Categories: Discussions

The vision for SUMO – Part 8: Live Chat

We’ve reached the last part of this comprehensive vision for the scope and role of the SUMO project moving forward. This part is dedicated to the most social form of text based user-to-user support: Live Chat.

If a problem isn’t yet covered in the Knowledge Base, or if the instructions in the article are too hard to understand, Live Chat is a powerful way for users to get touch with Firefox experts and get hands-on assistance in solving their problems.

Live Chat can also be a very fun way for contributors to provide support. Contributors helping out with Live Chat don’t just help users, they talk to each other in the backchannel as well, providing assistance to other helpers whenever needed. This means that although you’re usually the only one interacting with the user you’re helping, you’re never alone.

An important aspect — and one of the most interesting in my opinion — about Live Chat is that it is instrumental in detecting new and emerging support issues that are not yet covered in the Knowledge Base. The people helping out with Live Chat, along with helpers of the Support Forum, are those closest of all to what is going on — they really are the ones with their finger on the pulse, if you will.

However, without enough contributors, Live Chat has the risk of giving the user a negative experience, principally because of long waiting times or limited hours of operation. Without contributors, Live Chat can never be successful. We think the two keys to attracting contributors are:

  1. to make contributing as straightforward as possible;
  2. to minimize the load on the contributors by using their time as efficiently as possible – we do this with the support funnel. More specifically, we want to make sure that people search the Knowledge Base (and Support Forum) before turning to Live Chat, so helpers aren’t overwhelmed by the number of users requesting assistance.

I’ve already covered the support funnel in previous blog posts, so this time I’ll focus on how we can make contributing to Live Chat as simple and fun as possible. As always, we already have some ideas, but we really need your feedback as well.

Fully integrated help client

In order to help out with Live Chat today, you need to install an open source Java-based chat client called Spark. This means that unlike the Knowledge Base or the Support Forum, people need to learn how to install, configure, and use a separate application in order to get started with helping people with Live Chat. While the current solution certainly works, it is far from ideal for a number of reasons:

  • It’s an unnecessary technical barrier and yet another application installed on peoples’ computers.
  • It’s poorly integrated with the rest of SUMO, meaning, among other things, that people helping out with Live Chat may not be aware of the support funnel and the most commonly reported problems and solutions already documented in the Knowledge Base.
  • The software isn’t custom-made for our needs, meaning for example that it’s not straightforward for helpers to find the backchannel where other helpers hang out.
  • We rely on good documentation to ensure that helpers understand where they should look for assistance and relevant information.
  • We have a separate log in for Live Chat than the rest of SUMO.

Replacing Spark with something integrated into the SUMO website itself would give us a solution that is pre-configured for our particular needs, do away with the need for separate software, and seamlessly integrate information about weekly common issues and other things relevant for our helpers. The barrier of entry would be reduced significantly.

Simple scheduling solution

Today we have specific times of the day when we’re sure to be open — our “hours of operation” for Live Chat. This is helpful for contributors as they know when they can expect other people to be online. After all, just like many other things in life, Live Chat shouldn’t be a solitary activity.

The ideal solution would be if people could just log in and help out whenever they wanted to, but that would require lots and lots of helpers. We’re not quite there yet, and in the meantime we need a way to make the existing helpers gather around certain scheduled hours naturally. However, the hours of operation might not fit everyone, so we need something that will allow people to commit to specific time slots that are convenient for them, while still seeing which slots other people have offered to take or would like to take — think of it as a shared calendar.

This would give helpers the ability to plan ahead of time when they want to help out, based on what time slots other helpers have already committed to. We think this will encourage collaboration and make Live Chat even more fun. Furthermore, the official hours of operation could then be based on the confirmed schedule time slots, rather than predefined hours we’re using today. And from a user’s point of view, the relevant info is when Live Chat will open next, not a predictable weekly schedule.

Other improvements

Of course, integrating the help client and adding a capability to assist volunteers in being able to plan their contribution of time are just a couple of things we can do to make things simpler. Here are some other things we have in mind:

  • Support for other languages than English — Both users and helpers often speak more than one language. Ideally, users would simply select which languages they understand from the list of currently available languages. That list of available languages would itself be based on which helpers are currently logged on.
  • Add a lower limit of the number of helpers that needs to be logged on before Live Chat opens, to ensure a good support experience. This way, we can control the helper/user ratio and thus avoid helpers burnout. In other words, if there are lots of users in need of help, we’d need more helpers before we opened Live Chat, to ensure that the waiting times are not unreasonable.
  • Automatically save chat logs, along with user happiness rating and whether or not the problem was solved. This would allow us to make better use of the Live Chat stats about common Firefox problems. It would also allow us to tell which helpers are doing a great job, and which ones need a little more training.

