Hacker News new | comments | ask | show | jobs | submit login
Ask HN: room for a new forum platform?
8 points by retube on Aug 20, 2010 | hide | past | web | favorite | 17 comments
Stackoverflow obviously raised the bar for Q&A type forums. Whilst that format might not be suitable for all use-cases, so many sites seem to be implementing the old web 1.0 forum - e.g http://mailwatch.co.uk/forum/

I'm guessing many of these are self-hosted too. What opportunity do you think think there is for a updated, threaded, hosted forum service is? What kind of features would you want?

That's something I often thought myself, phpBB/vBulletin type of forums have been around for 10+ years now, and they haven't changed a lot since then.

I think there might be lots of things to rethink on the UX side now that we have so many new frameworks available.

If you aim to build a 100% hosted service, I feel like suggesting to let the client keep ownership of the database at least. I wouldn't like to give total control of all the posts on my forum to somebody else.

I'd say simplier & better looking interfaces + open customization features is the path to follow. Just my 2c.

Yes - it seems amazing that 90+% of web forums are the old php/vBulletin type. My major gripe with these is that they do not have threaded comments. Not to mention a very old-skool UI/UX as you say.

Interesting what you say about db ownership. And I guess for established forums there would be migration considerations too (maybe this is why they haven't upgraded?)

Customization would be no problem. You could allow users to source their own css for example.

Some anecdotes from running multiple forum sites over the years:

Most forums end up using flat replies because normal people don't understand threaded replies. A friend at YouTube recently said that their research backed up that anecdotal assumption.

A couple of the major forum vendors tried offering hosted solutions and it wasn't sustainable as a business model for them; they both went back to simply licensing the forum software itself.

Forum communities are like any other social system: network effects and "where your friends are" are what keep you from leaving. Most "power user" features go unused by the majority. Stats like only 3% in one community use the friends list, only 10% in another use the profiles, etc. It's never about the feature set, it's all about the community.

Existing forum software is basically a black box. It's hard to integrate with and expand. It's hands-off for the admin until they realize they want a site that's more than just a forum.

It's interesting what you say about flat replies. Although a "technical" person, I can't believe that users are not wanting to reply to individual posts. Perhaps it's a UI issue?

There are certain social dynamics that have grown up around this variety of single-thread-per-post forum: discussion is channeled into essentially one discussion per post (you start a new post if you want to fork it); threads are "bumped" when someone replies and the expectation is that it's a reply to the main thread of conversation; the pagination encourages discussion to move on (in a 50-page thread, you don't normally belatedly reply to something on page 8, unless you have a good reason for it); etc. Threaded discussions are very different in style imo, and have less of a "conversation" feel.

I tend to prefer threaded overall, but there are some non-threaded forums I'm on where I don't think the community would work as well with threaded. For example, I don't think the SomethingAwful forums would work well at all as a Slashdot-style (K5-style, HN-style, reddit-style) threaded forum. Actually, I think some parts of reddit would work better with unthreaded; there's often a sort of urban-sprawl feel to the thread branching there.

I was recently tasked with looking for forum solutions to replace phpBB for a contract client and yes there is definitely room. As many others have already stated phpBB is bloated and anyone who has had to dive through the code knows that it certainly doesn't have Wordpress-like extensibility. There's also the problem of user experience. Having been involved forums for the past few years I never noticed usability problems, but our less tech savvy users made sure to point out their gripes with it (bbcode, images, search). Stackoverflow and Quora certainly raised the bar but they are closed solutions that I still can't put on my own community site. Q&A sites also solve a different problem than forum sites. They're meant to answer questions, not foster discussion. This means that they lack several features of forums such as images and rich text editing. I also looked at Reddit type solutions but they have the same problems with discussions based solely around links. I would love to see a self-hosted solution that's somewhere in-between phpBB, stackoverflow and reddit.

There is a great market for that. Sparklit (http://www.sparklit.com/) has a hosted product (which looks might be derived from vbulletin), for example, but pretty much everything out there I've seen is old (1.0, as you put it). There is lots of room for innovation.

Data portability is still a huge problem. There is a massive lock-in once you get a forum going on a hosted platform because it's such a pain in the ass to get everything out.

I've considered rolling my own, but decent forums are deceptively complex. It would take way too much effort.

Go for it.

Quora has also dramatically raised the bar. If you haven't already, go join and ask / answer a couple of questions. It's slick, efficient and impressive to use.

My feeling is that, for forum software to really be useful (and lucrative), it needs to be usable by businesses. That means allowing customization, probably allowing file attachments, and integrating with a wide variety of software via APIs or installable plugins.

Hi Ben - a couple more questions: what kind of APIs do you envisage? And what sort of plugins?

I hear Quora is good. But they only seem to allow login/signup via facebook.

Or Twitter, I think. But yes, it's a very odd choice to not allow (1) OpenID logins, and (2) basic registration - particularly given their early adopter audience.

Maybe they don't allow it because of their early adopter audience: who among that crowd does not have a twitter or FB account?

