family search indexing marketing

I am a little disappointed with the current management of the Family Search Indexing project. I think that there is a lot more they could be doing to encourage greater productivity from their membership.

For one thing, they are a little closed. I think just some simple auto-updated statistics on the site would go a long way towards getting people fired up about the work. People like to think that they are part of something big, and from all I've heard, the amount of work that is getting done qualifies as pretty big. The more work people do, the bigger the numbers will climb, leading to greater excitement and levels of work...

To go along with the statistics, I also think rankings might be popular. You could see, for example, that last month you were the 3rd most active volunteer in your stake (or area for nonmembers), 247th in your state/country, and 8,914th in the world.

Another item that I feel has been over looked is the publication of a monthly newsletter. This should be easily found on the web site, but all members should also be asked if they would like to subscribe via email. Why? Simply because people get busy and forget about records extraction. It might not be that they don't want to or can't do it, its just that they don't think about it. The monthly newsletter would not only act as a gentle reminder, but it could also keep people updated on the new projects that have been added/completed, give tips for extracting, and just help people to feel included.

Right now, the extraction program is too much of a blackbox. A lot goes in, but seemingly very little comes out.

Anyway, those are my thoughts. I think the Distributed Proofing Project from Project Gutenberg could teach us a lot. http://www.pgdp.net . The first thing you see on their site is the number of books completed last month, and since the site began.

Aaron

That would be great if they were to put a running count of how many names were indexed in the indexing program, because we cannot always assume that the number of batches done equals x number of names.

For example, some death record batches had 24 total entries. 1900 Census batches have 50 lines of data. 1850 Census batches have 42 lines of data.

I'd like to see a stats page that is easily marked with a graphical, rather than textual, link on the homepage. Like you, I want to see similar stats, maybe not a ranking because if you are not doing a lot due to circumstances, it would look overly pessimistic.

Here are my ideas, I've numbered them for quick reference by anyone later in the thread.

1. Total entries completed (usually equals one name plus data, such as a census record line, etc).
2. Total batches completed*
3. Total users on system, how many logged in the day before, etc.
4. A few other interesting stats that might be fun for indexers to read about. Like most names in a day, week, etc., most names indexed by one person overall and by day, week, etc., without identifying the individual.

* The total batches completed line right now inadvertently appears that it might include partially completed batches that for whatever reason weren't entirely completed and had to be given to others to complete. I saw this myself already when I got too busy recently, the system still shows I 'completed' the number of names attributed to the batch, in this case a 1900 Census batch that had 50 entries.

Stake extraction directors already have access to stats for their stake members and others they act as group administrator for who live nearby on the indexing administration portal webpage. There needs to be a sort function on the page that shows the stats for each indexer so they can sort by field, I do this sometimes in the broadcast industry to get certain stats on radio ratings on an industry website, so it's not that hard to do.

Something else would be the ability to search the names already extracted, just so that those doing family history would be able to see the benefit of the project directly in their lives.

From what I've heard, the NFS development teams are only now getting to that part of it, figuring out how to integrate that data into the NFS project.

The test site is running with only limited data to see how this all works at present, eventually the goal is not only to show the indexed records as soon as projects are completed, but films that have been scanned but not sent to be indexed yet.

familysearchindexing.org is now showing completed projects, as of today there are about 46 completed items on that list.

Only the completed projects or also the work in progress? (As a project I would think the "Ohio 1890 Census" would be a project and "Book 1 Page 12 Side A Ohio 1890 Census" would be an in progress search)

Right now they only have the projects listed by state. To find out what states are being worked on all you have to do is look at the 'Current Projects' link, look at the 'Completed projects' link to see what states have been fully completed for a particular census, and there is also a link for projects being prepared for indexing.

You have an idea though, but it might take a little programming to get the system to show what pages on what project are being done right now.

if you want to see what has been done in the indexing project and is available to view, check out http://www.familysearchlabs.org/. You have to register, but it shows you how much of the record set is now online, and in many cases, is already connected to images.

I agree with all the ideas so far, I just have one comment:

JamesAnderson wrote:That would be great if they were to put a running count of how many names were indexed in the indexing program, because we cannot always assume that the number of batches done equals x number of names.

I think its important to make as much information as possible available without requiring that a person install the application. Part of the reason for the stats would be to help set the hook when people browse to the site.

I've started indexing, and I agree with all the comments, here. Successfully group projects have more feedback for the group and their goals. Sub-groups and teams also build. Think about getting a Priest Quorum to compete against the Laurels, etc. You wouldn't have to tie this to a specific ward. Your ward and Stake stats could be compiled in parallel to "teams".

Indexing is flying, now, but if we want to go supersonic we need to engage more than the members of the church.

Though a downside with competition is to get accuracy, when people are competing for numbers the quality might go down. If you could keep track of quality AND quantity then competition is great.

thedqs wrote:Though a downside with competition is to get accuracy, when people are competing for numbers the quality might go down. If you could keep track of quality AND quantity then competition is great.

Easy, you post the number of errors along side the corrects. Or a percentage. Or subtract the bad from the total and only show accurate numbers. Or reject a batch for rework. There are lots of ways to deal with quality and yet encourage competition.

Not so easy when you're dealing with data entry. All of the work is done by two separate people. An arbitrator views and discrepancies. More errors, more work for the arbitrator. There's also be possibility that the two data entry people are wrong - but agree. In that case the error may not get caught.

The work can also vary depending on what packets you get. Some are easy to read. Others are more difficult, but both are necessary contribution to the work.

I'd much rather see people take pride in their work then to engage in any kind of competition for numbers.

RussellHltn wrote:Not so easy when you're dealing with data entry. All of the work is done by two separate people. An arbitrator views and discrepancies. More errors, more work for the arbitrator. There's also be possibility that the two data entry people are wrong - but agree. In that case the error may not get caught.

The work can also vary depending on what packets you get. Some are easy to read. Others are more difficult, but both are necessary contribution to the work.

I'd much rather see people take pride in their work then to engage in any kind of competition for numbers.

Points well made. Competition could create headaches for the arbitrator.

Lets concentrate on better feedback. Let's provide this information to people: " I just did X names. How much closer are we to finishing the project?"

RussellHltn wrote:Not so easy when you're dealing with data entry. All of the work is done by two separate people. An arbitrator views and discrepancies. More errors, more work for the arbitrator. There's also be possibility that the two data entry people are wrong - but agree. In that case the error may not get caught.

The work can also vary depending on what packets you get. Some are easy to read. Others are more difficult, but both are necessary contribution to the work.

I'd much rather see people take pride in their work then to engage in any kind of competition for numbers.

The Irish Death lists have about 375 names in them and no line numbers. When one or both indexers get off the arbitrator has to fix the mess. I wish feedback to the programers was possible because there is a real need to be able for the arbitrator to insert a record in either the A or B indexing to clean up a missed name. My wife is an arbitrator and has spent up to 4 hours cleaning up the messes where A missed 4 records and B missed 7 all in the early part of a 375 record batch. If she could insert a Record in them they would have cleaned up quickly.

srweight wrote:The Irish Death lists have about 375 names in them and no line numbers. When one or both indexers get off the arbitrator has to fix the mess. I wish feedback to the programers was possible because there is a real need to be able for the arbitrator to insert a record in either the A or B indexing to clean up a missed name. My wife is an arbitrator and has spent up to 4 hours cleaning up the messes where A missed 4 records and B missed 7 all in the early part of a 375 record batch. If she could insert a Record in them they would have cleaned up quickly.

That's a weakness in both the software and the process. It is solved by fixes to the software and / or process.
1) Let an arbitrator return a set for "reconsideration"
2) Let the arbitrator insert records.

I've done a couple of Irish death indexes. They get tedious because they're so long. I was very careful, but when I checked by column, I found at least one error in every column, so it's easy to make a mistake, especially leaving out a line.

srweight wrote:The Irish Death lists have about 375 names in them and no line numbers. When one or both indexers get off the arbitrator has to fix the mess. I wish feedback to the programers was possible because there is a real need to be able for the arbitrator to insert a record in either the A or B indexing to clean up a missed name. My wife is an arbitrator and has spent up to 4 hours cleaning up the messes where A missed 4 records and B missed 7 all in the early part of a 375 record batch. If she could insert a Record in them they would have cleaned up quickly.

There is indeed a way to insert records and work with the record matching functions of the program. It's all under the Record Matching tab.

I likewise am disappointed with the lack of feedback on familysearchindexing. I have been a participant on a number of distributed computing and distributed work projects (like Project Gutenberg). Over and over, the most successful of these ventures are those that provide frequent feedback about progress. Some simple daily counts are a place to start. For example, freerice.com has a simple count of the number of grains of rice contributed in the past several days. I would like to see something very simple on the indexing home page that just showed three daily counts: number of records indexed, number of names indexed, and number of indexers who submitted. These could be graphed out showing daily, weekly, and monthly trends.

The bi-monthly memorandum that goes out is not adequate. For one thing it is too Church-centric. It will make no sense to folks who are not members of the Church and it is too out-dated. With the call invitation going out to all to join the indexing effort, we need to be more inclusive with our communication, especially about progress.

The 'projects completed' page is a good start. But I would like to see a 'progress' page that showed the number of records and number of names submitted on each page. No more math than simple counts would be sufficient to let indexers know what they are part of.

It seems that for an organization as obsessed with counting and reporting, we could be doing much better in this arena.

Has anyone considered nominating Family Search Indexing for a Duke's Choice Award?

Distributed computing and especially human computing are pretty hot right now and Family Search Indexing is the most substantial example of this that I'm aware of. It struck me that Family Search Indexing was rather apt for nomination and that it might result in some good publicity.

http://java.sun.com/community/dukeaward ... m_new.html

I personally would agree that statistics would great . . . maybe not on an individual basis.

I would like to see more publicity as well to get more involved whether within the church or without. I know that I have seen many projects out there trying to do similar things and have tried to get some involved in the indexing.

Finally, I would like to see more feedback from my own indexing. I can learn from the errors that I make especially if I find that I make the same mistake over and over again.

I would like to know if it would be possible to use the data extracted during the old extraction program, rather that indexing the same films again. Many of those extracted films are birth, christening and marriage records which I consider the backbone of genealogy. If those could be put on line in additon to the projects currently being indexed, it would make a huge dent in the need for indexing of films. Also, many of the extracted records not in the IGI were made available in the Vital Records Index series of CDs. And I assume there are also many that have been extracted that have not yet been made publicly available - i.e. those being used for temple ordinance projects. These extraction projects represents years of work already done, if there is a way to connect the data with the film from which it was extracted.

Now, that would be awesome! If we could get the old extracted data connected with the film in indexing. One problem that may exist may be that the current indexing projects may index more data than the old extraction projects.