Volunteer Response Team/Reports/2013
- 2013
Introduction
editThis is a statistical report of general activity relating to the operation of Wikimedia's VRTS, that This is a statistical report of general activity of emails sent to Wikimedia's VRTS, that was using at the time Open-source Ticket Request System to process emails. VRTS is used by the Volunteer Response Team to process emails from the public.
Emails sent to VRTS are assigned to a "queue" based upon the nature, the language and/or the Wikimedia project it might represent. This report reflects the general activity of the Volunteer Response Team. The statistics listed below are raw reports on account activity, administrative actions and ticket volume information for all language information queues and their subqueues. It also includes the same for sister projects, and permissions/photosubmission tickets.
After a request sat pending for about 3.5 years, our ticket system was finally upgraded to OTRS 3.2 this past August. With thanks to Martin Edenhofer, the creator of OTRS, who volunteered his time to assist Jeff Green and the Wikimedia Engineering team in the upgrade, our agents are able to more efficiently communicate with our customers. While we are still in the transition process and working out some bugs, we now have a more stable and secure environment to work in.
Account activity
edit- Accounts created/re-opened: 161
- Of these, 9 were for WMF employees, 33 were chapter representatives and 7 were solely for role-specific access, such as oversight, stewards, etc. The remaining accounts were for general information (and similar) queues. Of the new users, often more than one role is assigned upon account creation. As an example, someone from Commons might receive access to the sister projects, photosubmission and permissions queues.
- Accounts closed: 212
Closed accounts # Resigned 16 Inactive 133 Chapter related 7 No roles left * 34 Other ** 22
- * When a user no longer has access to any queues, their account is closed with this note.
** Other: 13 WMF staff accounts, 5 miscellaneous, 1 removed, 1 test account
Administrator activity
editNote: While a VRTS administrator may have zero/few logged actions, the majority of things they are tasked with are not logged. Examples include updating formatted replies, handling system maintenance tasks such as spam blacklist, message sorting, monitoring queues and agents for response time and quality, performing inactivity cleanups and creating statistics, orientating/coaching new (and existing) respondents, handling escalated tickets and issues with both agents and 'customers', recruiting agents and evaluate/process new applications. And of course, the VRTS administrators assist in answering tickets, just like all VRT members.
Current | Former | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Action | Cbrown1023 | Daniel | Jredmond | Keegan | Mailer diablo | Pajz | Raymond | Rjd0060 | Tiptoety | Abisys | JeLuF | Lyzzy | Sebmol | Total |
Accounts created | 29 | 13 | 0 | 16 | 0 | 8 | 38 | 47 | 7 | 0 | 0 | 2 | 0 | 160 |
Accounts closed | 59 | 23 | 0 | 8 | 31 | 5 | 21 | 90 | 2 | 0 | 0 | 0 | 0 | 239 |
Roles changed | 71 | 36 | 0 | 44 | 15 | 11 | 58 | 121 | 10 | 0 | 0 | 2 | 0 | 368 |
Addresses created | 3 | 0 | 0 | 4 | 0 | 0 | 8 | 3 | 0 | 0 | 0 | 0 | 0 | 18 |
Addresses disabled | 3 | 1 | 0 | 2 | 0 | 0 | 6 | 5 | 0 | 0 | 0 | 0 | 0 | 17 |
Queues closed | 9 | 2 | 0 | 1 | 1 | 1 | 2 | 4 | 0 | 0 | 0 | 0 | 0 | 20 |
Queues created/reopened | 3 | 0 | 0 | 2 | 0 | 0 | 5 | 4 | 0 | 0 | 0 | 0 | 0 | 14 |
Spam filter changes | 1 | 4 | 0 | 0 | 0 | 31 | 1 | 7 | 0 | 0 | 0 | 0 | 0 | 44 |
Total | 178 | 79 | 0 | 77 | 47 | 56 | 139 | 281 | 19 | 0 | 0 | 4 | 0 | — |
Queues
editNote: The OTRS software formats subqueues after two colons. For example, info-en::Vandalism is the Vandalism subqueue of the info-en queue.
