Re: [EE] Positive Notifications for Echo - EE

5 Feb 2013

Hi everyone,

Thank you for all your good suggestions and comments about positive notifications for
Echo, which mean a lot to us!

To sum up our discussion so far, here some of the key points and ideas we have heard in
this email thread and in offline conversations.

1. Positive notification ideas 
The positive notifications we have discussed so far fall in 7 main categories, based on
what triggers them. 

Here's a short list of positive notification types (see full list below):

* Contributions ('100 edits to a page you edited.')

* Contributors ('20 people contributed to a page you edited.')

* Mark as useful ('someone marked your edit as useful.')

* Mark as non-problematic ('Cluebot did not find a problem with your edit.')

* Pageviews ('a page you edited was viewed 1,000 times.')

* Visitors ('42 people read the page you edited.') 

* Feedback ('10 positive comments were posted on a page you edited.')

While most of these notifications could provide positive reinforcement to new users, some
of them are easier to do and more likely to work on all sites, as outlined below.

2. Some ideas are hard to build

Sadly, some of these ideas require a lot more development than we can afford for our first
release next month (we are looking for solutions that can be implemented in no more than a
week for this release).

For example:

* Mark as non-problematic:
This would require an API, so that tools like Huggle or Cluebot can communicate with Echo.
We have pushed back API development to future releases, because it would take
significantly more time than we can afford right now -- not to mention the extensive
coordination with third-party developers.

* Pageviews/Visitors:
This would require us to create a special database that would make pageview data available
in a structured format, so that a program could periodically extract that data on a
per-article basis, to determine whether or not a given threshold was exceeded. And visitor
counts would be even harder to collect. Sadly, these ideas appear to be out-of-scope for
this release.

3. Some ideas may not work for all projects  

Here are examples of ideas that may not work on all projects, because they would require
features that may not exist on these projects:

* Mark as useful
While the 'mark as useful' idea would provide just the kind of positive feedback
we are looking for and could be developed in about a week, not all projects may be willing
to add this new feature in their article history or diff pages. 

* Feedback
While many positive notifications can easily be provided from the article feedback
extension, not all projects may be willing to add this feature on their sites (e.g. the
English Wikipedia is now leaning against it). 

* Page reviews
Note that our current page review notification ('your page was reviewed') requires
Page Curation to be installed, which leaves out most other projects right now. So for all
intents and purposes, this notification will only help English Wikipedia users at this
time. 

4. Some ideas may not match primary user behavior  

We are primarily looking for positive notifications about article edits made by a new
user, because that primary behavior represents about 74% of their first actions. While
notifications about pages they created are nice to have, this only corresponds to about
16% of current user behavior, which makes them less important.

So to sum up, we're looking for ideas that match these criteria:
* makes the new user feel good about participating
* responds to an article edit they made recently
* can be developed in no more than a week
* is likely to be adopted by a number of projects

Given these criteria, the most promising ideas so far appear to be notifications about
contributions (or contributors) to a page you edited. We might also consider 'mark as
useful' or 'feedback' notifications for projects that seem likely to adopt
these new features, but those would have to be viewed as lower priority than the
contribution notifications. Other ideas may have to wait until future releases, sadly.

What do you think? Have we missed any ideas that could fit within our criteria? Or do we
have information that might shed new light on some of the more challenging ideas we have
discussed so far?

Thanks again for your great suggestions so far -- we're trying to solve a difficult
but important problem, which will require time and ingenuity from all of us. We hope that
with your help, we can develop practical methods for showing appreciation and gratitude to
new users in their first steps.

All the best,

Fabrice

_____________________

POSITIVE NOTIFICATION IDEAS 

Here is a full list of positive notifications we have discussed so far. 

They fall in 7 main categories, based on what triggers them, and are listed below in rough
order of feasibility:

1. Contributions:

* New edits to a page you edited:
"'100 edits have been made to this page since your last edit.'"

* New edits to a page you started:
"'100 edits have been made to a page you started.'"

2. Contributors:

* Contributors to a page you edited:
'20 people have contributed to this page since your last edit'

* Contributors to a page you started:
"5 other people have contributed to the page you created."

3. Mark as useful:

* Support in history or diff pages: (manual clicks)
'Kaldari marked your edit as useful'

… or: (alternative wordings)
'Kaldari thanked you for your edit'
'Kaldari liked your edit'

4. Mark as non-problematic:

* Manual evaluations in tools like Huggle:
'Okeyes did not find a problem with your edit'

* Automated evaluations by bots like Cluebot:
'Cluebot did not find a problem with your edit'

5. Pageviews:

* Pageviews on a page you edited:
"A page you edited was viewed 1,000 times"

* Pageviews on a page you started:
"A page you started just reached its first 1,000 page views"

6. Visitors:

* Visitors to an article you edited:
"30 people have looked at your article since you made your edit."

* Visitors to a page you started:
"42 people have read the page you started.'" 

* Visitors to your user page:
"20 people have visited your user page this week"

7. Article feedback

* Your useful feedback
"HowieF found your comment useful"

* Positive feedback for a page you edited:
"10 positive comments were posted on a page you edited."

