Wikia Search, the human edited search engine which we trashed at launch, continues to make incremental improvements (and thankfully they’ve turned down the “This is a Google-killer” hype machine).

Today they did something really smart – they released a Firefox Add-on that lets users add search data without going back to the Wikia Search site itself. That lets casual users who want to contribute to the project to do so with less effort, meaning they may contribute far more data.

The toolbar also alters search results pages from Google and Yahoo (sorry, Microsoft, you didn’t make the cut) and adds a rating widget and an Add button next to each result. Again, this makes it super easy for Wikia Search users to improve the project’s data without actually visiting the site.

Overall Wikia Search is starting to look like a real search engine, with decent results for a lot of queries. They’ve made a lot of progress in the last few months.

Wikia Search is finally ready to play with. Jimmy Wales admits that up until now his company’s project to apply the wiki concept to search:

Pretty much sucked. It has not been usable on a day to day basis.

We agree. But today, Wikia Search is beginning to suck a lot less. It has only indexed 30 million Websites, but it is finally rolling out a set of editing features that lets searchers reorder, add, remove, rate, annotate, and comment on results. It also makes it easier for anyone to try to game the search results. Although, as with Wikipedia, an spammers can be banned by the community. We should see some fierce edit wars on this one.

The Ajax interface lets you drag results up or down the page. Hover the mouse over a result, and a menu appears on the right. You can edit the title or description, and spotlight the result with a yellow highlight.

Here is what the editing mode looks like for a search result:

If you don’t like any of the results you see, you can add your own by simply pasting in the URL.

Wikia Search also allows people to “annotate” results. If you click on “annotate,” it opens up a window with the actual Web page in question. Anything you click on or highlight will be added to the result. For instance, I added Michael’s Twitter photo to the Twitter result for “TechCrunch”:

I also added a comment: “Best way to keep up with Arrington.” If other searchers think those additions are helpful they can rate the result highly, if they think it is spam, they can delete the additions. Wales is relying on the community that grows up around Wikia Search to ban anyone trying to game the system. He is also relying on it to create the best search results. He says:

This is Day 1 for that really active community participation. In terms of what happens next, we will find out in the next couple weeks. How broad, how deep, how much activity.

And if searchers don’t like what they find on Wikia Search, handy Google and Yahoo icons make results from those search engines just a click away.

Here’s a video with Wales demonstrating the new features:

Click To Play

As Google gets into the wiki space with Google Sites (the relaunch of Jotspot), all the other little wiki startups out there will need to keep one step ahead. Those includes Wikia, Socialtext, Wetpaint, and PBWiki. As it approaches 500,000 wikis, PBWiki is now putting the 2.0 version of its site into beta. The latest version includes an updated UI, folders, enhanced access controls and an easy way to customize the look and feel of your wiki page (see the screenshot of a customized TechCrunch page below or this generic demo page). The 2 million people a month that the company says visit PBWiki should like that.

Personally, I find the UI to still be something that an engineer would love more than a graphic designer. But it is an improvement. Adding skins is a move in the right direction. What’s your favorite wiki?

Wikimedia Foundation posted their audited 2007 financial statements (I’ve embedded the document below) last week. Their fiscal year actually ends June 30, so these are already almost eight months old, but they reveal some interesting information about the entity that controls Wikipedia nonetheless.

Generally Wikimedia publishes these five months or so after the end of the year; this year they took eight months. Total donations and other income increased from $1.5 million in 2006 to $2.7 million (the period covered is prior to their recent fundraising effort). Donations of Google stock actually made up a material portion of contributions – 681 shares were donated in fiscal 2007 (worth about $315,000 based on the current stock price).

Travel expenses jumped significantly from $140k to $264k. Given that this period included time when Jimmy Wales was pitching Wikia Search around the world, some conspiracy theorists are speculating that travel expenses related to the for-profit Wikia (which Wales founded) were being reimbursed by Wikimedia Foundation. Wales, however, told me via email that the foundation does not reimburse him for any travel expenses at all, even for pure Wikipedia events, in order to remove any doubt about mixing funds between the entities. “I fund all that myself, out of my own pocket personally,” he said.

