[rc5] gaining and retaining users

andrew meggs insect at antennahead.com
Fri Jun 20 23:40:53 EDT 1997


At 6:26 PM -0700 6/20/97, TC Lai wrote:

>With regards to personal interest, I can't comment for other people, but
>the DESChallenge effort kept on gathering steam, even after several months
>of key-crunching.

My gut instinct says that unless changes are made, bovine will have more of
a problem with this than deschall. I just find the web site sort rather
confusing, relative to deschall's and solnet's. The statistics, especially,
seem unclear and fragmented (for example, the "Executive Summary" item
under "Hot Links!" suggests "white paper" to me, not "aggregate
statistics"), and overall I feel like I'm getting too much information that
isn't important and not enough interesting data. Since progress statistics
are the only form of positive feedback that we can give users to hold their
interest and create excitement, this is a bad thing.

>  I was running clients for over a month, maybe two, and
>while many people did drop out, a good number stayed in and many others
>joined.  If you look at the DESChall stats, the general trend was to
>continually shorten the average time to 50% completion.

I'm including something below that went out internally on the
deschall_developer list. I think all of the issues raised apply equally to
bovine.

.............begin quoted text............

With new clients (finally!) looking ready to go, one of the things that I
would now be working on if this were at my "real" job would be to put
together a profile of a typical user experience using a focus group and a
bit of role-playing on my part. (My degree is out of the psych department,
not CS.)

We don't have a focus group, but here's what I see when I play "novice user":

A small percentage of the people who hear about deschall are going to get
extremely interested in it and subscribe to the mailinglist. A small
percentage of those subscribers will participate in the discussion. A
somewhat larger percentage will follow along but not post. But most of the
subscribers will find the discussion to be a bit over their head, or find
it boring when the majority of the messages are about other platforms (even
Win95 occupies less than 50% of the discussion), or simply not like the
volume of mail, and before long they'll fire off an unsubscribe message to
either majordomo or the posting address itself.

To us, the developers, the tiny subset of those people who 1) subscribe to
the list, 2) stay on the list, and 3) actively post to the list, might
sometimes seem to represent the pool of client operators because of their
visibility, but we should never lose sight of the fact that they aren't --
they're the hardest of the hardcore. The people who stay on the list
receive a constant barrage of encouraging progress reports. The people who
get off the list or more likely never subscribe to even the announce list
don't have anything but this little program that merely spits out dots or
fills progress bars, and they're going to make up the majority of client
downloads.

Dots and bars do very little to hold one's interest. And that's a problem,
because the people who have the least interest in the project to begin with
are the people receiving nothing but this minimal encouragement and
feedback. It's especially a problem because these non-hardcore users are
very likely to be on single-user GUI-based machines that get rebooted
regularly, so the client will be visible when it's running, and require the
user to actively restart it regularly for it to be running at all. They
aren't on multiuser remote-login systems that stay up for months, where you
can just nohup the client and forget about it.

So, we likely have a very big retention problem. Looking at the web page, I
note that over 50,000 IP addresses have participated since the start, yet
only 10,000 did so in the last 24 hours. I can't do a log analysis to
estimate how much of that is due to changing dialup IP's, but even if we
write off 75% of it to dialup, that would still mean we've lost over half
of our clients during the course of the project.

What can we do about this? First, we need to make it as easy for users to
get clients running and *keep* them running as possible. The Read Me for
the MacOS update includes ultra-simplified instructions for setting up the
program to re-launch automatically on bootup; do the clients for other
frequently-rebooted OSes include such instructions? Would including DESGUI
directly in the Windows zipfile be good or bad?

Second, we need to keep people interested and encouraged, so they *want* to
keep their clients running. A weekly report of "We're xx% done and expect
to finish in y weeks, GO TEAM GO!" to the announce list might help,
especially if we managed to get a lot more of the non-hardcore people
subscribed to it. A quick blank on the client download form to fill in your
email address and automatically subscribe could probably get a lot of names
added.

Also, we've already got a way that's less obtrusive than a mass email to
send regular encouragement to 100% of the client operators. I'm talking
about "M" messages from the server.

  [ note -- M messages were a way of making clients display
    human-readable text messages. Deschall never used them,
    but I think we should have, if only to announce the
    availability of client updates. ]

                                    If we decided to use that method, there
would be an extremely delicate balance between giving encouragement and
giving users a big-brother style propaganda ticker, but I think it could be
managed. Positive factual reports like "The entire project checked nnn
trillion keys in the last 24 hours" or "Someone is expected to find the
winning key within nn days" would probably be safe, although I tend to find
even the implied demand for support in "Thanks for your support" to be
potentially offensive. Simply keeping people aware that the program over in
the corner was involved in something larger than spewing out dots is very
likely the most that should be attempted through this kind of means, but
even that could be very effective. At the very least I think the increased
retention of the people who don't especially care about deschall would far
outweigh the increased erosion of militia members who found the messages
ominous, since the apathetic come in far greater numbers.

My apologies for the long-winded letter, but this is sort of what I do
professionally, and a 5:1 ratio between total clients and current clients
demands some attention to our retention. Do other people have thoughts on
these issues?


____________________________________________________________________________
Andrew Meggs, content provider                  Antennahead Industries, Inc.
<mailto:insect at antennahead.com>                 <http://www.antennahead.com>


----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list