Opened | Closed |
---|---|
|
|
Traffic
edit- Notes
- Totals include all tickets closed between 2013-01-01 00:00:00 and 2013-12-31 23:59:59 UTC.
- In OTRS, the "root" or "parent" queue can also contain tickets. As every queue has its own address, sometimes messages are not filtered ahead of time to a subqueue and end up in this root queue. They can be moved to a more appropriate subqueue if it is exists, but sometimes they are just responded to from within that parent queue.
- If you come here from the 2012 report, you might notice that we have dropped the distinction between different close states (i.e. +, - and nrn). This is because the majority of agents do not make use of these in a standardized way, so providing these figures would not be indicative of anything.
info queues
editThese queues are the general information addresses that are displayed on the Contact us pages on the Wikipedia projects. (See the English Wikipedia's page, for example.) They usually deal with questions from readers of Wikipedia and from the subjects of articles. They are typically divided based on the language that they are written in, not the project that they relate to, because the queues are staffed by Wikimedia volunteers speaking a certain language, not from a certain project.
Tickets closed | Agents with access * | |||||||
---|---|---|---|---|---|---|---|---|
Queue | 2012 | 2013 | Change (%) | End 2012 | End 2013 | Change (%) | Per agent ** | Notes |
info | 5 | 16 | 220 | 108 | 96 | -12.1 | 0.2 | |
info-af | 18 | 4 | -77.8 | 1 | 1 | 0 | 4 | |
info-als | 1 | 0 | -100 | 0 | 0 | 0 | — | |
info-ar | 164 | 474 | 189 | 7 | 11 | 57.1 | 52.7 | |
info-ca | 90 | 56 | -37.8 | 3 | 3 | 0 | 18.7 | |
info-cs | 419 | 419 | 0 | 6 | 7 | 16.7 | 64.5 | |
info-da | 115 | 128 | 11.3 | 6 | 5 | -16.7 | 23.3 | |
info-de | 7647 | 7128 | -6.8 | 47 | 44 | -6.4 | 156.7 | |
info-el | 125 | 203 | 62.4 | 2 | 4 | 100 | 67.7 | |
info-en | 20112 | 29590 | 47.1 | 206 | 210 | 1.9 | 142.3 | |
info-es | 2399 | 2062 | -14 | 22 | 25 | 13.6 | 87.7 | |
info-et | 22 | 4 | -81.8 | 1 | 0 | -100 | 8 | |
info-fa | 32 | 400 | 1150 | 1 | 7 | 600 | 100 | |
info-fi | 31 | 44 | 41.9 | 3 | 4 | 33.3 | 12.6 | |
info-fr | 1423 | 1164 | -18.2 | 27 | 28 | 3.7 | 42.3 | |
info-he | 198 | 195 | -1.5 | 10 | 7 | -30 | 22.9 | |
info-hr | 4 | 3 | -25 | 4 | 2 | -50 | 1 | |
info-hu | 183 | 256 | 39.9 | 14 | 15 | 7.1 | 17.7 | |
info-it | 520 | 420 | -19.2 | 10 | 8 | -20 | 46.7 | |
info-ja | 384 | 242 | -37 | 6 | 9 | 50 | 32.3 | |
info-ka | 0 | 0 | 0 | 1 | 0 | -100 | 0 | Created 08/2012 |
info-ko | 82 | 37 | -54.9 | 2 | 2 | 0 | 18.5 | |
info-nds | 17 | 3 | -82.4 | 0 | 0 | 0 | Inf | |
info-nl | 2647 | 2583 | -2.4 | 17 | 19 | 11.8 | 143.5 | |
info-no | 278 | 212 | -23.7 | 5 | 3 | -40 | 53 | |
info-pl | 763 | 1285 | 68.4 | 12 | 9 | -25 | 122.4 | |
info-pt | 169 | 42 | -75.1 | 9 | 9 | 0 | 4.7 | |
info-ro | 165 | 156 | -5.5 | 3 | 3 | 0 | 52 | |
info-ru | 955 | 759 | -20.5 | 15 | 17 | 13.3 | 47.4 | |
info-scn | 17 | 0 | -100 | 0 | 0 | 0 | — | |
info-sk | 2 | 0 | -100 | 1 | 1 | 0 | 0 | |
info-sl | 0 | 3 | Inf | 0 | 4 | Inf | 1.5 | Created 11/2013 |
info-sr | 49 | 50 | 2 | 2 | 1 | -50 | 33.3 | |
info-sv | 332 | 252 | -24.1 | 3 | 2 | -33.3 | 100.8 | |
info-tr | 159 | 196 | 23.3 | 5 | 3 | -40 | 49 | |
info-uk | 0 | 6 | Inf | 0 | 2 | Inf | 6 | Created 12/2012 |
info-vi | 24 | 11 | -54.2 | 1 | 1 | 0 | 11 | |
info-zh | 178 | 118 | -33.7 | 13 | 12 | -7.7 | 9.4 | |
Total | 39729 | 48514 | 22.1 |
* This number is slightly lower than the actual as some agents have support access that covers multiple queues and these are not reflected in the table. These users do not usually answer tickets regularly though, for example VRTS administrators who have access to all info queues for administrative purposes.
** The figure is based on the (arithmetic) mean of the number of agents on Dec. 31, 2012, ("End 2012") and the number of agents on Dec. 31, 2013 ("End 2013").
- Breakdowns
Note that info queues other than those listed below do not make use of subqueues (hence no breakdowns exist).
info-en
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
info-en | 7317 | 14952 | 104.3 | |
info-en::Copyvio | 521 | 562 | 7.9 | |
info-en::Courtesy | 6972 | 7948 | 14 | |
info-en::Mobile | 195 | 0 | -100 | Closed 04/2013 |
info-en::Organizations | 9 | 713 | 7822.2 | Created 11/2012 |
info-en::Quality | 2316 | 1975 | -14.7 | |
info-en::Reuse | 400 | 461 | 15.2 | |
info-en::Schools | 56 | 67 | 19.6 | |
info-en::Tech-issues | 273 | 610 | 123.4 | |
info-en::Urgent | 7 | 41 | 485.7 | |
info-en::Vandalism | 1619 | 2196 | 35.6 | |
info-en::Voice messages | 376 | 41 | -89.1 | |
info-en::info-en (Mohammed) | 51 | 24 | -52.9 | |
Total | 20112 | 29590 | 47.1 |
info-fr
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
info-fr | 1423 | 1163 | -18.3 | |
info-fr::Posters | 0 | 1 | Inf | |
Total | 1423 | 1164 | -18.2 |
info-nl
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
info-nl | 1334 | 1325 | -0.7 | |
info-nl::wikiportret | 1313 | 1258 | -4.2 | |
Total | 2647 | 2583 | -2.4 |
info-pl
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
info-pl | 680 | 1240 | 82.4 | |
info-pl::baninfo | 83 | 45 | -45.8 | |
Total | 763 | 1285 | 68.4 |
info-zh
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
info-zh | 99 | 66 | -33.3 | |
info-zh::info-zh-hans | 34 | 23 | -32.4 | |
info-zh::info-zh-hant | 45 | 29 | -35.6 | |
Total | 178 | 118 | -33.7 |
Sister projects
editThough the general information queues may serve all projects, the majority of the queries received tend to be related to Wikipedia or the Wikimedia projects as a whole. As the numbers below show, there are not many tickets that relate to non-Wikipedia projects, but the Sister projects queues were created to give them special attention so that they could be handled by people who know more about the specific project.
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
Sister projects | 1 | 0 | -100 | |
Sister projects::Commons | 378 | 234 | -38.1 | |
Sister projects::Wikibooks | 80 | 92 | 15 | |
Sister projects::Wikidata | 0 | 9 | Inf | Created 12/2012 |
Sister projects::Wikinews | 63 | 14 | -77.8 | |
Sister projects::Wikiquote | 91 | 59 | -35.2 | |
Sister projects::Wikisource | 7 | 4 | -42.9 | |
Sister projects::Wikiversity | 12 | 7 | -41.7 | Created 04/2012 |
Sister projects::Wikivoyage | 1 | 7 | 600 | Created 12/2012 |
Sister projects::Wiktionary | 98 | 82 | -16.3 | |
Sister projects::Wikispecies | — | 0 | — | Created 02/2013 |
Total | 731 | 508 | -30.5 |
Unlike in info queues, there is just one access right for all "Sister projects" queues. Between 2012 and 2013, the number of agents with access rose from 73 to 95 (+ 30.1 %). *
* These numbers are slightly lower than the actual as some agents have support access that covers multiple queues and these are not reflected. These users do not usually answer tickets regularly though, for example VRTS administrators who have access to all info queues for administrative purposes.
permissions
editSometimes Wikimedians would like to use images in articles that do not have clear cut licensing information. These images might not be licensed already, might have a license that is not permissive enough for use on Wikipedia, or might otherwise need confirmation that the copyright holder has agreed to the license shown on the projects. Wikimedians often contact the copyright holders of various images and ask them to license them so that they can be used on Wikimedia projects. The permissions team on VRTS is the one that processes these "permissions" e-mails and attempts to verify that the copyright holder releases the images under an allowable license, that they do so clearly while understanding what that means, and that they actually are allowed to license the images in such a way (e.g. they hold copyright). These e-mails are then stored in the ticket system so that they can be verified by anyone with appropriate access. For more information, see commons:Commons:Permission.
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
permissions | 60 | 194 | 223.3 | |
permissions::permissions-commons | 3572 | 4227 | 18.3 | |
permissions::permissions-commons::permissions-commons-de | 741 | 717 | -3.2 | |
permissions::permissions-commons::permissions-commons-es | 123 | 198 | 61 | |
permissions::permissions-commons::permissions-commons-fr | 182 | 338 | 85.7 | |
permissions::permissions-commons::permissions-commons-he | 933 | 1159 | 24.2 | |
permissions::permissions-commons::permissions-commons-pl | 140 | 134 | -4.3 | |
permissions::permissions-commons::permissions-commons-pt | 93 | 107 | 15.1 | |
permissions::permissions-de | 2096 | 2047 | -2.3 | |
permissions::permissions-el | 69 | 65 | -5.8 | |
permissions::permissions-en | 2549 | 2897 | 13.7 | |
permissions::permissions-eo | 2 | 0 | -100 | |
permissions::permissions-es | 144 | 135 | -6.2 | |
permissions::permissions-et | 33 | 58 | 75.8 | |
permissions::permissions-fi | 14 | 20 | 42.9 | |
permissions::permissions-fr | 646 | 506 | -21.7 | |
permissions::permissions-hu | 295 | 309 | 4.7 | |
permissions::permissions-it | 258 | 309 | 19.8 | |
permissions::permissions-ja | 0 | 16 | Inf | |
permissions::permissions-ml | 4 | 21 | 425 | |
permissions::permissions-nl | 223 | 174 | -22 | |
permissions::permissions-pl | 178 | 180 | 1.1 | |
permissions::permissions-pt | 134 | 97 | -27.6 | |
permissions::permissions-pt::permissions-pt-nf | 4 | 4 | 0 | |
permissions::permissions-ru | 729 | 589 | -19.2 | |
permissions::permissions-sk | 40 | 52 | 30 | |
permissions::permissions-sk::erby-sk | 0 | 0 | 0 | |
permissions::permissions-sl | 9 | 44 | 388.9 | |
permissions::permissions-sr | 46 | 43 | -6.5 | |
permissions::permissions-tr | 26 | 72 | 176.9 | |
permissions::permissions-uk | 30 | 134 | 346.7 | Created 10/2012 |
permissions::permissions-vi | 2 | 3 | 50 | |
permissions::permissions-zh | 45 | 13 | -71.1 | |
permissions::permissions-zh::permissions-zh-hans | 79 | 81 | 2.5 | |
permissions::permissions-zh::permissions-zh-hant | 64 | 35 | -45.3 | |
Total | 13563 | 14978 | 10.4 |
Unlike in info queues, there is just one access right for all "permissions" queues. Between 2012 and 2013, the number of agents with access rose from 350 to 378 (+ 8 %). *
* These numbers are slightly lower than the actual as some agents have support access that covers multiple queues and these are not reflected. These users do not usually answer tickets regularly though, for example VRTS administrators who have access to all info queues for administrative purposes.