The financial statements also note, though, that Wikia and Wikipedia do share some infrastructure costs, assets, employees and expenses:

The Organization shares hosting and bandwidth costs with Wikia, Inc., a for-profit company founded by the same founder as Wikimedia Foundation, Inc. Included in accounts receivable at June 30, 2007 is $6,000 due from Wikia, Inc. for these costs. The Organization received some donated office space from Wikia Inc. during the year ended June 30, 2006 valued at $6,000. No donation of the office space occurred in 2007.

Through June 30, 2007, two members of the Organization’s board of directors also serve as employees, officers, or directors of Wikia, Inc.

Financial statements for 2006 are here.

Wikimedia Foundation, Inc 2007 Financial Statement

The organizers of the DLD conference in Munich put on a great show today. One of the more lively sessions was called “Humans Disrupting Algorithms” and featured Wikipedia/Wikia Search’s Jimmy Wales and Mahalo’s Jason Calacanis, moderated by Fortune’s David Kirkpatrick.

Jimmy and Jason each gave a brief overview of their human powered search engines. Jason railed on Google and other big engines, saying algorithms have failed to control spam and SEO gaming, and that humans must be involved to get good results. Jimmy was more circumspect, and spent most of his time arguing that large numbers of people will be willing to spend time helping Wikia Search develop good results.

Perhaps the most interesting moment, however, was when Google’s VP of Search Product and User Experience Marissa Mayer commented on human v. algorithmic search results from the audience.

ReadWriteWeb’s Marshall Kirkpatrick, who didn’t attend, has a good basic transcript of the session (proving to me once again that it is often easier to cover a conference remotely instead of batting crowds and dealing with terrible Internet coverage). I was able to take some video of a couple of interesting segments, though, embedded below.

In the first segment Wales gives the audience his overview of Wikia Search, and Calacanis jumps in with a few observations as well. The second is Marissa’s comment on what she sees as a false dichotomy – Google Page Rank, she notes, is based on real humans linking to sites on the web. Listening to her felt like a cold shower after a night of heavy partying.

As an aside, the DLD conference is clearly one of the better events I’ve attended in the last few months.

Many of us have waited a year as the Jimmy Wales hype machine promised a human powered search engine that could take on Google. Tonight that search engine launched at alpha.search.wikia.com, and it may be one of the biggest disappointments I’ve had the displeasure of reviewing.

First of all, it’s barely a search engine at all. It’s based on the open source Nutch software and contains an index of web pages created by Grub (a company Wikia acquired last year). The search results are poor and thin, as would be expected if not for the huge expectations that have been set. Absolutely no one is going to use this to search the web, until (and if) it is greatly improved.

But beyond the poor search results, there is really no “human” element to the engine at all. That functionality will come later, says Wikia CEO Gil Penchina. For now, users can add keywords to their profile – things that they are interested in, etc. When a search is conducted by others on those terms, the user’s picture is shown in the right hand column. Eventually, users will be able to edit and improve results for searches they are interested in. But currently, all users can do is add keywords to their profile that they might someday be interested in, and/or contribute to a “mini-article” that appears at the top of search results for queries (example).

And about those profiles. As anticipated, Wikia Search is yet another social network. User profiles include basic elements like a photo, adding friends, and information about interests and skills. And in a direct rip off of Facebook, Wikia Search profiles contain an activity stream of stuff you and your friends have been up to over the recent past.

Wikia search would be a disappointment even without the massive hype we’ve had to endure. And taking that hype into account, this product is an inexcusable waste of time.

To be fair, CEO Gil Penchina warned me it wouldn’t be a great product at launch. It’s simply a proof of concept of what can be created using open source software and little money, he says. Fair enough. But it’s time for Wales to be quiet, let this thing evolve or not, and eventually let the software do the talking. Eventually Wikia will make the index available to third parties. But the index needs to be reasonably decent before anyone will want it. Wikia has a long way to go to get there.

Update: Good debate in the comments below, including a couple from Jimmy Wales.

