Enterprise 2.0: Corporate Wikis reviewed: Confluence, JotSpot, WetPaint, Socialtext

(For the most recent articles on Web 2.0, check out my full Web 2.0 articles category.)

(This original post was written in July 2006. There have been MANY updates. Please read all the way through to the end of the article for all updates as my opinion has changed over time as the vendors have updated their offerings.)

(Most recent update: Feb 28, 2007 is at the bottom)

Wikis are on the rise in corporations. And it’s about time. They are light weight replacements for heavy weight knowledge management systems and are also a way for your user community to generate content that is better, faster, and probably easier to read than you can as a vendor. One way to enable them to contribute would be to build a wiki and let them flesh it out. Some good examples are coming up in this article: “Corporate wikis breaking out all over: MSDN Wiki” by Dion Hinchcliffe. (He has another great post as well called “Exploiting the Power of Enterprise Wikis“)

Quote of the day: “Not leveraging the contributions of a company’s most impassioned and enthusiastic customers is starting to be seen as a significant oversight in many business circles.”

It appears in the article that eBay is using Wikis to better communicate between their users, partners, and suppliers. Now MSDN (Microsoft Developer Network) is using their pages to improve the quality of their developer documentation with the MSDN Wiki. THAT is a great usage. Your users often know your product better than your engineers and product managers because they have to live with it day to day. And guess what? If they tell the truth about some part of your product being broken – that’s a GOOD thing.


I have looked at a lot of the corporate wiki tools in the past six months. Unfortunately we are still at the awkward pre-pubescent stage. I haven’t yet found the best tool that has all of the following characteristics. I want a tool that:

  • is beautiful
  • is easy to use
  • integrates with our corporate authentication system
  • allows users to “email to a page”
  • has rock solid WYSIWYG editing (including tables!!)
  • can absorb Microsoft Word’s messy HTML without barfing
  • has great management tools for pruning and gardening the wiki
  • allows for attachments of files
  • includes great threaded conversations on a page
  • is hosted but can also be used in an appliance inside the firewall
  • has granular user & group security privileges (ideally tied into Active Directory)

Atlassian’s Confluence is the best of them so far. Pros: the overall design is clean, it has advanced management tools, good security, and simple attachments.Its email function has to pick mail up from a POP box which makes it a little bit less ad-hoc but still functional. And most importantly, it also has great tools for moving pages around. Cons: Text editing, like with most apps these days is a bit dodgy, and pasting in blocks of text from Word is likely to cause problems. The pricing model is reasonable but for some reason (possibly because they’re from Australia), they still don’t have a directly hosted option so you have to use somebody like Contegix or deploy it on your own box. This seems to be a big and obvious oversight on their part these days. Also, their pricing model doesn’t encourage small deployments right off the bat. I think this is the one that we’ll use more of internally at the company where I work.  Summary: The best of the enterprise wikis today, and one of the best options for scalability.  [Update: see updates at bottom –  they now have a hosted option and low-end pricing.]

Next would be JotSpot. I can only recommend JotSpot for really small deployments though. Pros: The text editor is pretty good and the tables even work reasonably well. The pricing model is good, the setup is quick, it’s pretty easy to use and understand, and you can also buy it in an appliance that you keep inside your firewall and integrate with your security infrastructure. Cons: The design is quite clunky, the font sizes are huge, there is a lot of wasted screen real estate, the “applications” they can install are really lightweight and not that useable (what’s up with that awful blog module Joe???) There are also bad page renaming bugs that can and do bite you in the ass. Summary: I have built several wikis in JotSpot and it’s probably the best for simple wikis for smaller teams with a few pages. But I wouldn’t want to scale it out very far. It isn’t robust enough. [Note: Jotspot was acquired by Google and has closed off the ability to register but has also stopped billing its current clients. See below for more information.]

WetPaint is a newcomer that is doing some interesting stuff and that might be a better bet than JotSpot. Pros: The design is beautiful, the tool is very easy to use, the text editor is one of the best I have seen. Cons: I’m not clear on their entierprise suitability and it’s not really their target market. It didn’t appear that they had much in the way of administration tools, granular security, or any way to integrate into a back-end authentication system. Summary: I met one of the WetPaint guys at Gnomedex but he didn’t seem to know the product very well. Hopefully next time, they’ll put somebody more knowledgeable at their booth who knows the product in more detail. I think they’re worth watching to see what they do in the next few months.