* Useful feedback for a page you edited:
"10 useful comments were posted on a page you edited."

* Feedback satisfaction for a page you edited:
"85% of readers found what they were looking for on a page you edited."

The most promising ideas will soon be added to our feature requirements page:

http://www.mediawiki.org/wiki/Echo/Feature_requirements#Features_under_cons…

On Feb 2, 2013, at 5:03 AM, Oliver Keyes wrote:

...

 On 2 February 2013 12:43, Benoît Evellin &lt;benoit.evellin(a)wikimedia.fr&gt; wrote:
 Hi everybody

 I have various observations for all of your ideas.

 * Useful edit notification : this idea may be a good one, if the wording illustrates a
Jedi/padawan relation instead of an editor-in-chief/freelance relation. We want equality
between all editors. We all know that is not true, so we mustn't dig the trench
deeper.

 * Contributions since the last edit : I completely agree with Chris experience and Liam
suggestions. Be careful again in the wording : articles are the property of no one. 

 * positive notifications and bot notifications : how will it work on Wikipedias without
theses features ?

 The idea is that instead of hooking it into specific services (ClueBot, Huggle) we'll
have a "silent" notification - something that exists but is not triggered by
MediaWiki, and can instead be triggered through the API. So, when ClueBot finds an edit
does not meet its standards for reverting it, it would poke the API to send $notification
to $userwhomadeedit. Because it's not service-specific, other wikis with their own
automated or semi-automated tools could also hook in using the same process.

 The problem with tying the notification to something in MediaWiki is that MediaWiki
itself really doesn't have a way of recognising edits as 'good' or
'bad' - that's always been handled through bots and user extensions.

 (I'd actually argue that this is a good illustration of why a decade of correcting
for MediaWiki's flaws by way of the TS, bots, API calls, etc has substantially harmed
the efficacy of our product(s) - but this is an essay-sized rant for another day :))

 Benoît

 2013/2/2 Sage Ross &lt;ragesoss+wikipedia(a)gmail.com&gt;
 On Sat, Feb 2, 2013 at 6:18 AM, Oliver Keyes &lt;okeyes(a)wikimedia.org&gt; wrote:

 On 2 February 2013 03:33, Liam Wyatt &lt;liamwyatt(a)gmail.com&gt; wrote:

 I'd like to give a giant +1 to Chris's suggestion - telling (potential)
 editors how many other people have read the article is a big motivator. It's
 logical really, we know this from the Education outreach projects and also
 from all the GLAM content donations: people REALLY are motivated by the fact
 that *their* writing and multimedia is being seen by lots of people.

 Currently that information is rather hidden away in a link to the
 toolserver via the History tab. If you could bring that information more to
 the fore it could be really satisfying. For example:
  "30 people have looked at your article since you made your edit." or,
 "350 people have seen this article in the last month" or even "6 other
 editors have changed this article and 500 people have read it since you last
 helped edit it". Perhaps you could even give some more complex breakdowns
 with pageviews by continent?
  The problem with this (or potential problem) is twofold: first, with a large
 number of pages it could get spammy. Second, to my knowledge the toolserver
 and stats.grok.se sites are not run off any kind of live data; they're
 reliant on database dumps. We'd either be plugging into third-party services
 of unknown viability or need to make a request to analytics for them to make
 this kind of data more internally available and transparent, which could be
 a pile of work.

 The traffic dumps have been running pretty reliably on a daily basis,
 so it's close enough to live for this purpose.

 Making that more internally available and transparent would be well
 worth a modest pile of work, as this is data that we know is very
 powerful motivation for many contributors (new and experienced alike).

 It would take some experimenting to see what kinds of traffic-related
 data are effective in Echo notifications, but the basic concept has a
 lot of potential. (And getting article-level traffic data integrated
 into our internal infrastructure would be an important step forward
 even beyond usage in Echo.)

 -Sage

 _______________________________________________ 

...
  On 2 February 2013 08:38, Chris McMahon
&lt;cmcmahon(a)wikimedia.org&gt; wrote:

 On Fri, Feb 1, 2013 at 1:02 PM, Fabrice Florin &lt;fflorin(a)wikimedia.org&gt; wrote:
 Hi guys,

 We would be grateful for your advice on how to give more positive notifications to new
users after their first edits.

 We're looking for notification ideas that could lead new editors towards a
"happy path" to encourage further contributions. 

 What about "20 people have visited your user page at User:Yourname this week"
or similar? I think it points up the interconnectedness of wiki users. 

 In general, I think notifications about people and what people do will be more welcome
than notifications about pages and what is done to them.  This is true of the suggestions
I've read, but I think it is worth noting specifically.    

 _______________________________________________ 

On Feb 2, 2013, at 3:17 AM, Oliver Keyes wrote:

...

