We tested several approaches to Direct Messaging before arriving at what shipped. One of the initial concepts we tested allowed users the ability to message other users that were anonymous. You would only be able to initiate the conversation from a post or a reply. We provided that initial piece of content as the context for the chat.

The second part of the problem led to us building Direct Messaging which required a handle to be functional. For now I’ll focus on the first solution: handles.

Our users were used to posting content anonymously and we didn’t want to take that away. While there was plenty of content that users would want to post and not be accountable for, there was also great content that they wanted associated with their persona. We were seeing signals from users that this was a feature that they would find useful. There were users that had already adopted a pseudo-identity that typically related to the content created. They were known in their communities for things like:

• Posting high quality memes
• Providing utility to the community at large with things like information on weather or campus events
• Just being interesting characters that others in the community wanted to keep up with. Yak Famous if you will

For the content these users wanted to take responsibility for, they would insert some sort of attribution into their post.

As we worked through the feature, I put together a ton of mocks which we tested with users. I was able to sit in on a number of these tests, and what we uncovered was that users were generally ok with the concept of handles so long as they were optional. They understood the value that a handle added but only for specific types of content.

Users shared that even if this feature were available, there would still be times they'd want to post anonymously. Some of the consistent examples given were:

• I don't want people to know how many times I post to the feed
• I'm going to say something that might upset people
• I want to post about something I don't usually post about. If it doesn't get any interactions, I don't want that to reflect poorly on my persona.

The underlying need in all of these examples was control. Users liked the feature but they wanted to use it at their discretion.

We had to be vague about how often users would be able to change their handle. In keeping with the vision of identity in the app, we wanted users to maintain their handle / persona for as long as possible. The ability to change your handle was only achievable through contacting customer support at this time.

Another reason we had to be vague about handle changes came from lessons we’d learned from an earlier feature called MyHerd. Users had to opt into MyHerd. The opt-in experience had language stressing that once a user selected their MyHerd, they would not be able to reset it in the future. This brought anxiety to the decision and had a negative impact on adoption of the feature.