How do we know if Discourse is right for us?

In order to properly try Discourse, we should establish:

  • What we want to get out of Discourse;
  • What success would look like;
  • What an exit strategy would look like in case of a non-successful trial run.

For the first, I would like to get a better platform for talking to our users and identifying bugs and hitches in installing and usage that don’t fit particularly well in GitHub issues. As well, I would like to lower barriers of entry; people may not post in ipfs/go-ipfs as it is a scary, mature project - here, it may be easier to ask silly questions that they may have.

Success is not clear. I would like to specifically see if we get more users and more questions than we do on ipfs/ipfs, ipfs/support, and ipfs/faq combined. This requires tracking how many people interact here; does anyone have suggestions on how to do that?

If we don’t get more engagement with our userbase, I would consider Discourse a failed experiment.

Finally, I am hesitant to remove GitHub repositories or links to ipfs/faq and ipfs/support in favor of DIscourse, because it will be a hassle to add those back in in case we go back. I think it should be enough to leave those, add a note encouraging that users come here. And, if this fails, I think copying over issues as questions in the FAQ would be the best way forward.

What do you think?

I’m gonna throw-in my two cents:

Discourse is not going to work if we don’t fully push for it. That means:

  • Moving this to a proper domain under ipfs.io
  • Killing about every repo which people are using to ask questions/get support/discuss
  • Making sure people dont ask questions in Github, closing any issues asking for support, anything non-technical, not related to the repo, and redirecting everything here. Repos should be about code as a general thumb rule.
  • FAQs should have their own static website (because last thing a user wants is to scroll through 20 posts when trying to find a simple answer to a common question)
  • Drop the exit strategy and make it the support/discussion platform: start thinking about how to make it work and not how to undo things if it does not work
  • Focus efforts in streamlining the user support. A growing user base + bad support = headaches and bad image, something we need to avoid in the future. Going mainstream will mean a large influx of non-technical users for which we need to be prepared. Github is not the best environment for them.

Discourse is a great piece of technology, it is well proven and does a very good job. Search works great. I don’t think though that we can expect a community of users and developers to keep track of Discourse+a large number repos with similar purpose, specially with links everywhere pointing to different places, and people posting content and questions where they should not.

If we take a strong policy/organizational stand to make discourse work, the forum will thrive and will do its job as community catalyzer. If we keep “testing” we will not reach that point, we will only add confusion (specially to newcomers) and the test will fail. In that case, Discourse will add little value to what we already have.

4 Likes

Discussed in todays all-hands, how do we evaluate if Discourse actually works for us?

And what is the timeframe for figuring this out?

It’s a little hard to get a metric for its success, since (from what I can tell) its audience is supposed to be newcomers and non-developers, but at the same time you don’t want newcomers coming here while it’s being trialled, thinking it’s the “official forum” or anything.

Developers won’t use it because they have Github and IRC – so if the forum’s still a ghost town in two, three months’ time, will that be because it’s not the right fit for the community, or because the community never saw it in the first place?

1 Like

Discourse supports adding hooks for google analytics. Should we set those up (sooner rather than later) so we at least have that metric? People might read stuff without posting (or even creating a user account)

There’s a very similar topic:

http://ipfs.trydiscourse.com/t/measuring-success-of-discourse-trial/34

I have moved the messages in the other thread under the same topic (mostly to test how this worked in Discourse). Me likes.

@hector What was the title of the other conversation? What was the title of this one? I wonder if @richard was trying to encourage two distinct conversations. If we combine them into one thread, I want to make sure we address both topics.

“How do we know if Discourse is right for us and how long time will it take to figure out?”
VS
“Measuring Success of Discourse Trial”

Think it’s about the same.

This was the previous topic, slightly surprised this info isn’t shown anywhere in the resulting conversation’s GUI.
http://ipfs.trydiscourse.com/t/how-do-we-know-if-discourse-is-right-for-us-and-how-long-time-will-it-take-to-figure-out/24

@hector has listed things we should do if/when we have decided to use discourse. I agree that those are all things we should do once we have made a decision. The functional question is/was (before we consolidated 2 threads):

What do we need to to in order to make this decision?

Breaking that into sub-questions:

How long should this trial last?
We don’t want to spend a long time being non-committal

How will we know whether this is the right tool for us?

  • What needs is it addressing?
  • Is it addressing those needs well? How will we measure that?
  • Does it address those needs better than the other tools we’ve considered (like github)?

How will we make the decision to pull the trigger?
When the trial period is over, how will we make the decision? Once we make that decision we will either shut down discourse or do all the things you’ve listed. How will we decide which action to take?

Sorry this was me not doing things correctly. There is no “merge” as such, only “move posts to” and while it is mentioned in the other thread where some posts went to, I should have mentioned it here that I had moved them from there.

I agree with pretty much everything @hector says about discourse not working if we don’t go all in. I also agree that we should kill ipfs/support and ipfs/faq if we are going to trial this; having both will not work, at all. Moving issues here, or just closing stale issues, seems to me to be the smartest thing to do.

Making FAQs into their own static website makes a lot of sense to me, too, and I think it is high time we actually do that. However, that is going to take a lot of work. I won’t be able to do it alone.

Matt’s questions make sense, too.

How long should this trial last? Let’s give it until the end of Q1. So, let’s re-evaluate in four months. Does that work for everyone?