The long wait for Wikia Search is over with the alpha version of the service now live.

As we wrote December 23, Wikia’s Jimmy Wales is pitching the search engine as a Google competitor; its a strong statement to make but does Wikia Search Alpha take the good fight up to Google? Not yet.

I ran a couple of different search terms into Wikia with mixed results. A search for TechCrunch for example shows TechCrunch.com as the first result, but type in Michael Arrington and the only TechCrunch result is at 7, and it’s TechCrunch France aside from something called a “mini-article”. I tried a couple of other people in Wikia Search as well, the service has no issues with finding sites/ companies by name, but really struggles when you’re searching for people.

The user interface is clean and uncluttered, but the choice of font colors has a lot to be desired.

Ultimately you can judge for yourself at alpha.search.wikia.com

Update: as pointed out in the comments, Wikia Search also doesn’t suggest results based on incorrect spellings, for example Wikiq doesn’t recommend Wikia, either in the results itself or with a “did you mean Wikia” line as Google would do.

Update 2: see Michael’s thoughts here.

Well, the waiting appears to be over, and the promised 2007 launch date was technically achieved. Wikipedia/Wikia Founder Jimmy Wales has publicly announced the private beta for Wikia Search – right now. And the public launch is set for January 7. In a note to the Wikia Search email list a few minutes ago, he wrote:

From: jwales@xxxxx.com
Subject: [Search-l] private pre-alpha invites available
Date: December 23, 2007 7:04:01 PM PST
To: search-l@wikia.com
Reply-To: search-l@wikia.com

Ping me if you want one…. we’re launched.

I’m going to be letting people in slowly over the next few days and we
are aiming for a January 7th public launch. We want to run over the
system with help from people to complain about what is broken…

Best way to ask is by email, but please don’t be offended if I don’t
answer right away. I am expecting a bit of a flood here.

–Jimbo
_______________________________________________
Search-l mailing list
Search-l@wikia.com
http://lists.wikia.com/mailman/listinfo/search-l
Change options or unsubscribe: http://lists.wikia.com/mailman/options/search-l

I spoke with Wikia CEO Gil Penchina on the the rules around the beta – users are being asked politely to withhold posting any information about the beta until the public launch on January 7. Hopefully people will respect that – there are bound to be some major hiccups and Wikia deserves a chance to iron those out before what is sure to be a ton of attention on the product.

We’ve waited more than a year for Wikia to launch their human powered search engine. The project was first announced in December last year by Wikipedia/Wikia founder Jimmy Wales. The promise was to return better results than Google and other search engines, using humans to make quality decisions:

“Google is very good at many types of search, but in many instances it produces nothing but spam and useless crap. Try searching for the term ‘Tampa hotels’, for example, and you will not get any useful results…Essentially, if you consider one of the basic tasks of a search engine, it is to make a decision: ‘this page is good, this page sucks.’ Computers are notoriously bad at making such judgments, so algorithmic search has to go about it in a roundabout way…But we have a really great method for doing that ourselves. We just look at the page. It usually only takes a second to figure out if the page is good, so the key here is building a community of trust that can do that.”

A lot has happened since that announcement. Mahalo, a Sequoia backed startup with their own approach to human powered search results, launched in May and is showing promising early growth. Meanwhile Google, perhaps somewhat annoyed by Wikia Search as well as Wikipedia’s ongoing refusal to add Google ads to their pages, announced Knol earlier this month – clearly a shot across the Wikipedia bow.

Not much on Wikia search, however. They’ve set up a page to discuss the project. In July Wikia announced the acquisition of Grub, which had technology to allow distributed web crawling by users. And an early screen shot, showing a Facebook-like profile page, was shown in South Africa in November.

Wikia Search In 2007 Or Not? Jimmy Wales Say Yes.

But the promise has been to launch Wikia Search this year, and time is fast running out. There’s just one week left in 2007.

Today a report was published that Wales, in an IRC chat, promised to make the end-of-year launch date: “ the search engine *will* launch before the end of the year, probably in private beta first, and then open to the public in early january. No specific dates are certain yet. But sooon.”

