How can we address BOINC's retention problem?

Message boards : Promotion : How can we address BOINC's retention problem?
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Sid
Avatar

Send message
Joined: 29 Jun 08
Posts: 30
United States
Message 18248 - Posted: 5 Jul 2008, 15:11:38 UTC




Currently BOINC boasts of 1,417,601 registered members, but only 342,528 can be categorized as active.

That a problem that needs to be addressed.
If everyone is thinking alike, someone isn't thinking. - General George Patton



ID: 18248 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15002
Netherlands
Message 18250 - Posted: 5 Jul 2008, 15:42:41 UTC - in response to Message 18248.  

Perhaps when electricity isn't so expensive, people aren't going on summer holidays, houses aren't washed away by hurricanes/tornados/levees breaking, etc. etc.

Otherwise, why?
ID: 18250 · Report as offensive
Sid
Avatar

Send message
Joined: 29 Jun 08
Posts: 30
United States
Message 18251 - Posted: 5 Jul 2008, 16:04:41 UTC - in response to Message 18250.  

Perhaps when electricity isn't so expensive, people aren't going on summer holidays, houses aren't washed away by hurricanes/tornados/levees breaking, etc. etc.

Otherwise, why?




Seems to me that it is futile to promote BOINC only to watch 75% leave.

Seriously, we need to explore why roughly 3 out of 4 people who register for BOINC just walk away from it.


If everyone is thinking alike, someone isn't thinking. - General George Patton



ID: 18251 · Report as offensive
John37309
Avatar

Send message
Joined: 28 Jul 07
Posts: 91
Ireland
Message 18255 - Posted: 6 Jul 2008, 1:42:48 UTC - in response to Message 18251.  
Last modified: 6 Jul 2008, 2:07:51 UTC

Perhaps when electricity isn't so expensive, people aren't going on summer holidays, houses aren't washed away by hurricanes/tornados/levees breaking, etc. etc.

Otherwise, why?




Seems to me that it is futile to promote BOINC only to watch 75% leave.

Seriously, we need to explore why roughly 3 out of 4 people who register for BOINC just walk away from it.


I agree 100% Sid. This is a very serious issue and it needs to be addressed, its not a joke.

This topic was briefly discussed last year at the BOINC workshop in Geneva and Bruce Allen of the Einstein@home project actually had a more accurate figure on this. Bruce said that the average retention rate from the Einstein project was about 30 days. That means that the average user signs up to Einstein and quits crunching roughly 30 days after starting. This figure of about 30 days was confirmed by other projects also.

Bill Bovermann of the world community grid project said they were definitely getting a much higher retention rate by sending out a "Please come back" email to users about 30 days after they stopped crunching.

I think we should compile as much information about this topic as we can. We need to think of all the possible reasons why people leave BOINC projects and study the information. If we can make some sense out of our collection of ideas, we can see what solutions we can offer. Then we write it out properly and add it to the trac wiki for project admins to find.

We need be methodical about this. There is no point in asking someone else to do it. We need to do the work here otherwise this chat area just becomes a talking shop where nothing gets done. So lets compile this list with everything we can think of that might be a possible reason for people leaving BOINC projects.

This is the current trac wiki page dealing with this topic; http://boinc.berkeley.edu/trac/wiki/VolunteerRecruit

Sid, i like the way you called this thread "How can we address BOINC's retention problem?". How can we address the problem is a very productive way to look at things!

John.
| Irelandboinc.com | PaddysInSpace.com
ID: 18255 · Report as offensive
Profile Ananas

Send message
Joined: 27 Jun 06
Posts: 305
Germany
Message 18257 - Posted: 6 Jul 2008, 8:37:37 UTC

One thing might be helpful to know :

Is there a connection between the time people stay with a project and their membership to a team?

Do people in teams stay longer than single crunchers or is there no difference?

If team people stay longer (in average), projects could focus more on presenting teams and encouraging people to join teams.
ID: 18257 · Report as offensive
Sid
Avatar

Send message
Joined: 29 Jun 08
Posts: 30
United States
Message 18259 - Posted: 6 Jul 2008, 14:09:19 UTC - in response to Message 18255.  