How will we know if this works: if we find that we are able to help users answer their questions easily, if threads here don’t go stale, if we do better than we are currently doing with ipfs/support (not well), than I will consider this a success. It might not be worth it to set up explicit metrics, but to simply gauge how the community is after a couple of months. We can do this in a group thread. Right now, I think there is a good deal of confusion around ipfs/support and ipfs/ipfs. I hope that this can help with that - as in, I hope to see less “I opened an issue and no one responded.”

I think we can make the decision to shut down Discourse based on a group call. I’d be happy to have a majority vote to keep it going, or a majority vote in favor of some modification (which I can’t predict) to make it work better. Do you think that that would work, Matt?

Matt (@flyingzumwalt) is on vacation right now. If he wasn’t on vacation he would observe that we’re talking about two different phases of trial, which is causing confusion. He would then offer this clarification:

Phase 1: Kick the tires on ipfs.trydiscourse.com

We are in Phase 1 right now. The primary purpose of Phase 1 is to decide whether to proceed to Phase 2. It’s easy to back out of Phase 1 if we decide this was a bad idea, but we want to keep Phase 1 as short as possible because it puts us in a strange in-between state.

What happens in Phase 1:

  1. Run the forums at ipfs.trydiscourse.com so it’s clear we’re just kicking tires
  2. Create user accounts, create some threads, try to host a few discussions
  3. Leave existing repos for faq, support, community, etc intact and continue maintaining them
  4. Decide whether to proceed to Phase 2

Some people are really eager to get out of Phase 1. Other people have expressed hesitation. The primary thing to do in phase one is to address everyone’s concerns before proceeding.

Phase 2: Make a Real Go of it on community.ipfs.io

Phase 2 of the trial is what you’re all discussing, where we would do things like @hector lists

  • Moving this to a proper domain under ipfs.io
  • Killing about every repo which people are using to ask questions/get support/discuss
  • Making sure people dont ask questions in Github, closing any issues asking for support, etc
  • etc…

As @hector and @richard have both said, in order for Phase 2 to work, we will need to be “all in”. Because of that, we need everyone to agree that we should do it. Also, backing out of Phase 2 would be a lot harder, so we should pause to make sure we want to proceed before we do it.

How do we decide whether to Proceed to Phase 2?

Given this clearer language, the question we need to answer is How do we decide whether to Proceed to Phase 2? If you feel we have already made this decision, please say what were the factors that drove this decision.

Some things that definitely need to happen before we proceed to Phase 2:

  1. Make sure @jbenet has weighed in. He has expressed misgivings in the past.
  2. Make sure all of the Project Leads have agreed to be “all in” with this, treating threads on discourse with as much attention as they would give to an issue in ipfs/support (or more!)
  3. Write down the things that drove this decision, and get a record that everyone agreed – this might mainly be links to existing discussions in Github or discourse. Put that info into https://github.com/ipfs/community/issues/190

Is there anything else that needs to happen before we say it’s time to start Phase 2?

(Since @flyingzumwalt is on vacation, he’s pre-emptively giving a :+1: for proceeding to Phase 2)

2 Likes

I agree with @flyingzumwalt about moving to Phase 2.

I am all in, and am ready to move ipfs/support and ipfs/faq here.

Can we have feedback from the Project Leads about using discourse, and checking it daily? This means: @dignifiedquire, @Vic, @hector, @flyingzumwalt, who I believe are all in. It also means: David Dias, Haad, why, and jbenet. Anyone else I am missing? Let’s direct them here.

Apologies for the pre-emptive posting - I am not used to Discourse’s autocomplete, which is slow and doesn’t support tab enter to select, but instead posts. So far, this is the most annoying thing about Discourse.

1 Like

Sorry for the delayed response. I am all for moving to Phase 2 as soon as we have made a good plan of how to migrate the content we want to migrate over.

Migration Plan

  • Copy all open ipfs/faq issues over here.
  • Build a document that coalesces the ‘answered’ FAQ questions into a single document. Post it on the website.
  • Mention Discourse on ipfs/ipfs, ipfs/community, ipfs/pm.
  • Close some of the open long-standing issues on ipfs/community which are served by this repo.
  • Close ipfs/support, close all issues, and say that it is now Deprecated. Actively gut the README.
  • Post a link to Discourse on the IRC channel.

I think that covers it?

1 Like

I would like to have this as an FAQ post in Discourse as well. [quote=“richard, post:17, topic:34”]
Copy all open ipfs/faq issues over here
[/quote]

What about closed ones?

What about open issues there, should we migrate them actively here?

Discourse should be included in the topic, I guess that’s what you meant with this.

What about other announcement channels, like the blog and twitter?

I was thinking closed or answered questions in the FAQ would stay there as reference; we would only move over the unanswered, open issues.

Regarded open issues on support, I would either close them there (with a week or two response suggested), or move them here.

We should also announce it on the blog, Twitter, and in the sprints.

1 Like

[quote=“richard, post:19, topic:34”]
I was thinking closed or answered questions in the FAQ would stay there as reference; we would only move over the unanswered, open issues.[/quote]

General rule of FAQs is “check if your question was already addressed before”.
If you don’t move everything (at least in some form), Discourse /search will provide only partial results, decreasing overall discoverability,

This is a good idea. It will mitigate the “discoverability-by-search problem” I described.