Socialtext: I have tried on at least four separate occasions to use and like Socialtext but I can’t. Pros: Ross Mayfield, the founder, “gets it”. He understands that lightweight wikis will be important in the enterprise. It includes “email to a page” functionality and you can buy the Socialtext appliance and integrate it into your back-end. Cons: the text-editor is the weakest of the bunch – they only just added WYSIWYG in the last few months and the table stuff is quite awful. The word HTML handling is poor. There are few management tools. Summary: I just can’t use this application. I don’t know where they’re putting the money that SAP invested in them because I can’t see evidence that it’s improving the application. I can’t recommend it for anybody. Go with one of the other three above instead. [UPDATE: See Feb 28, 2007 notes below.]

I didn’t cover all of the many open source wikis here: MediaWiki, XWiki, Twiki, etc. because most, if not all, have to be downloaded and installed and I’m more interested in enterprise quality wikis with the feature set I noted above.

This post brought to you by Joy from the album “Thankful” by MaryMary

UPDATE: I modified my summary line above for Confluence to make it more complementary. I was feeling a bit sour on the whole category when I wrote it before and had written “It is the best of the worst” – a phrase I often use when discussing the best in a category that has significant problems. But it’s more fair and accurate to just say “the best so far.”

I also saw a note from Jon Silver at Confluence on their blog responding to my comment on hosting.

He states:

“The question of hosting comes up now and again. To set the record straight: it’s not an Australian thing—there are many good examples of Aussie companies bringing hosted software to the world! Software as service has gained a lot of attention by great companies like Salesforce.com. It’s a good model. But we’ve found that our customers really enjoy being able to host the application themselves and fully customise it to meet their needs. For some organisations, their security policies require them to manage software themselves. Nearly 1,500 customers in small, medium, and large organisations use Confluence inside their firewalls today. As pointed out in the review, Contegix (and other Atlassian partners) are available for people who need a hosted wiki solution.”

I disagree very strongly and here is why. Australia is far behind the North American market in terms of broadband penetration so your view is probably skewed by this. I’m also not stating that it’s binary and that it must be one way or the other. I’m commenting on the fact that you have too many barriers to adoption that are precluding you from making the money that you could make if you would get out of your own way. First barrier: I can’t just fill out a form and start using it. Second barrier: I would have to have the IT resources to find a box to run it on. Third barrier: I would have to download and install it and get it working – I have ten years of computer consulting and I gave up on your install process because it wasn’t a point and click install – you’re still selling to geeks. Fourth barrier: Your pricing model puts Confluence out of the range of people who want to start small and grow.

Fix those problems and you have a chance at dominating this space. But right now Socialtext and JotSpot which are far inferior to your product, are probably going to beat you simply and easily because they are hosted, they are easy to test, they are easy to use, and they have great and simple pricing models. Which will be too bad because once again, better marketing (focusing on the basics of removing barriers) will win and the better technology (your product) will lose.


Jotspot was acquired by Google for an estimated $50M on an estimated $3-4M of trailing revenue. NICE MULTIPLE. Sheesh. Can’t help but wonder if those numbers are real. 12x trailing revenue??? Wow. Not sure of those numbers though in any case.

Wetpaint raised a second round of financing. The plot continues to thicken.

And Atlassian is releasing a hosted version! At the risk of sounding like a PR flak, I’m going to give a lot of kudos to the Atlassian team, some of whom I have come to know over the past six months. They are a nice bunch of people and are very smart. They are finally launching a hosted version of the wiki and were so kind as to let my company test it out over the past few months while they worked the kinks out of it. Within my company, four different project teams independently came to the same conclusion as to which was the best enterprise grade wiki (I led one of those teams) and we have now standardized on Confluence as our enterprise wiki tool. So far, the adoption is reasonable, but we haven’t done a ton of marketing yet and had too many sign-up / authentication barriers in the way that we are now removing. The pricing looks like it will give them a great low-end feeder product that could help them achieve the much-desired “foot in the door without IT involvement” in small or large enterprises. It may end up simply being lead-generation rather than profit center but that’s not a bad thing.

Confluence still looks and feels like an application sold to early-adopter geeks but the team at Atlassian is very aware of the user interface deficiencies and are working to remedy that.

Way to go guys. I hope you have an amazing 2007!

February 27, 2007 update is here

(19) Comments

  1. Interesting take. Some notes on Confluence – Adaptivist.com offers Confluence hosting and why does your top choice get “best of the worst”?