I asked Wikia CEO Gil Penchina if the quote was accurate and whether to expect a launch in the next few days. His response was “Can’t comment on exact timing.”

It won’t be important a year from now if Wikia Search launches this year or early next year. But it is time for the product to be judged on the merits of the search results created by it, not on a series of press leaks and hazy screen shots. I look forward to the launch, whether it be this year or (hopefully at the latest) next.

News of Google most recent project, Knol, came out late last Thursday without, as far as I can tell, much in the way of press pre-briefings. All the major publications were late to the story. Blogs hit it fast, but had nothing to go on other than the brief blog post put up by Google’s Udi Manber announcing the project. Our initial story on Knol is here.

From a product perspective, Knol is not much different than existing products like Squidoo and Hubpages. It’s a new knowledge base for authors. Anyone, eventually, will be able to write on any topic they choose. Google will provide authoring tools, store the information, allow others to comment and suggest edits, add ads with the author’s approval, and provide traffic via their search engine.

But Knol isn’t really aimed at Squidoo and Hubpages. It’s much more likely that Google is jealously eyeing the massive traffic that flows through its search engine to Wikipedia. As Nick Carr has noted, Wikipedia continues to climb and climb in search results for many top search terms.

More Ad Inventory Needed, ASAP

Wikipedia, a non-profit, has stubbornly resisted any efforts to monetize its pages. Google would kill to supply ads to Wikipedia. Barring that, competing with them makes a lot of sense.

Google needs to grow revenue to support their valuation. And for that, they need ad inventory. It wasn’t surprising when Google started hosting news directly and allowing comments (that = page views). So the idea of them hosting a knowledge base shouldn’t be surprising, either.

Authors have a choice – they can have ads or not. But if they have ads, they can only choose Google. Many authors are going to include ads, and Google will get extra inventory.

Delicious Timing

Wikipedia has caused more problems than just refusing to take Google’s ads. They are also launching a much anticipated search engine this month via their for-profit arm, Wikia (Wikipedia founder Jimmy Wales hates it when Wikia is called that, but it’s damned hard to tell where Wikipedia ends and Wikia begins sometimes). Google isn’t likely to be particularly scared of Wikia’s new search engine, but it has probably been a little annoying for them to watch all the press about the upcoming “Google Killer.”

Google doesn’t usually pre-announce products before launch. in this case they did. Why? Perhaps as a reminder to Wikipedia that competition can flow both ways.

Anyone Remember Google Base?

As a content management system, Knol is a kissing cousin to Google Base, a classified ad platform that Google launched in late 2005. Google Base has gone exactly nowhere – if anything it’s a spam farm and nothing more. But at the time of its launch the New York Times and others heralded it as a major disruptive force to the classified ads world. Knol may be Google Base with a little more strategic thought applied.

No Conflicts Here, Move Along

Google says that Knol pages will be indexed into their search engine but will have no special ranking. That’s a little bit untrue, since they’ll be hosted by Google and will have the advantage of Google’s hefty PageRank to lift them in search results. And since no one will be auditing Google to ensure that Knol pages are treated just like everyone else, there are bound to be claims of conflict of interest. The fact is, Google will make money from Knol, and so they’ll have a financial interest in moving people to those pages. That makes them less believable in the role of a neutral gatekeeper.

Google is now synonymous with search. Offering Google Knol and putting it in the search results is analogous to Microsoft offering Office for the Windows platform. Sure, anyone can compete with Office, but Microsoft has a natural advantage and finds ways of keeping market share. The Knol team will likely do the same over time.

Wikipedia v. Knol

Wikipedia gets massive support from the community because it’s non-profit. Google can’t compete with that, so they’re focusing on putting the authors’ names in lights and giving them a little cash on the side, too. That should help them pull some heavy Wikipedia contributors over to their project.

Very soon we are going to see a lot of Wikipedia content moving wholesale to Knol. Wikipedia content is basically free to use, redistribute, copy, whatever, under the GNU license:

All text in Wikipedia is covered by GNU Free Documentation License (GFDL), a copyleft license permitting the redistribution, creation of derivative works, and commercial use of content while authors retain copyright of their work.

Anyone writing for Knol is likely to at least peruse Wikipedia content before publishing. And if they see anything good, they are at liberty to simply lift and copy it over to Knol, and get a adsense check for their time.

So, in a way, Google has found a way to monetize Wikipedia content after all.

In a poll on Friday, TechCrunch readers narrowly said Google hasn’t overstepped its boundaries with Knol. If Knol is a success, those results may be a little different a year from now. In fact, the more successful Knol is, the more uncomfortable people are going to be with Google as gatekeeper and content provider.

New wiki-based search engine Mahalo is launching social networking features today at the LeWeb3 conference in Paris.

Mahalo is a search engine that focuses on user link submissions and an editorial process to theoretically produce better search results than algorithm-only engines like Google. It first launched in May 2007.

The company already pays users for quality submissions. Today, they are adding user profiles and other social networking features to further incentivize users to submit quality content.

Editors decide if links submitted for a particular search term should be inlcuded in search results. If a submission is accepted, the user gets credit and a higher score. If it’s banned, the user’s score takes a hit. All of the results are shown on a user profile (click image above to see larger version), so heavy users will be inclined to add new links carefully and increase their score. The service has around 26,000 pages of search results, and each one represents 10-30 different search queries. If a search returns no Mahalo pages, results from other search engines are shown. 1,000 or so new pages are created each week on the site, says CEO Jason Calacanis.

Mahalo has direct competition coming from Wikia. Based on early screen shots that show user profiles, Wikia seems to be taking a nearly identical approach to maximizing user participation.

Mahalo is doing very well based on early Comscore statistics, which report 2 million monthly page views and 874,000 unique visitors. More importantly, the trend is clearly towards fast growth. Compete statistics agree.

Mahalo has raised around $20 million on two rounds of financing. Another rumor says their last round was valued at over $100 million.

It was eleven months ago that Wikipedia founder Jimmy Wales first mentioned his vision for a people-powered search engine that would eventually launch under his for profit startup, Wikia.

Not much has happened since then, other than a lot of chatter on an email discussion list, and the small acquisition of Grub, a distributed web crawling company, from Looksmart. The official site for Wikia Search is here.

But the promise has been for Wikia Search to launch this year, and it appears to be on track. Yesterday Matthew Buckland reported that Wales showed “some of the first screen shots” of the new project (the first, as far as I know).

The main screen shot is a profile page for a user (see above) that looks surprisingly like a Facebook profile. It was taken by Nic Haralambous.

The Man v. Machine debate as it applies to search is about to begin. By this time next year we should have lots of data on the performance of Wikia Search, as well as the new startup Mahalo which is also in this space. Until then, we can spend our time speculating and, I guess, continuing to live with Google for our search needs.

Seattle-based Wetpaint, which launched in June 2006, is a hosted wiki site that focuses on great looking sites and making the user interface as easy as possible. A number of wikis have popped up around popular pop culture stuff, as well as more private sites.

Tonight they added new feature that should generate a lot of page views – they have fully integrated a forum/message board into every wiki.

This isn’t Tangler-level forums (which we consider to be the bleeding edge), but they’ve put a lot of thought into the feature set around these message boards. Posts can be tagged, the view expanded/contracted, there are email notifications of new messages, and the search feature works well. Any forum thread can also be turned into a wiki with a couple of clicks.

CEO Ben Elowitz says the two products go together well – wikis are great for evergreen content but don’t allow for good conversation. Forums allow great conversation but aren’t great for new readers. The hope is that by combining them they’ll allow for better content for all users. And in the process get a lot of page views.

Other startups innovating in the forum space (besides Tangler, mentioned above) are Meetro and Grouply.

The hosted wiki space is crowded, and Wetpaint competes with Wikia and PBWiki, among others. Comscore shows Wikia in the lead with over 3 million monthly uniques, followed by Wetpaint with 1.3 million and PBWiki with 770k (Wikipedia, of course, is the 800 pound gorilla, with 228 million unique monthly visitors):