Why would Discourse's business model work when Stack Exchange v1's failed?


(Helder Ribeiro) #1

Stack Exchange v1’s business model was “build your own community, we’ll give you the software on a subscription basis”.

That model failed because communities require critical mass, and achieving that is really hard. All but one community created at that time (Math Overflow) got any significant traction, if I recall correctly. SE eventually decided to shut down the SaaS model and to launch instead their own communities via Area 51.

Discourse compares itself to the wordress.org model because it’s open source and anyone can host their own, but aside from the self-hosting part, the model resembles SEv1 a lot more than WP. There are a lot more solo bloggers out there than forum organizers, and they can start blogging (and paying) from day 1, without needing to first build up a community with critical mass to make it something they can use.

So what’s different this time around?

Btw, I’m assuming that charging end-users for hosting their forums is a major part of the revenue stream, because if the plan is just selling to big enterprises I don’t even see why charge end-users in the first place.


Building a plugin to add Stack Exchange features
#2

Are they actually ever charging users? I haven’t seen anything along those lines so far (though maybe I missed it)


(Chris Hanel) #3

Discourse is much more useful to smaller communities as a forum than SEv1 would be as a question aggregation tool. It requires much less traffic to sustain itself, so I don’t think it’s a true equivalent.


(Helder Ribeiro) #4

It’s not a “question aggregation tool”. It’s a forum in the Q&A format. Why would that require more people to have a critical mass than a Discourse forum? (I’m assuming you meant Discourse requires much less traffic.)


(Chris Hanel) #5

Yes, Discourse requires a lot less traffic to function successfully as a forum than a StackExchange site does.

Discourse also has much broader purpose for a wider variety of online communities, so it has that going for it as well. Considering there’s a much wider potential market, I don’t think it would have much chance of suffering the same fate.


(Helder Ribeiro) #6

You’re just restating that without justifying it. Where did you get this from? There’s certainly no data on it, but why do you have this opinion?


(Philip Durbin) #7

Very little. Good point. I hope their business is sustainable. I’m rooting for them.


(Helder Ribeiro) #8

Apparently, yes (bold emphasis mine):


(Chris Hanel) #9

Sorry, you ended your last post indicating I might not have been clear so I was making sure.

I don’t really think I’m going out on that far of a limb making the conclusion that a more standard discussion forum like Discourse has much wider potential and a lower bar for sustenance as a community than a StackExchange site.

StackExchange has two things going against it that Discourse does not: First, is narrower in purpose. It’s built to be able to ask direct questions to a community at large and hopefully receive a timely, verified answer. The number of websites or communities that would see a StackExchange instance as being something useful to them is small.

Second, the ingrediants required for a StackExchange site to prosper include a topic that guarantees the continued need to ask questions (and require human responses rather than google searching), and enough people that are knowledgable on the subject and invested enough to share what they know. If everything worth covering has been covered, activity dies. If there isn’t enough variety in the userbase to have an answer for the full range of questions, then activity dies. If nobody has anything worth asking, then users aren’t going to stick around and monitor for questions they can answer.

Meanwhile, Discourse as a discussion forum has much broader use. Just from my own vantage point in this corner of the internet, I could rattle off an endless number of gaming groups, webcomics, dev teams, podcasts, support groups, sports clubs, etc. that have use for Discourse that would never dream of bothering with a StackExchange system. There are forums with 10-15 people on them that have been alive for a decade, for them the system is serving its purpose. I can’t think of a situation where a userbase that small would get the desired output of StackExchange.

That’s why I think your concern about the hurdle of building up to critical mass in order for Discourse to be useful is misplaced. I believe that this new project will serve the needs of far more customers, and those customers will have to work a lot less in order to see that effort result in something sustainable.


(Adam Davis) #10

I strongly suspect that the potential for SE to make money managing the sites themselves was much greater than the potential letting other people manage them, and paying SE a fixed cost per month.

There were many sites that were successful in their niche, which are cash cows, such as moms for moms, gaming sites, etc. The people that started them would have been only too happy to pay SE and make the money off advertising and referral links. Many of them created their own places using clones, in fact, and are doing well.

I believe it’s because they chose to do closed source, and keep everything under their wings that things didn’t take off nearly as much as they could have for the platform, but it probably was the right business decision. It doesn’t benefit the community as much as it could though.

But there are many successful businesses that host forum software, so we know there’s a market. Further there are several businesses that sell forum software, so we know there’s a need for good software that is under continuous development.