 On 2 February 2013 00:20, Steven Walling &lt;swalling(a)wikimedia.org&gt; wrote:

 On Fri, Feb 1, 2013 at 2:26 PM, Oliver Keyes &lt;okeyes(a)wikimedia.org&gt; wrote:
 So, I feel strongly that Huggle and ClueBot integration is frankly all the positive
notification we need for edits alone. I can gather some data on it, but I'm pretty
sure they cover the vast majority of incoming edits. I'm also wary of sticking another
interface element in page histories or diffs (already crammed), since it increases the
footprint of Echo and the chances people might grumble about it.

 I think Oliver is probably correct that hooking in to these pre-existing queues is the
right approach. The one question I'd have is: what volume of edits are actually being
marked as helpful? 

 I can get some numbers on this - I'll poke the Huggle and Cluebot teams today. I
think it's less 'mark as helpful' and more 'mark as non-problematic'
at the moment, but they're for all intents the same (An edit that does not need
reverting). 
 I would say focusing on one method where we know users are getting positive feedback, and
then iterating on that, is a good approach. Also: am I correct in assuming that we have
the "marked as patrolled" notification for page creators? 

 I would say that adding the "mark as helpful" button is a larger  change than
you think, and likely to cause a stir. I would proceed with caution there, because in
addition to the noise you'll generate from putting _anything_ new on histories and
diffs, we need to build a positive feedback mechanism that is going to work in the long
term, and which we think experienced users will actually use in a large scale way. This is
not a trivial ux problem, as you can see from the tale of the MoodBar/Feedback Dashboard
story: we can always get newbies to generate a new queue of activity. But changing the
habits of experienced editors with lots to do is hard. 

 -- 
 Steven Walling
 https://wikimediafoundation.org/ 

On Feb 1, 2013, at 12:15 PM, Luke Welling WMF wrote:

...
  Article Feedback integration, ie "A page you
edited 'List of Ancient Jedi' was rated 4* for trustworthiness"

 "5 other people have contributed to the page you created 'The
Wombles'"

 I think the best positive vanity metric would be "42 million people have read the
page you created 'Dr Who'" but getting the data would be difficult.

 On Fri, Feb 1, 2013 at 3:02 PM, Fabrice Florin &lt;fflorin(a)wikimedia.org&gt; wrote:
 Hi guys,

 We would be grateful for your advice on how to give more positive notifications to new
users after their first edits.

 We're looking for notification ideas that could lead new editors towards a
"happy path" to encourage further contributions. Many studies have shown that
positive reinforcement plays an important role in increasing a user's productivity,
and we would like to provide at least one or two good solutions to support that goal in
the first release of Echo at the end of March.

 Here are some of the ideas which we have brainstormed to date, on our Echo feature
requirements page:

 http://www.mediawiki.org/wiki/Echo/Feature_requirements#Positive_notificati…

 They include:

 * Useful edit notification
 This feature would add a 'Mark as useful' button in article history and/or diff
pages, to invite experienced editors to mark edits they find useful, so they can send
positive feedback to new editors. This could be done through a simple text link next to
each edit (e.g. next to 'Undo'): 'Mark as useful' (or 'Useful' for
short). When an editor clicks on that link, this notification would be sent to the new
editor: 'Kaldari marked your edit as useful on Golden-crowned Sparrow. <View your
edit>'

 * Huggle/Cluebot Notifications
 This feature would hook into third-party anti-vandalism tools like Huggle or ClueBot.
Some of these tools could be adapted to provide an "approve of edit" button
(Huggle already has such a feature), which could send a notification to the user who made
the 'good' edit saying something like "this was good! Keep it up".

 * Contributions since your last edit
 Another feature we discussed is giving new users a bundled notification that '20
people have contributed to this page since your last edit'. While this type of
activity is currently handled by the watchlist, this type of 'contributions'
notification could have a positive impact in getting the new user to go back to a page
they edited earlier (particularly if they don't yet feel motivated to use the
watchlist). 

 * Positive notifications for active new users
 We have already identified a number of positive notification ideas for active new users,
which include: WikiLove, Page Link, or Page added to WikiProject. While these positive
notifications are likely to motivate active new users, a challenge we face is how to give
positive reinforcement to new editors immediately after their first few edits, before they
become active enough to start new pages or be noticed for a WikiLove message.

 What do you think of these first ideas? None of them may be perfect in their current
formulation, but with your help, we could be improve them to provide a practical solution
that helps engage new users to participate more productively. With everyone's
guidance, we can do better than only send them negative notifications when their edits are
reverted (which is like a slap in the face) -- or sending them no notifications whatsoever
after their first edits (which is what we are doing now). 

 Do you have other ideas for positive notifications we could be sending to new users? 

 You are welcome to respond in this email thread, or add your comments or suggestions on
this Echo talk page.

 Thanks for any tips or ideas you can provide to help us with this important editor
engagement goal!

 All the best,

 Fabrice

 _______________________________

 Fabrice Florin
 Product Manager, Editor Engagement
 Wikimedia Foundation

 http://en.wikipedia.org/wiki/User:Fabrice_Florin_(WMF)

_______________________________

Fabrice Florin
Product Manager, Editor Engagement
Wikimedia Foundation

http://en.wikipedia.org/wiki/User:Fabrice_Florin_(WMF)

Donate to keep Wikipedia free:
https://donate.wikimedia.org/