  2. The reason I said, “best of the worst” is that none of the apps do it all. All of them have brutal text editors (all web 2.0 apps have brutal text editors). I would put Confluence at about 7/10 for enterprise use. The rest are far lower on the scale.

  3. Great points. I probably should have replied to your review a little differently yesterday. As for the hosting model, we think it’s a great model, and I (and many others here) agree with you on the barriers to entry (at least #s 1-3). In fact, we may decide sometime in the future to offer a hosted service ourselves. It’s something we’ve discussed, not only amongst ourselves but also with customers, bloggers, and other wiki afficianados. The people who have adopted Confluence so far have been very pleased, but as you point out, it requires resources and a degree of technical proficiency that not everyone possesses. If IT services go the way that Marc Benioff predicts, we’d find ourselves out in the cold. Even the good folks in Australia are aware of the conversation, and actually, for the record, I’m in San Francisco… far from the Outback ;). As for being trounced by the competition, time will tell. We very much respect and like the people we’ve met at both SocialText and Jotspot, we admire their evangelism of wikis, and they seem to be doing very well. If you’re ever in the area of San Francisco or Sydney, please stop by and say hi. I think you’ll find we’re not that out of it. 🙂

  4. Troy, I like every word of your update re. Confluence. In fact I just met some of the Atlassian folks in their SF office, heard their point of view about enterprise customer’s preference for a solution behind their firewall – that is eventually, WHEN they become enterprise customers – and I told them the same.

    What’s great about the “Web 2.0 model” is that you can forget the wasteful enterprise sales model – the technology sort of sells itself, by small user groups starting to use it, without IT “blessing”.. then it spreads, becomes viral, eventually the CIO gives in and blesses it to the entire corporation 🙂 This bottom u p adaption by non-technical business users is what’s at stake by not offering hosted access. That says, Confluence is doing very well, financially beating the other two together… so who knows how much more they could do by offering both on-demand and on-premise.

  5. Another “wiki without the wiki” to consider for business and enterprise use is http://www.centraldesktop.com We’ve adopted it in our small business (we are a small advertising firm) after trying several the wikis you review in your post (except for Confluence…it appeared like it was “too much” for our small business.

    I agree with Zoli….WetPaint has no business in this list.

  6. Another “wiki without the wiki” to consider for business and enterprise use is http://www.centraldesktop.com We’ve adopted it in our small business (we are a small advertising firm) after trying several the wikis you review in your post (except for Confluence…it appeared like it was “too much” for our small business.

    I agree with Zoli….WetPaint has no business in this list.

  7. Hi,

    I’ currently trying to start a blog on the subject of corporate wikis (WikiBC) and hence conducted some research on the topic. I think that 2 things are to be considered about your post :

    First, hosted wiki solutions can evolve easily (I am a SocialText user, they just started their interface 2.0, far better than the first, and it only improved my situation : the only drawback is that I couldn’t accessed it for 4 hours) and should be evaluated on the basis of expected potential mutations rather than their present state only. Whatever the wiki you choose you will find features you either do not like or you miss. The best wiki company will be the one that address its consumer needs the soonest and this does not necessarily mean the most advanced one today.

    Second, a successful wiki is not about features, it is about employees adoption and usage. Even with a basic structure, if you can convince people to use it by showing them what they can get out of it, it will work. The wiki offers a potentiality of knowledge sharing and structuration, of improved communication. Getting people to use it has far more to do with a good formation and motivation process than with the product itself (as long as it works).


  8. I agree with your analysis of our wiki Confluence, your take on our being a bit late on hosted, and especially the priority you give to UI and the need to look “beautiful”. The last area — UI and usability — is one we need to give higher priority now that our enterperise features — security, search, etc. — get strong reviews from customers.

    Our big enterprise effort was the release of Confluence Massive this month which delivers clustering for infinite scalability. We did this because we were getting a lot of pressure from companies to roll out to huge user populations: IBM has 50K users on one Confluence intranet, and SAP’s Developer Network which uses Confluence has a 500K user data base. Also a number of universities including MIT intend to roll Confluence out to the entire student population.

    Now that Massive is out, expect to see us concentrate on shorter release cycles again. UI matters, and I appreciate your critique.

    Your take on Australians and hosting is intriguing. [I’m an American so not being defensive here]. The reason we concentrated on installed software, up until now, is most buyers of wikis have been IT people, and most companies have preferred this option, if you compare — I will humbly point out — our sales vs. other commercial wikis. Atlassian also built an open architecture and a plugin library that lets customers build on top of the wiki, which installed software fits.