I think we should compile as much information about this topic as we can. We need to think of all the possible reasons why people leave BOINC projects and study the information.



John:


I believe we would do better to examine why people get involved in BOINCing in the first place.

I believe that people get involved in BOINC projects with many motives, some being:



  • They are interested, even excited, about the underlying science.

  • They feel that they are making a difference and participating in real, relevant scientific research.

  • They want to contribute to worthwhile causes and BOINC provides an avenue.

  • They are technojunkies and BOINC proves that size matters.

  • They have time and computers on their hands, and BOINC provides a little bubblegum for the mind and excuses to buy more computers.

  • The want to explore online social networking and BOINC teams stroke those people.

  • They join teams to network with people with similar interests and to participate in competitions.




I believe the real reasons why people get into BOINCing resolve down to their desire to participate, to contribute, to come away with a sense of individual accomplishment and sense of purpose.


The more we can identify and stroke these motivating factors and minimize the negative factors, the higher our retention factors will be.


If everyone is thinking alike, someone isn't thinking. - General George Patton



ID: 18259 · Report as offensive
John37309
Avatar

Send message
Joined: 28 Jul 07
Posts: 91
Ireland
Message 18262 - Posted: 6 Jul 2008, 23:05:33 UTC - in response to Message 18257.  
Last modified: 6 Jul 2008, 23:16:00 UTC

One thing might be helpful to know :

Is there a connection between the time people stay with a project and their membership to a team?

Do people in teams stay longer than single crunchers or is there no difference?

If team people stay longer (in average), projects could focus more on presenting teams and encouraging people to join teams.

Ananas,
Its a good point. Around this time last year there was a big push from David A and a collection of team founders to try to improve the "Tools" that team founders have, to see if we could improve the way teams work. Many good new features were implemented as a result of the discussions but things fizzled out and development in this area stopped.

I think it very much varies from team to team, i have no way to measure the statistic but i do think teams are more likely to create "something extra" to keep people involved and crunching.
John:


I believe we would do better to examine why people get involved in BOINCing in the first place.

Sid,
These are two separate things and when trying to learn from this, we should not confuse the 2 issues of "What attracts you in the first place", and "Why people leave".

So this is a list of reasons why people join and is separate to the retention issue. Sid i think you should post this as a separate topic, we need to chat more about why people get started in the first place. We need to keep the topics separate.

I believe that people get involved in BOINC projects with many motives, some being:



  • They are interested, even excited, about the underlying science.

  • They feel that they are making a difference and participating in real, relevant scientific research.

  • They want to contribute to worthwhile causes and BOINC provides an avenue.

  • They are technojunkies and BOINC proves that size matters.

  • They have time and computers on their hands, and BOINC provides a little bubblegum for the mind and excuses to buy more computers.

  • The want to explore online social networking and BOINC teams stroke those people.

  • They join teams to network with people with similar interests and to participate in competitions.




I believe the real reasons why people get into BOINCing resolve down to their desire to participate, to contribute, to come away with a sense of individual accomplishment and sense of purpose.


The more we can identify and stroke these motivating factors and minimize the negative factors, the higher our retention factors will be.



It is difficult to separate the two issues. But many of the things you have listed will be on both lists because "Getting started" and "Retention" are interlinked.
| Irelandboinc.com | PaddysInSpace.com
ID: 18262 · Report as offensive
John37309
Avatar

Send message
Joined: 28 Jul 07
Posts: 91
Ireland
Message 18263 - Posted: 7 Jul 2008, 3:12:54 UTC
Last modified: 7 Jul 2008, 3:16:46 UTC

Here is a link to a message where Paul D. Buck talks a lot about what we are discussing here in this message; http://setiathome.berkeley.edu/forum_thread.php?id=46181

Same message, this particular post is good; http://setiathome.berkeley.edu/forum_thread.php?id=46181&nowrap=true#739206

Its worth reading!

John.
| Irelandboinc.com | PaddysInSpace.com
ID: 18263 · Report as offensive
SekeRob

Send message
Joined: 25 Aug 06
Posts: 1596
Message 18264 - Posted: 7 Jul 2008, 10:11:50 UTC - in response to Message 18263.  

Actually, I think WCG sends out a reminder once after 3 weeks of result return cessation. Quarterly Newsletters are important and progress updates from the scientists and a little homework to give that forward going feel:



Also, i think the standard BOINC projects thoroughly fail mostly to convey how much work was done by the members.... not just the credits.... tell of the results and the cpu time.... tflops is such a hmmm value for anyone not into computing.

Some teams have automated their member contributions monitoring and send a reminder/warning mail, which is helpful particularly for those that run BOINC as a set-and-forget service.

just a few cents worth
Coelum Non Animum Mutant, Qui Trans Mare Currunt
ID: 18264 · Report as offensive
Sid
Avatar

Send message
Joined: 29 Jun 08
Posts: 30
United States
Message 18265 - Posted: 7 Jul 2008, 12:05:48 UTC - in response to Message 18264.  
Last modified: 7 Jul 2008, 12:06:27 UTC

Also, i think the standard BOINC projects thoroughly fail mostly to convey how much work was done by the members.... not just the credits.... tell of the results and the cpu time.... tflops is such a hmmm value for anyone not into computing.



Agreed, the average BOINCer must realize that HIS/HER contribution [and several 100K like them], even if it is only a few hours a day on a single core computer, is the real strength of the BOINC system.

The highly visible and vocal group of frothers with enough overclocked servers to heat an office building form a tiny minority.

This vast majority of BOINC participants need to be convinced that their contribution matters and that participation, contribution and a feeling that what they are doing is significant are what is important. . .

. . . leave the back alley, broken bottle competitions to the few who are so inclined.
ID: 18265 · Report as offensive
John37309
Avatar

Send message
Joined: 28 Jul 07
Posts: 91
Ireland
Message 18285 - Posted: 8 Jul 2008, 1:31:19 UTC
Last modified: 8 Jul 2008, 1:32:26 UTC

OK, i will start a list of reasons why people leave BOINC projects. You can add to the list by replying to the message and adding what you think.

The first 5 reasons i have listed here are direct feedback from a user that quit SETI@home. He ran SETI for 2 and a half months and quit in May 2007. I just spoke to him a few minutes ago.

Reasons why people leave or quit BOINC projects!
1. Uses too much resources - I need them for other things
2. Not exciting – no competition
3. Nothing to gain – nothing in it for me
4. Electricity cost - too expensive
5. Don’t believe in the science or don't really care
6. Too complicated - could not get it to work properly
| Irelandboinc.com | PaddysInSpace.com
ID: 18285 · Report as offensive
SekeRob

Send message
Joined: 25 Aug 06
Posts: 1596
Message 18288 - Posted: 8 Jul 2008, 7:33:06 UTC - in response to Message 18285.  
Last modified: 8 Jul 2008, 7:35:04 UTC

OK, i will start a list of reasons why people leave BOINC projects. You can add to the list by replying to the message and adding what you think.

The first 5 reasons i have listed here are direct feedback from a user that quit SETI@home. He ran SETI for 2 and a half months and quit in May 2007. I just spoke to him a few minutes ago.

Reasons why people leave or quit BOINC projects!
1. Uses too much resources - I need them for other things
2. Not exciting – no competition
3. Nothing to gain – nothing in it for me
4. Electricity cost - too expensive
5. Don’t believe in the science or don't really care
6. Too complicated - could not get it to work properly

7. Purpose of a project, humanitarian or math or space science
8. Project updates absence is for sure a consideration. What to expect of e.g. SETI.... yesterday nothing interesting reporting, forecast, not expecting much today!
9. Downtime frequency and duration
10. Security issues with AV & Firewall or just plain concerns (extension of 6)
11. Public domain or for profit unclarity.
12. GM concerns.
Coelum Non Animum Mutant, Qui Trans Mare Currunt
ID: 18288 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15002
Netherlands
Message 18293 - Posted: 8 Jul 2008, 15:20:26 UTC - in response to Message 18288.  