photosubmission
editWikimedia projects often lack good, freely-licensed images for certain articles, especially biographies of living persons. To mitigate this, the photosubmission queues were created on VRTS. They allow readers of Wikipedia and article subjects (or their agents) to easily submit photos to illustrate their articles. Similar to with the permissions queues, the photosubmission team verifies that the person submitting the article is capable of freely licensing it, that they clearly state their intent to license it under a specific free license, and that they understand what they are doing. The team then helps to add the images to the relevant articles.
Queue | 2012 | 2013 | Change (%) | Notes |
---|---|---|---|---|
photosubmission | 686 | 810 | 18.1 | |
photosubmission::photosubmission-ar | 0 | 0 | 0 | |
photosubmission::photosubmission-da | 0 | 8 | Inf | |
photosubmission::photosubmission-de | 31 | 26 | -16.1 | |
photosubmission::photosubmission-es | 17 | 29 | 70.6 | |
photosubmission::photosubmission-fr | 0 | 0 | 0 | |
Total | 734 | 873 | 18.9 |
Unlike in info queues, there is just one access right for all Photosubmissions queues. Between 2012 and 2013, the number of agents with access increased from 105 to 108. *
* These numbers are slightly lower than the actual as some agents have support access that covers multiple queues and these are not reflected. These users do not usually answer tickets regularly though, for example VRTS administrators who have access to all info queues for administrative purposes.
Spam
editOf the roughly 1,600 tickets created everyday, about 1,400 are spam and about 180 legitimate tickets end up in an info, sister projects, permissions or photosubmissions queue. This leaves approximately 50 legitimate tickets in other queues.* Prior to the upgrade, weak spam filtering led to almost half of all spam emails getting through into the queues everyday on a regular basis. To alleviate the workload associated with sorting out spam, the process of handling such emails was modified in the course of the upgrading process. The configuration of the spam filter (SpamAssassin) was changed and learning capabilities were added, and the filter now sends potential spam either directly to the "Junk" queue or to a new "Probably Spam" queue which is manually reviewed. In addition to that, we have established a semi-automated process for closely monitoring false negatives in November 2013, and we use address- or content-specific rules to catch emails we identify as not properly handled by SpamAssassin. That way, by the end of 2013, we could drastically reduce the number of spam emails making it to our ordinary queues to below 100 per day while another 110–120 emails are sorted into "Probably Spam" and assessed by volunteers.**
* Rough estimate. The number of spam emails is based on a two-week sample during the composition of this report (spam emails are deleted after 10 days and we do not keep track of the amounts). The other figures are averages based on all traffic in 2013. Since a small amount of queues (e.g. for oversighters) is not searchable by VRTS administrators, the number of legitimate tickets in other queues was increased by 10.
** Values for "Probably Spam" reflect the average from January 16, 2014, through March 8, 2014, as data for 2013 is not available.
Outlook
editThe major event of 2013 was the upgrade to OTRS 3.2, and the VRTS administration plans on making regular updates to the software a high priority in their requests to the Wikimedia Foundation. We also will continue to seek technical support for bugs and patches. Other support opportunities that we hope to undertake in 2014 include comprehensive agent onboarding, training, and continuing education. We also will work toward providing additional and improved documentation and support for all languages on the agents' internal wiki.