    None of this changes the need for us to offer hosting now. The wiki market is very much a business and an IT market now. Thanks for the analysis. — Jeffrey

  9. Jeffrey, thanks for the comment. Much appreciated that you stopped by. As you know, I’m a fan of your team and what you’re doing there. Here are some follow-on thoughts.

    I’m really excited for you guys that you have been able to get Massive out the door. You’re way up there at the high end of the market when you’re talking about those kinds of audience sizes. I have no doubt that your app will be solid there.

    The Australia thing is probably no longer relevant. For many years, Australians were known for having little presence on the net because the pipe to Australia for the net in the early days was seriously under the necessary capacity. I’m sure that has all been fixed with the massive telco / optical build out of the 1995-2000s though so I was probably out of touch to make that comment.

    I’m aware that your sales eclipse that of the other wiki vendors – and that rocks. Your focus on revenue and profitability is admirable and is a lesson many others in this current gold rush would do well to emulate. This is why I think you’ll be around for a while. “Good on ya” as they say in Oz.

    I would also like to applaud the plug-in architecture that you’ve built – that was one of the selling points for our team when they looked at your software for sure. Having said that, many of the plug-ins really end up solving stuff that should be done by the app. Not being able to copy/paste Microsoft’s crappy HTML directly into some sort of parsing field at this stage of the game seems WAAAY overdue. The XMetal guys built some kick-ass tools for that over the past few years. Let me know if you want an intro. But then their XHTML editor would take another $1M to rebuild from scratch – it was fifteen years of development but could be rebuilt now in about 6 months. Text editors as you will learn are a sore point with me.

    Regarding the move to hosting, you and I discussed this once before but I think my thinking has evolved in a different direction than it had at the time we last spoke. At that time, I was concerned that you were missing the bottom of the market and would lose them permanently as they jumped on the disruptive, cheaper, hosted platforms below you. A standard innovator’s dilemma – you have overshot the low-end who just want a quick, light-weight wiki with low barriers to entry.

    And that’s still true. But the biggest gain I think from having software hosted vs. not is the ability to go from 6 month release cycles to “Flickr time” cycles of 15 minutes. The number of potential mutations of your software can be thousands of times more. The other critical piece of that is that on premise software generally isn’t instrumented to track attention data. Hosted software can be designed to both mutate quickly but equally important to track every single thing a person or group does in real time. That is the other half of the evolution cycle. Test, learn, adapt, test, learn adapt. With hosted software and an extreme focus on test, learn, adapt, a vendor can mutate their software from “it’s okay” to “this rocks and they keep coming up with stuff exactly when I want it!!!” I think that a lot of vendors will move from on-premise to hosted and will not pay attention to the attention data – they will consider it a business model and delivery change and will miss entirely that it is an enabler for radically fast product innovation. I’d be keen to hear your thoughts on that.

    Anyway, as noted before, I wish you and your team the best for 2007. I expect it will be a heck of an interesting year ahead for you!


  10. Thanks for the great post. We are seriously considering Confluence as an enterprise wiki for our software firm.

    Based on your experience with it, what kind (qty, quality) of human resources would you say this tool requires? The installed version, not the hosted.

  11. Hi there. Well, it definitely needs a good technical person who can manage the application. If you have Active Directory, then your tech person needs a bit more knowledge again.

    You need a “Wiki champion”, not a “wiki bully” to lead the efforts by finding and using those specific instances where a wiki really is the best way to achieve a goal, and who leads people there by making their lives easier, not by berating them to use yet another tool they don’t want.

    You need somebody who can do the gardening of the wiki, who can manage the chaos a little bit as the pages start to accumulate.

    And somebody to check back with users to find out what’s working and what isn’t.

    These might all be the same person but more likely it’s a small team: a technical person and a more user-focused person could probably wear all the hats.

  12. Your desired characteristics with the criteria “is beautiful”, which immediately brought to mind American Idol. In music, the geeks are the ones with perfect pitch, who can count time, and know what key you’re in. I then recalled that Mozart never made any money until his vaudeville period, and what could be more vaudeville than absorbing Microsoft Word’s messy HTML without barfing. Bobby McFerrin was a good musician for a long time before he pandered to the masses with “Don’t worry, be happy”. One hopes Confluence remembers its geek roots long enough that “enterprise” doesn’t end up serving as a synonym for skin-deep.

Comments are closed.