The negative ones then.

13. Fighting/flaming on the forums (and through PM and email).
14. Getting banished (for whatever reason fits).
15. Getting spammed.
16. Invisible project developers/administrator.
17. Over-intrusive Administrator.
18. Over-intrusive moderation.
19. Constant losing of credits.
20. Not enough credits.
ID: 18293 · Report as offensive
Chris Sutton

Send message
Joined: 29 Aug 05
Posts: 117
Message 18296 - Posted: 8 Jul 2008, 18:54:49 UTC - in response to Message 18285.  

OK, i will start a list of reasons why people leave BOINC projects. You can add to the list by replying to the message and adding what you think.

You might find some of the data collected here Survey results useful.
ID: 18296 · Report as offensive
Sid
Avatar

Send message
Joined: 29 Jun 08
Posts: 30
United States
Message 18297 - Posted: 8 Jul 2008, 19:13:59 UTC - in response to Message 18296.  

OK, i will start a list of reasons why people leave BOINC projects. You can add to the list by replying to the message and adding what you think.

You might find some of the data collected here Survey results useful.



Thanks for that link, here's a pearl:


3:05 PM UTC, February 14 2008

1. I had a TERRIBLE time figuring out which projects would run on my Mac. I had to visit each site, check specs, and even then it wasn't always clear, and STILL isn't. is there an 'official' list of which BOINC projects run on Macs and if so, WHERE IS IT??? If none, PLEASE create and publish as a bulleted item on your web site - in this PC-centric world, it would be a real boon to new and even older contributors. This simple but chronically overlooked issue NEEDS to be CLEARed up, and only YOU can do it! (I can't even find this issue adequately addressed on any BOINC help site, either.)

2. Most projects are puzzlingly SLOW to credit work to my machine. It simply should not take sometimes a week or more to turn this data around and post it. Sometimes I wonder whether my results have really gotten through.

3. It isn't clear to me how fair and balanced the crediting system among projects really is. Are there standards which BOINC enforces?

4. I have shut some projects down because they feed me enormous, choking chunks of data that take literally many days of computing time to complete. If a computational error occurs, all that computer time is lost. Projects very much need to download data segments that take NO MORE than maybe half a day to complete. If an unfortunate error occurs, at least all of the other running projects don't suffer as well from the time lost.

5. Thanks and regards, Darryl W. Dockins dwdockins@charter.net



Some key points that realy need more examination:


I really didn't realize that BOINC was that problematic to the Apple delegation.

Projects that drag on for weeks before assigning credits to finished work.

I agree, any WU that runs over 12 hours is a downer.

Some sanity with BOINC credits needs to be enforced.

If everyone is thinking alike, someone isn't thinking. - General George Patton



ID: 18297 · Report as offensive
enels

Send message
Joined: 3 Jul 08
Posts: 6
United States
Message 18298 - Posted: 8 Jul 2008, 20:03:31 UTC

I've also lately been thinking about the poor retention rate of DC participants. It could be the electricity costs drive people away. 30 days is long enough for Americans at least to get an electricity bill. 100 watts 24/7 would cost about $7-10 month. Many people will notice this increase.

Even if they can afford the electricity, they might be turned off by the resulting heat. I've read in more than one forum people turning off their machines because their computer room is just too hot. I can understand this. Some years I've quit participating in DC projects when the air conditioning is running.

Even if they can afford the electricity and stand the heat, they may be turned off by the resulting fan noise. Instead of building computers to be relatively quiet at all times, computer makers have designed computers to be relatively quiet when idle. The fan noise some computers emit when running full blast is simply unbearable for many people.

But I am speculating. I don't know of any survey of ex DC participants inquiring about why they stopped. But it is a great question and something that should be investigated. Something attracted them to DC, but why only for 30 days?
ID: 18298 · Report as offensive
Profile KSMarksPsych
Avatar

Send message
Joined: 30 Oct 05
Posts: 1239
United States
Message 18300 - Posted: 8 Jul 2008, 22:50:51 UTC - in response to Message 18297.  