Have you tested Live Chat already? If so, what did you like about it, and what would you want to change? We’d love to hear your thoughts, and whether or not you think our ideas are in line with what you’re seeing, or if you have other ideas we haven’t covered here.  And if you haven’t tried it yet, consider this a written invitation. :)

4 comments on “The vision for SUMO – Part 8: Live Chat”

  1. Michael Lefevre wrote on

    Well, first I should say that the reason I’m not more involved in support is because I haven’t got the spare time for it (and especially what spare time I might have doesn’t tend to be predictable enough to commit to specific times, although I completely understand why that is necessary, and having an easy system for doing it can only be good).

    In general, I think you have expressed pretty clearly in this post the things I thought could be improved. Despite my being familiar with a whole range of software, including several IRC and IM clients, it took me a while to work out how to use Spark, and I hadn’t quite got to grips with all the fastpath bits after a couple of hours. After a few weeks, I removed it from my system cause I like to keep things tidy and I wasn’t using it.

    The other issue I had while doing live support was with the SUMO knowledge base. I’m sure I would get better at finding things on SUMO with time, but actually I’m afraid I found answers more easily with just a Google search, and from that either managed to find a SUMO KB article by refining my search based on the content of the answer I found elsewhere, or by using Google restricted to the SUMO site. The site funnel search needs to be at least as good as the results of a Google search restricted to the site (and if it isn’t, then you could always use a Google search instead…).

    Finally (and I mean this to be an example, but it wouldn’t be bad to know about the specific case…) Someone came to live chat with a question, and I found some answers on the forum and an external site. As requested, I noted it in the contributors forum – I am afraid I never followed up on it (sorry about that!). But assuming I never do, shouldn’t there be some way of keeping track of that?

    I guess it’s down to better integration, but it should be easier to get from the point I was, finding an answer elsewhere on the web that should be in the KB, to getting it in the KB. I had to ask via chat whether I should be putting a stub in the wiki, filing a bug, or making a post in the forum (is that documented somewhere?). Then I posted in the forum and someone suggested I drafted something in the wiki and then I guess I would go back to the forum to ask for review?

  2. David Tenser wrote on

    Michael, good to hear from you again! Thanks for your comments. For people like you with less predictable spare time, helping out in the forum or the knowledge base is probably easier than live chat. Just five minutes per day in the forum can really be helpful for lots of users, as well as giving other contributors a better overview of what our most commonly reported issues are.

    Our search function is definitely one of the things we’re focusing on next. In fact, we are using Google today, but it’s not working very well for us. We’d like to better integrate the KB and forum search, making sure we can weight articles independently to make sure new and critical issue get promoted in the search results, and many other things. That said, if you can show examples of when a site specific Google search on SUMO is better than the search bar in SUMO itself, I’d be interested in seeing that to get a better understanding of what’s going on.

    About your contributor’s forum thread, I think you did exactly what you should do there, which was to raise the issue and check what the next step should be for something that is frequently (or seemingly frequently?) reported. We’ve just recently started to compile lists of which issues are indeed the most frequently reported ones, based on stats from the forum, Live Chat, Hendrix, etc (see For any issue that ends up in this list, there should be a Knowledge Base article. In other words, we need to know whether or not it’s actually a common issue before we write a Knowledge Base article about it. Back in June, we had just started to look at these lists so we relied more on anecdotal evidence (e.g. one contributor reading two forum threads about an issue and thinking it’s a frequent issue based on those two threads). We need to make this process more obvious to contributors, to reduce confusion.

    Keep the feedback coming, and great to have you trying out SUMO!

  3. Atul wrote on

    Great article! Live support is awesome.

    Regarding suggestions, one thing I was surprised to find out when talking to Cheng Wang at the Mozilla summit is that there’s no way to see how much time a volunteer has contributed to helping people. One big benefit of volunteering to contribute to a written article is that you have something tangible that you’ve contributed to the community, which others can see and thank you for; when it comes to live chat, though, the only way for me to find out that Cheng had dedicated countless hours of his time to LC was by asking him.

    So I think that having a “leaderboard” showing/ranking how contributory volunteers have been, along with perhaps public comments left for them by people they’ve helped or worked with (as you mentioned in your post) could both be useful as a way to provide transparency and accountability, while also providing a social incentive to encourage more people to join in.

  4. David Tenser wrote on

    Atul, thanks for sharing great ideas about how to encourage people to join SUMO. I like the idea of showing how much time someone has spent helping, and also the public comments/testimonials.