lycoming County Pennsylvania. I never get a response and the records are never corrected. Ancestry discourages direct communication, requiring answers to a long questionnaire before allowing a question to be posted. Of course, the question is never answered.
I hope this dialogue will be helpful to Ancestry in responding to user interests rather than glitz and overflow of useless hits.

47 judy adamsAugust 5, 2008 at 3:40 pm

jerry #30 and carole #31

you mentioned opening the dot com page several times so you can search. jerry you found that you have a bit of difficulty. as i said in my post above if you open both dot com and dot co dot uk you will find no problem as they seem to work independently of each other. and YOU will be able to search for both usa and english records in dot co dot uk for best results there use OLD SEARCH. i expect if you open up other country versions eg australia , canada ectra it will work also.

oh yeah anne i will post this in capitals again because i am some what anoyed again

I LOGGED IN AGAIN TO DAY AND WHAT DO I FIND

YES YOU GUESSED IT

THE DREADFULL UNWANTED POOR FUNCTIONING

NEW SEARCH

despite having returned back to OLD SEARCH on my previous visit

I DOT WHANT NEW SEARCH

i do expect that once i have made a choice to stay with OLD SEARCH i should find that when i log on i am in OLD SEARCH still and not to have to make the choice again at least this time the button was where it should have been.

48 Joseph S. HaasAugust 6, 2008 at 8:51 am

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 http://comments.ancestry.com/mbexec/message/an/PEOPLE.193408910/1 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).

49 Anne MitchellAugust 6, 2008 at 9:02 am

Since this conversation has moved over to the posting:

The new search interface

I’m going to close this posting for comments, and ask that you share your thoughts there.

I have trouble keeping up with more than one post at a time, and I don’t want to miss your comments and ideas.

About the Ancestry.com blog

Here you will find informational, and sometimes fun, posts from the folks behind the scenes here at Ancestry.com. We hope you’ll notice just how passionate we are about family history and about the products we’re building to help connect families over distance and time.

Visit Ancestry.com
Notifications

Receive updates from the Ancestry.com blog Learn more