I'm obviously not an early adopter :) Actually, in seriousness I'm probably not. But still, I like to remain anonymous and only reveal personal data when I have absolute control and trust in over who/where I am releasing it to (and I don't mean I don't trust Quora, it's exposing myself to their users that I have an issue with)

I didn't like Quora at all. It allows users to modify your question and tags and some of them misunderstand your question or just don't want it there. One of my questions received no answers, many followers hoping for an answer, and 13 edits removing tags and changing the question to be something completely different from what it was. In other words, it can be abused.

thanks for letting know about Quora

As someone who looked into phpBB and managed a vBulletin installation for a company I worked at, I will say that yes there definitely is room. I was also one of the first couple hundred users and one of the early active participants in two forums that grew to thousands of users (head-fi.org, thewolfweb.com), so I've experienced first hand the primary problem of forum admins: growing the community and handling that growth

Here are the problems with the current options:

-- phpBB and vBulletin are self-hosted. The new platform needs to be SaaS.

-- Both platforms suffer from feature bloat, especially vBulletin. Forums aren't just products, they are tools for managing communities. Whatever product is designed to replace phpBB and vBulletin needs to begin with the basic premise that the forum participants are the users and not the forum administrator.

-- phpBB and vBulletin right out of the box come with too many useless features enabled. Most are features you don't need or that you don't need until the forum already has a large number of users. They complicate the forum in the early days and hinder growth. A forum is like a shark. If their isn't movement/activity it dies.

-- One key mistake almost all first time admins make is making too many forum "rooms" too early in the game. A forum should start with no more than maybe two to three "rooms". The software shouldn't prevent you from making more, but it should advise you against creating more rooms when the "activity level per room" is still low. The company who builds this software should actually discover the ideal amount of activity per room level to promote growth via aggregate A/B testing across all its clients forums.

-- In fact, this previous suggestion is the first in a long list of features that help "prevent new forum administrators from shooting themselves in the foot" kind of like a safety switch on a gun.

-- The new forum needs to be designed first and foremost to help its admins grow the user base and help stimulate user participation. This is key. Most new forum administrators don't have a clue on how to do this. I would say that the startup that tackles this problem would need to try a Mint Blog approach, which the blog educates the product's users on the topic of community management.

-- Were I the product manager for such a product, I would read Clay Shirky's "A group is its own worst enemy" several times before starting and regularly while designing the product. The idea in that essay is really important to understanding the state of forum software today and the problem you are trying to solve. Avatars are an example of a problem with forum software today. The avatars in both programs can be modified to be any size and can include animated gifs and whatnot. Basically, the avatar features are currently configuration over convention. On top of that the current defaults for avatars leads to a situation where avatars are so big that they create a lot of whitespace between responses that interrupt the feeling of the conversation threads. Allowing animated gifs results in distractions from the primary activity.

-- Invite features - Last time I checked, none of the current forum softwares are designed for viral/social growth.

-- Migration tools are necessary. You should make it easy for phpBB users and vBulletin users to migrate to your new forum platform.

-- Social network integration - The software should be designed from day one to have an API and to integrate with Facebook, Twitter, etc.

-- Constant email notifications by default to stimulate user participation just like Quora does. However this needs to be simpler to "dial down" as a user becomes more active as it eventually clogs up the email inbox of that user. Quora's control panel for this isn't very user friendly as it is very granular. A better approach would be to sort the types of notification emails by historical frequency from those that user has gotten the most of to those that the user has gotten few or none of. This makes the decision very easy for the user and promotes them to only turn off the notifications that are actually inconveniencing them.

-- I can cite at least one example that demonstrates that this is a product missing in the market: Zendesk. Zendesk comes with forum features, not because forum features are a necessary feature of a helpdesk, but because companies that need help desks typically need forums and there was no company that Zendesk could plug into via API for plug'n'play forum support. The forum software in Zendesk is adequate for most users' needs, but it doesn't scale well because it was always designed to be bolt on to the helpdesk itself.

-- I haven't used GetSatisfaction yet, but I would imagine that this is a shared issue with them as well.

-- Forums basically have two personas: hobbyists and shared-interest communities (i.e. head-fi.org, stevehoffman.tv, thewolfweb.com, church groups, etc.) and companies.

That all being said, I would say that an similar product that is also needed, especially inside companies, is QaaS. Quora as a Service.

A lot of companies want a Quora type solution first and foremost than a forum. I would say that Zendesk needs a product like a Quora-as-a-Service instead of a forum.

I'm currently working on my own startup right now, but if someone were to want to build this, I would be very happy to serve as an advisor/board member for a small stake of equity. This is a problem that I've given a LOT of thought to and that I would love to see someone solve.

Thanks for the very detailed response. My comments - and some are "devils advocate":

- Agreed that a Saas offering has lots of cost/admin benefits. But - it makes it difficult to integrate with a sites existing signed-up member-base or social network.

- What are examples of feature bloat? What are the key/core features?

- What would be the equivalent of a room on Stackoverflow? I think that SO is really just one room, with a lot of questions in that one room (ignoring tags)?

- Invite mechanism - yes. - FB/Google login/interface - yes. - Email notifications - yes

GetSatisfaction - can't comment. I thought this was more about feedback wrt to a website or business, as opposed to a fully fledged discussion thread, hosted say on a national newspaper (an ultimate customer here)

I've never used Quora (fb only sign-up). What's the key differences between Quara and a forum or Stackoverflow?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact