Meetings 2020 08 24 (Read Unread)

Read/Unread Resarch & MVP Discussion - August 24, 2020

Overview

Attendees: Allie, Mickael, Erik, Nina, John

Agenda:

• recap our understanding of what problems we are solving
• capture our takeaways from UX research so far
• seek consensus on scope for MVP & refinement of acceptance criteria

Anticipated follow-up:

• Update https://github.com/freedomofpress/securedrop-client/issues/187 with acceptance criteria
• Update technical document (draft): https://docs.google.com/document/d/1HwHIXQ5RaOCYyaDafxtgLfzLqPBEg-QkZIJvZ-EaOfI

Feature/problem mapping

Which problem does a global read/unread state solve?

(Nina) Does: Does somebody need to look at this Source to review new activity. Does Not: Who is working on this, have any of our Journalists yet sent a Reply, is there a new Reply, Who else has seen this, have documents been reviewed, who is assigned to this story, what external IJs are working on this, has the Source logged in since their last submission.

(Erik) For SecureDrops with a single user: "What's new to me, what have I looked at before" For SecureDrops with more than one user: "What has anyone on my team already looked at"

(Allie) I think the biggest argument for global bolding/unbolding is that folks are used to this in the JI and they are used to inferring that something is probably being worked on by another journalist if it's unbold. During research we learned they still have to check with teammates and track if their assumptions are true. Also in the JI it takes a long time to download individual messages/replies/files, whereas on the client all you have to do is scroll around on the source list, so this will be less of an indicator of what's being worked on and more of an indicator that someone else clicked on a source.

• (Erik) Note that in the web-based JI, simply starting the download of a source document will mark it as "read", even if you have not stepped through the process of viewing it

(mickael) Has someone else at my organization clicked on this source (and maybe viewed the contents of the messages/attachments)

Agreement:

• "Has anyone else in the organization has clicked on a SecureDrop source?"

Which problem does a per-user read/unread state solve?

(Erik) For SecureDrops with a single user: "What's new to me, what have I looked at before" For SecureDrops with more than one user: "What's new to me, what have I looked at before"

(Allie) The per-account read/unread state makes it easy for the logged-in user to know what is new to them.

(John) Prevents others' activity from causing me to miss new sources/submissions/replies. Enables the waterline feature.

Which problem does a "mark as unread" feature solve?

(Erik) "I've casually looked at this without really dealing with this, want to leave it for others" - only useful in combination with global state

(mickael) revert accidental or intentional marking as read (or marking as unread as reminder for self in future)

(Nina) Does: I want to review everything, but not mess-up what my colleagues see as new or still in-need of their attention or triaging. Does Not: I want to assign a journalist a case to own, I want to take my name off this Source as having viewed it since the last activity, I want to flag to a Journalist that I want them to respond to this, I want to send documents to a Journalist, I want my admin to export documents for me.

(Allie) Mark as unread is a feature that solves the problem of clicking on a source that you didn't intend to read.

(John) I'm not sure. The spreadsheet suggests that people expect and prefer read state be updated automatically, and there was at least some feedback that people use starring to mark sources they want to come back to.

Which problem does "new" waterline feature solve?

(Erik) "What's new to me?" (only useful in combination with global state)

(Nina) Does: What has happened (new messages, files, Sources, and Replies) since I last checked my SecureDrop? Does Not: What is a new Source contribution that nobody has seen yet?

(Allie) The new waterline solves the problem that global read/unread doesn't: it shows what material is new to the logged-in user so that they don't miss a new source or new content.

(mickael) What sources/submissions are new to me?

Which problem isn't solved by above ^? (Erik) A global read/unread isn't a robust "assigment" feature but more akin to a shared IMAP inbox with casual, informal ways to manage shared attention.

(Nina)

• The ability to attribute open "tags" to Sources (topics, journalists, field offices)
• The ability to leave notes for others in my newsroom, around a specific source.
• Assignment (ownership) of a Source to other journalists
• Importance (of a Source) relative to other Sources.
• Who (specifically) has "seen" a Source since the last activity
• Who (specifically) has reviewed messages from Sources
• Who (specifically) has opened/reviewed a file
• Who (specifically) has reviewed/seen replies from colleagues
• When each individual action within the Source, happened (when older messages were sent, read, reviewed).