If anything, I think the business case for discourse is stronger than the business case was for SE at the start.

But I don’t think SEv1 failed. I think it was actually too successful, and was yanked so it could be made more profitable, rather than the other way around.


(Helder Ribeiro) #11

True, there’s a market, but when you raise a large amount of VC money, it’s not enough to have some market, it has to be something huge. Is there a $1B+ market for forum software?

There were only a handful. Math Overflow, moms 4 moms, and I can’t think of any other, and I don’t know that any made any significant money. From Joel Spolsky himself on the StackOverflow blog:

We got a lot of ghost-town sites that nobody visited. We also got a lot of duplication: multiple sites on the same topic, competing for the same people and preventing one other from hitting critical mass.

Bottom line, it just wasn’t working. We’ve been in beta for half a year now, and we only have a handful of sites that get enough traffic to provide quality, timely answers to difficult questions.


(Helder Ribeiro) #12

I also feel that if the platform had been open source from the start, they might have seen more traction grow organically over time. I wish there were high-quality, well-maintained, liberally-licensed open source Q&A software like theirs for me to run a few things on.


(Adam Davis) #13

It didn’t grow the way Joel wanted, yes, but they also felt that having ghost town sites reflected badly on the company, which wasn’t good for investors. If they had let them grow slowly and organically, yes, they would have had a few hundred ghost towns, but they also would have had dozens of really good sites that grew from nothing.

The ghost town sites certainly weren’t hurting the community. Duplication was causing some issues, there were two good electronics communities that started, but they could have kept going their separate ways and been fine. The idea that SE had to step in and force the communities to adapt to their definition of success in order to save the communities is ridiculous. They were not doing it for the communities, they were doing it for SE. And once they decided to go that direction, they had to shut down v1 in order to get people to move over to area51.

I think they did what was best for them and their investors, and they are very successful at making money, and creating value for millions of people on the Internet, which can use the service for free. It’s great. But it should be abundantly clear that SE makes the majority of its decisions based on the bottom line, rather than how useful a platform they can make their software for the world.

Regardless, I don’t think the two situations are comparable enough to draw any sort of useful conclusion.


(Lucas Nicodemus) #14

If I had to give my two cents, it’ll be successful on a hosted model because of the irregularity of the stack it’s hosted on. If it gains popularity with end users, communities will want to adopt it. Not many places host this kind of stack, and if it’s priced properly (say $20/mo for 1000 users), people will revert to buying it rather than muddling with it themselves.

I’d say based on requirements alone it would be nigh impossible to host this on shared hosting easily. Hosted WP is successful because it’s easy to sign up and run with - and even then it’s relatively simple to host on shared (even free) hosting compared to this.


(Jeff Atwood) #15

Well, lots. Pretty much everything.

Stack Exchange does not do discussion. What Stack Exchange does is focused, fact based Q&A, which is the opposite of discussion, and they’ll be the first to tell you that. More specifically:

  • Stack Exchange was intended to displace experts-exchange and other “evil” or extremely low quality Q&A sites. Topics are generally scientific or technical in nature, things that can be verified with facts, and strictly limited to Q&A.

  • Stack Exchange is primarily designed to service search engines and produce useful artifacts. For example, when searching for a programming problem, Stack Overflow is one of the first hits (often the first!); users click the link, read what people have to say, then go back to work. Write once, read thousands or millions of times.

  • Discussion is strongly suppressed, except insofar as it enables better questions and answers. Only the minimum amount of discussion necessary to support Q&A is allowed. Changes are best presented as edits to the underlying question and answers, not a long chain of back and forth text.

  • Moderation is necessarily strict; only practical, answerable questions that support the topic and are useful to future visitors are allowed. The difference can be subtle, particularly to users unfamiliar with the Stack Exchange model of high signal, low noise Q&A

  • Popularity is frequently at odds with the goals of Stack Exchange.

  • You spend a lot of time educating the community to encourage behaviors that are conducive to learning, not necessarily entertainment. It takes a lot of ongoing discipline from moderators and experienced community members.

  • Not all topics and communities can sustain this technical Q&A format where the best answers should be verifiably and provably correct, at least in some small way.

Stack Exchange is not open source. And I’d argue it doesn’t need to be, and shouldn’t be open source, since the model requires such strong rigor and discipline. If you open sourced SE, all you’d accomplish is thousands of failed communities pop up all over. Have any of the OSS clones of SE succeeded? I would argue none of them have. Because strong curation and centralization is necessary for the model to survive.

This is why Area 51 exists, to point experienced Stack Exchange community members at new sites so they can be seeded with people who grok the model – and most importantly, have the discipline to enforce it.

Q&A is a small part of what many communities do. For example take LEGO. How much of LEGO is strict, fact-based Q&A? Most LEGO communities would need the following:

  • look at this cool LEGO thing I built!
  • buy and sell rare LEGO kits to each other, people who get it
  • share an interesting LEGO story
  • ask about the best ways to build LEGO
  • ask about the best places to find LEGO
  • hanging out and talking about non-LEGO stuff with other LEGO enthusiasts

None of that is allowed on Stack Exchange, by definition.

So there’s your 3 key differences:

  1. Stack Exchange does not do discussion, by definition.
  2. Stack Exchange is not open source.
  3. Q&A is a small part of what many communities do.

Now consider how Discourse works for comparison:

  • Discourse is intended to displace all existing forum software, and become a part of the fabric of the Internet. Topics are determined entirely by what the community will allow, limited only by the community’s imagination.

  • We expect people to find discussion topics via Google, but what Discourse encourages most of all is participation to build a community. We want people to identify with a forum. Rather than presenting the answers to their questions, Discourse provides a platform for engagement (what do you think of, what’s the best, how do you like to…) and a long lasting experience.

  • Discussion is encouraged, for any reason at all, for as long as everyone likes. As people respond, the whole discussion flows downstream chronologically in traditional fashion.

  • When everything is “on topic”, moderation is far easier, because the community only needs to decide when users are trolling or being overtly rude, not make value judgments on the relative utility of contributions as they do on Stack Exchange.

  • in Discourse, popularity can be encouraged without reservation, because it means the discussion is working and people are having fun. It’s OK to produce only 5% or 10% useful content to the outside world.

  • As long as community members aren’t being mean to each other, they don’t have to spend a lot of time educating other users about how Discourse works; they are free to engage in whatever topics they find enjoyable and interesting.

  • Any community, provided it is of reasonable size, can sustain a forum where its users share their thoughts, opinions, and feelings.

Really, Discourse and Stack Exchange could not be more different. In practically every way I can think of they are in fact opposites.


(Helder Ribeiro) #16

You only mentioned cultural/format differences, but the main concern is about the business model (I’m really starting to like Discourse, same as I did SEv1, and I very much want you guys to make shitloads of money).

The business model doesn’t seem to be too different between Discourse and SEv1: host your community => give me money on a monthly subscription.

Based on what you said (and @ChrisHanel on a post above), the hope is that since the format is less constrained, a forum would need orders of magnitude less people to be useful (i.e. have critical mass), and would thus appeal to orders of magnitude more forum admins who would become paying users. Is that the line of thinking, business-wise?


(Helder Ribeiro) #17

This is the whole point of this thread. Not how successful SEv1 was in making useful software for the world, not how many sites could’ve slowly or organically grown, but how that business model worked out for their bottom line and their investors. Discourse also has both (bottom line and investors), hence the original question. You’re saying yourself that SEv1 wasn’t a good enough money maker, and that’s exactly my point.


(Grandy Peace) #18

In my mind, there’s nothing about Stack Exchange that can make it properly a third place. I love various Stack Exchange sites, and Stack Overflow, but in fulfilling their narrow purpose so well they sort of rule out being a third place.

Web forums are a possible third place . The body of existing web forum software is inadequate, however. I find it inadequate from a “forum participant” standpoint, and I find it inadequate from a “forum admin” standpoint. I’ve used quite a few of the major offerings over the years. I don’t yet know if Discourse is the next generation or not (I do like some of what I am seeing, though). But I agree with Jeff that forum software needs a champion, and it is in bad need of having new life breathed into it.


(Jeremy Boyd) #19

Another major difference is the price point. IIRC SEv1 cost over $1000/mo for a community to organize a Q&A site, which not many could afford, let alone had the community to need that many questions and answers. So SEv1 cost too much for you to not use it.

Where as if this is priced in the $10-99/mo range, you wouldn’t need a very active community to get $50/mo worth of usage out of it. After all, what is $50 - Taco Bell and a movie. With that low of an entry cost, way more communities will be able to be hosted, and since they wont need many resources, they can be hosted shared, costing way less than an install of SEv1.


(Helder Ribeiro) #20

Small factual check: It was more like 200 USD/month. But that’s still one order of magnitude above the number Discourse is flying around, so I think yours is a valid point.