[UX]Counting questions and answers for EN forums

[RGB ES <rg...@gmail.com>]

A really quick "study" based on number of post on the EN forums that
shows, IMO, some interesting information.

Fist the numbers, then some "analysis" on those numbers. Note, long
post below...

========================================

Number of Topics: T
Number of Posts: P
Ratio Posts/Topics: R

Classified By App

Writer
T = 11984
P = 61938
R = 5.17

Calc
T = 9072
P = 45646
R = 5.03

Base
T = 4858
P = 22895
R = 4.71

Impress
T = 1718
P = 7197
R = 4.19

Draw
T = 634
P = 2930
R = 4.62

Math
T = 198
P = 1009
R = 5.1

Particular Forums

Beginners
T = 502
P = 2200
R = 4.38

Setup and Troubleshooting
T = 10048 (MS: 4058; Linux: 1324; OSX: 1119; non classified: 3547)
P = 47644 (MS: 19362; Linux: 6559; OSX: 5352; non classified: 16371)
R = 4.74 (MS: 4.77; Linux: 4.95; OSX: 4.78; non classified: 4.62)

Macro
T = 3423
P = 15571
R = 4.55

Discussion Forums

General Discussion
T = 1017
P = 6283
R = 6.18

Site Feedback (including Forum Governance)
T = 344
P = 2635
R = 7.66


Totals
T = 44468
P = 225242
R = 5.07

========================================

Some quick observations:

- The most used forums are Writer, then Setup and Troubleshooting,
then Calc, then a big gap, then the rest
- There are around 5 messages on each thread on forums dedicated to
the different apps
- Setup problems on Linux need more work than on the other platforms
- Beginners forum have the lowest Post/Topic ratio: 4.38
- Discussion forums have the highest Post/Topic ratios: 6.18 and 7.66

Some quick conclusions from these observations:

- The Post/Topic ratio is always high, which means that interaction
with users is always needed. Solving AOO problems always need a lot of
feedback and investigation. The immediate conclusion is that systems
like stack overflow, yahoo answers and similar will never work here:
those systems are useful when questions have unique answers that you
can evaluate, vote and classify, not when feedback (more information,
sample documents, different trials...) is needed.

- Setup and troubleshooting numbers are perhaps too high: this show
that users have big problems not only using the software, but on
making it run properly. To have so many installation problems, spell
check problems, crash problems... is a bad sign: to take the software
and make it run should be the easiest part. But on the positive side
we have that the ratio is lower than on the other forums: setup
problems are easier to solve than, say, Writer problems.

- As already mentioned, the ratio for Linux setup problems is the
highest. That's not necessarily a problem with our Linux builds: Linux
distros always made a big mess when packaging their highly patched
versions of OOo and now LibO. Problems to install AOO on ubuntu or
fedora because the way they packaged LibO is a good example.

- The low ratio on Beginners forum is a good signal: it means that
real beginners come with simple problems that have an easy solution.
More advanced users have more demanding problems that need more
interaction resulting on the higher ratio on the other forums. But
perhaps the ratio en Beginners forum should be even lower: ideally,
simple problems should have a ratio close to 2 (one question + one
answer).

- Higher ratios on discussion forums is good, but maybe not good
enough: most forum users do not engage on discussions, they are only
interested on solving their problems.

- Something else?

Regards
Ricardo

Re: [UX]Counting questions and answers for EN forums

[RGB ES <rg...@gmail.com>]

2012/6/27 Rob Weir <ro...@apache.org>:
>
> What time period are these counts for?  All time?


All time. Those numbers are a sort of "snapshot" of today situation.


>> Some quick observations:
>>
>> - The most used forums are Writer, then Setup and Troubleshooting,
>> then Calc, then a big gap, then the rest
>> - There are around 5 messages on each thread on forums dedicated to
>> the different apps
>> - Setup problems on Linux need more work than on the other platforms
>> - Beginners forum have the lowest Post/Topic ratio: 4.38
>> - Discussion forums have the highest Post/Topic ratios: 6.18 and 7.66
>>
>
> Do you have a sense of whether these ratios are different with AOO
> than they were with OOo?


Difficult to say. The forum is working since 2007 and from an user
point of view AOO have only a few weeks, so I do not expect big
changes on the short term. But beside the quickstarter bug, the
problems with English thesaurus and the lack of en_GB localization, I
don't see "unusual" threads, which is a Good Thing.


>> Some quick conclusions from these observations:
>>
>> - The Post/Topic ratio is always high, which means that interaction
>> with users is always needed. Solving AOO problems always need a lot of
>> feedback and investigation. The immediate conclusion is that systems
>> like stack overflow, yahoo answers and similar will never work here:
>> those systems are useful when questions have unique answers that you
>> can evaluate, vote and classify, not when feedback (more information,
>> sample documents, different trials...) is needed.
>>
>
> It is an interesting point.  However, StackExchange allows
> clarification of the original question, as well as multiple
> "solutions" that are then rated.  But that only suits some problems.
> For example, it would work well for questions about how best code a
> particular macro task.  But not so good for something that needs
> debugging, as you note..  On the other hand, a site that rates
> previous answers and gives them greater position in search might help
> the user solve their own problem without even needed to make a post.
> But that is hard to measure.


Sadly, on my experience "normal users" do not search. At all. So I
don't think a rating system will help.


>> - Setup and troubleshooting numbers are perhaps too high: this show
>> that users have big problems not only using the software, but on
>> making it run properly. To have so many installation problems, spell
>> check problems, crash problems... is a bad sign: to take the software
>> and make it run should be the easiest part. But on the positive side
>> we have that the ratio is lower than on the other forums: setup
>> problems are easier to solve than, say, Writer problems.
>>
>
> Is it worth collecting the "top 10" issues with 3.4 and making an FAQ
> for this?  We could put it on the download page and the support page
> and pin it as a topic on the forums.


Beside some annoyances that will be fixed on 3.4.1, fortunately we
don't have "new problems" on 3.4. But of course a list with the "most
common problems and their solutions" is a good idea to consider.

BTW: on EN forum there is from now a new section called "User
Experience (UX)" so it's time to increase the ratio on the discussion
forums... :)

http://user.services.openoffice.org/en/forum/viewforum.php?f=106

Regards
Ricardo

Re: [UX]Counting questions and answers for EN forums

[Rob Weir <ro...@apache.org>]

On Wed, Jun 27, 2012 at 2:53 PM, RGB ES <rg...@gmail.com> wrote:
> A really quick "study" based on number of post on the EN forums that
> shows, IMO, some interesting information.
>
> Fist the numbers, then some "analysis" on those numbers. Note, long
> post below...
>
> ========================================
>
> Number of Topics: T
> Number of Posts: P
> Ratio Posts/Topics: R
>

What time period are these counts for?  All time?

> Classified By App
>
> Writer
> T = 11984
> P = 61938
> R = 5.17
>
> Calc
> T = 9072
> P = 45646
> R = 5.03
>
> Base
> T = 4858
> P = 22895
> R = 4.71
>
> Impress
> T = 1718
> P = 7197
> R = 4.19
>
> Draw
> T = 634
> P = 2930
> R = 4.62
>
> Math
> T = 198
> P = 1009
> R = 5.1
>
> Particular Forums
>
> Beginners
> T = 502
> P = 2200
> R = 4.38
>
> Setup and Troubleshooting
> T = 10048 (MS: 4058; Linux: 1324; OSX: 1119; non classified: 3547)
> P = 47644 (MS: 19362; Linux: 6559; OSX: 5352; non classified: 16371)
> R = 4.74 (MS: 4.77; Linux: 4.95; OSX: 4.78; non classified: 4.62)
>
> Macro
> T = 3423
> P = 15571
> R = 4.55
>
> Discussion Forums
>
> General Discussion
> T = 1017
> P = 6283
> R = 6.18
>
> Site Feedback (including Forum Governance)
> T = 344
> P = 2635
> R = 7.66
>
>
> Totals
> T = 44468
> P = 225242
> R = 5.07
>
> ========================================
>
> Some quick observations:
>
> - The most used forums are Writer, then Setup and Troubleshooting,
> then Calc, then a big gap, then the rest
> - There are around 5 messages on each thread on forums dedicated to
> the different apps
> - Setup problems on Linux need more work than on the other platforms
> - Beginners forum have the lowest Post/Topic ratio: 4.38
> - Discussion forums have the highest Post/Topic ratios: 6.18 and 7.66
>

Do you have a sense of whether these ratios are different with AOO
than they were with OOo?

> Some quick conclusions from these observations:
>
> - The Post/Topic ratio is always high, which means that interaction
> with users is always needed. Solving AOO problems always need a lot of
> feedback and investigation. The immediate conclusion is that systems
> like stack overflow, yahoo answers and similar will never work here:
> those systems are useful when questions have unique answers that you
> can evaluate, vote and classify, not when feedback (more information,
> sample documents, different trials...) is needed.
>

It is an interesting point.  However, StackExchange allows
clarification of the original question, as well as multiple
"solutions" that are then rated.  But that only suits some problems.
For example, it would work well for questions about how best code a
particular macro task.  But not so good for something that needs
debugging, as you note..  On the other hand, a site that rates
previous answers and gives them greater position in search might help
the user solve their own problem without even needed to make a post.
But that is hard to measure.


> - Setup and troubleshooting numbers are perhaps too high: this show
> that users have big problems not only using the software, but on
> making it run properly. To have so many installation problems, spell
> check problems, crash problems... is a bad sign: to take the software
> and make it run should be the easiest part. But on the positive side
> we have that the ratio is lower than on the other forums: setup
> problems are easier to solve than, say, Writer problems.
>

Is it worth collecting the "top 10" issues with 3.4 and making an FAQ
for this?  We could put it on the download page and the support page
and pin it as a topic on the forums.


> - As already mentioned, the ratio for Linux setup problems is the
> highest. That's not necessarily a problem with our Linux builds: Linux
> distros always made a big mess when packaging their highly patched
> versions of OOo and now LibO. Problems to install AOO on ubuntu or
> fedora because the way they packaged LibO is a good example.
>
> - The low ratio on Beginners forum is a good signal: it means that
> real beginners come with simple problems that have an easy solution.
> More advanced users have more demanding problems that need more
> interaction resulting on the higher ratio on the other forums. But
> perhaps the ratio en Beginners forum should be even lower: ideally,
> simple problems should have a ratio close to 2 (one question + one
> answer).
>
> - Higher ratios on discussion forums is good, but maybe not good
> enough: most forum users do not engage on discussions, they are only
> interested on solving their problems.
>
> - Something else?
>

Interesting data, thanks.

-Rob

> Regards
> Ricardo