(mickael) per-journalist is the straightforward "new to me" whereas global read/unread has further implications/meaning

(Allie) Global read/unread does not allow us to show bolding of sources when there are new replies because the sender of a reply has already seen it, which global read/unread will mark as seen by everyone. Account-level read/unread can show a source as bold if there are new replies from your teammates. This makes global read/unread more complicated because if we want to show when there are new replies from other teammembers that you haven't seen, we'll have to track this part of read/unread at the individual level.

(Allie) We still need to solve the problem of how we show unopened files, new replies from other teammates, and assignment.

Key research takeaways

Do users understand the SecureDrop inbox as a whole to be personal to them, or global to all users?

(Erik) It seems clear that users understand the inbox to be shared, based on the existing paradigm.

(mickael) They expect it to mirror journalist interface

(Allie) The users I interviewed seem to be familiar with the Journalist Interface and expect it to be similar. But the client is different in that it is very quick to click on a source and scroll down the list (all the messags and replies are loaded and marked as read quickly).

Agreement: To the the extent that they have experience using SecureDrop in a shared inbox setting, they expect the SecureDrop Workstation to mirror its behavior

Do users expect read/unread status to be personal to them, or global to all users?

(Erik) A bit ambiguous, but my interpretation is that they similarly expect it to be global based on the shared inbox mental model.

(Nina) Users need to coordinate with their broader newsrooms, first and foremost.

(mickael) I think this is hard to test in a prototype where there is no usage concurrency, and where we are "simulating" a single session - As it gets easier to use SecureDrop for journalists, there may be some concurrent/overlapping usage

(Allie) The users I interviewed seem to expect that we have few resources and to manage with whatever features we give them. If we tell them we can't implement a way to show what their teammates are looking at, then there's a bigger argument for global read/unread so that they can see what their teammates have clicked on and know what they should look at next.

If read/unread is global, do users expect an additional way to identify material "new to them"?

(Erik) Unsure, my interpretation is that an ephemeral notification when stuff comes in (loading animation etc.) or a very subtle indicator may be sufficient

(mickael) in the only interview i participated in, waterline appeared to confuse the user

(Allie) I believe this is the complication with global read/unread: there needs to be a way to show users what's new material for them, but mixing bolding and newlines in the same list is confusing. We can show what's new easily with account-level bold/unbold.

Agreement: more research required to better understand user needs in context of global read/unread

If read/unread is global, do users desire a method to review material without marking it as read for everyone?

(Erik) I think so, some team members want a way to quickly inspect things, and offline mode is not sufficient for that (you may have to download material, which is only possible in online mode)

(mickael) I think so -- In some organizations, some users have different roles and with the workstation making it much easier/quicker to view files, maybe they want to peek at the submission without committing to the full processing/triage process

(Allie) One of the users in our user interviews said that since the client is so much faster than the JI that he checks it a lot more and therefore sees himself as having to use the toggle to undo the global state change of read/unread so that his teammates can see what's new for them.

Agreement: Some users would likely appreciate a "Mark as unread" in the context of a global read/unread feature

If read/unread is global, do users desire a method to mark material as "dealt with" or "assigned"?

(Erik) It's unclear to me whether a formal method like this is actually desired.

(mickael) Global read unread seemed to me like it was a synomym for "dealt with"

(Allie) The users I interviewed are used to keeping track of who's working on what by texting each other and keeping records outside of SecureDrop. Again, users seem like they expect little change from the way things work now. No one said "I would like to stop having to text and track "dealt with" outside of sdc. But I bet it'd be used in sdc if we added this, and this would improve security.

If read/unread is global, do users desire additional "seen by" information?

(Erik) Research seems limited in answering this since the prototype only had a very limited build-out here. My interpretation is that it could have significant value because teams know each other's strengths and weaknesses, and knowing "so-and-so has looked at it" can convey a lot of information.

(Allie) During user interviews one user seemed excited by this as a new feature. This doesn't exist in the JI so it tells you more about who looked at something.

If read/unread is per-journalist, do users expect an additional way to identify material "viewed/treated/triaged by another journalist"?

(Erik) I think so, given that they've spoken to how a global read/unread could help them to answer the same question.

MVP acceptance criteria discussion & next steps

Agreement:

• Initially we'll implement a global state
• Interpret is_downloaded Boolean from the web-based JI in Client (TBD: Should "read/unread by the Client" be reflected in the JI?)
• "Mark as unread" out of scope for now, to be considered along other user needs identified via MVP response
• Small work session on refining acceptance criteria consistent with the above (Mickael/Allie/Erik)