I agree, any WU that runs over 12 hours is a downer.



The science done dictates WU length. Some things can't be broken into nice 1 hour chunks. CPDN is the best example, but there are others out there. QMC can have some massive work as well as Orbit (so I'm told).

It would be nice if projects would make this clear, but then again, some people today seem to have problems reading what is right in front of their noses.

Kathryn :o)
ID: 18300 · Report as offensive
FangRust

Send message
Joined: 14 Jul 08
Posts: 19
United States
Message 18559 - Posted: 17 Jul 2008, 3:02:15 UTC

“The science done dictates WU length. Some things can't be broken into nice 1 hour chunks. CPDN is the best example, but there are others out there. QMC can have some massive work as well as Orbit (so I'm told).

It would be nice if projects would make this clear, but then again, some people today seem to have problems reading what is right in front of their noses.”


Why not form “Massive WU Teams”? Have the org say we need x computers for y time and have a pool of people who will donate 24/7 for the duration. Shirly there are enough volunteers to manage short term goals.

Pro-active, interactive, some kind of “active” is necessary to keep people interested. Look at how many skins are available for the BOINC manager. Clubs, groups, orgs, etc. have the same age old problem: the same people volunteer over and over and over and… to fill the positions necessary for an org to function.

A lot of people are cut off at the ankles with the “tecky” complexity that exists with BOINC, Grid Republic, BAM, etc. Quite a few letters say so or complain about WU not being available or other issues.

One thing is certain, it will never get done if it’s just talk. Come up with a game plan. Implement it. Fix it along the way. Get it done.

Just a thought.
ID: 18559 · Report as offensive
John37309
Avatar

Send message
Joined: 28 Jul 07
Posts: 91
Ireland
Message 18706 - Posted: 21 Jul 2008, 3:24:47 UTC - in response to Message 18559.  


Pro-active, interactive, some kind of “active” is necessary to keep people interested. Look at how many skins are available for the BOINC manager. Clubs, groups, orgs, etc. have the same age old problem: the same people volunteer over and over and over and… to fill the positions necessary for an org to function.

A lot of people are cut off at the ankles with the “tecky” complexity that exists with BOINC, Grid Republic, BAM, etc. Quite a few letters say so or complain about WU not being available or other issues.

One thing is certain, it will never get done if it’s just talk. Come up with a game plan. Implement it. Fix it along the way. Get it done.

Just a thought.

I currently working on the pro-active, interactive element. This "Promotion" forum is a good start.

I'm working on several idea's at the moment. They will be released soon.

John.

| Irelandboinc.com | PaddysInSpace.com
ID: 18706 · Report as offensive
Sid
Avatar

Send message
Joined: 29 Jun 08
Posts: 30
United States
Message 20723 - Posted: 9 Oct 2008, 16:21:13 UTC - in response to Message 18297.  
Last modified: 9 Oct 2008, 16:22:27 UTC


3. It isn't clear to me how fair and balanced the crediting system among projects really is. Are there standards which BOINC enforces?

Some sanity with BOINC credits needs to be enforced.





The subject of cross-project parity is too often just another flame war.

A Cobblestone of computation is supposedly 0.86 trillion floating-point operations, but the reality bears only casual resemblance to the theory.

I thought that the abuses of Reisel Sieve and Cosmology would be enough to bring some attention to this problem. . .

. . . but Ramsey confirms that it has just gotten worse.

I have been having some success in team building, trying to marshal team members to participate in crunching competitions only to be told that they couldn't tolerate the loss some projects would mean to their RAC and TC standings. . .

Apparently before the project credits are applied to overall statistics, a correction factor needs to be applied.

. . . only then would overall standings have any credibility.


Until this elephant in the living room is addressed, the BOINC enterprise suffers greatly.
If everyone is thinking alike, someone isn't thinking. - General George Patton



ID: 20723 · Report as offensive
1 · 2 · Next

Message boards : Promotion : How can we address BOINC's retention problem?

Copyright © 2022 University of California. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.