New features for the Mollom module for Drupal

We have just released new versions of the Mollom module for Drupal 6 and Drupal 7. In addition to various bug fixes, as well as usability and API improvements, we have included two new end-user features. First, we've provided the ability to control the strictness of the text analysis. This allows you to control how aggressively Mollom should show CAPTCHAs and block spam. Second, we ported the profanity checking from the Drupal 7 version of the Mollom module to the Drupal 6 version. This means that you will be able to choose to use Mollom to block obscene language in addition to spam. Progress!

On social

I’m sure that no one would disagree that for the last several years, the word most used in Web 2.0 is “social”. Social, as in social networks, social games, social feeds, social bookmarking, social search, social websites, or really, social anything! The thing is, in fact, it took quite a bit of time before companies stopped the talking the social game, and really started walking it. Of course, many large companies have strongly embraced “social”, including Facebook ads, fan pages, apps, Twitter buzz, LinkedIn groups, and others in their marketing mix. Additionally, brand websites and corporate websites are becoming more social as well: visitors can register and create a profile, blogs with commenting plug-ins are implemented, Twitter feeds are added and sometimes these sites feature full blown fora, message boards or even come close to creating proprietary social networks.

This is an interesting trend for both consumers and brands alike. These online channels drive scalable interaction and communication between consumers and brands, in both directions. Companies can more easily identify their so called brand ambassadors and treat them in appropriate ways. Consumers finally see the barriers to interact with huge corporate organizations lowered and are happy to have a channel to ventilate their personal opinion on products or services they use daily. Everybody’s happy. Win-win!

However, all great (and small) evolutions come with negative aspects. Opening up your corporate websites and allowing anyone to share opinions, musings, rumors or just anything they like to share, makes a brand vulnerable. Since any product, any service, or any brand will have both lovers and haters, the discussion between the two camps could be beneficial for the brand, especially if the discussion is steered in a positive way. This is called “brand caring” and it’s something I believe every company should do.

However, the more active your blog, forum or discussion board becomes, the more visitors it attracts, and the more likely it will be that you’ll bump into, well, … spammers. Those “contributors” do not care about your brand, products or services and likely don’t even know on which domain they’re posting comments. Usually they’re just “bots” that identify interesting targets, sometimes create a profile and start spamming your website, ruining the experience for other visitors and damaging your brand reputation. Deleting these spam entries is just too time consuming and too costly – especially when you can allocate your human capital to other projects while letting Mollom’s spam filter do the work for you. Don’t worry about accuracy: for over 99% of all posted content, Mollom makes the right decision (and either classifies the message as spam or ham). So for every site owner thinking about joining the “online” social revolution: don’t hesitate to do so, but be prepared to counter spam attacks if your site is getting more and more popular. Mollom is a tool that helps you do that.

Call for testers for Mollom's hosted moderation interface

As the amount of user-generated content on the web increases exponentially, the need for responsive, swift and effective moderation of that content also increases. Spam is not only annoying to a site’s visitors, it has a substantial influence on both visitor behavior and site reputation.

• A classifieds website containing many fake items for sale, for instance, will cause a potential customer to either refrain from buying an item or from posting a classified ad herself.

• A discussion forum filled with spam directly degrades an otherwise high quality discussion.

• Reputable news sites lose credibility when they do not - or cannot - block spammers from commenting on news items.

• Fans of great performers and artists become increasingly frustrated when their blog posts are lost in an avalanche of spam messages.

To date, there exist no off-the-shelf, useful services that enable moderators to control comments, blogs, forum posts and other user-generated content in an effective and efficient way.

There are no systems available that easily manage and moderate content that is spread across several sites, that switch from one site to another in a fraction of a second using a single backend and with a consolidated user interface.

There is no intelligent service on the market that can accurately scan huge amounts of content and preemptively classify it according to predefined variables, making the work of moderators easier and faster.

Or is there?

• Imagine a publisher of several dozen websites being able to steer its moderation team, track team performance and empower moderators to control content, while that team works more quickly and effectively than before, while using one single backend for all of its sites.

• Imagine a social network empowered to track and manage the reputation of each individual user, from the posting IP address to the quality and nature of the content itself. Imagine if such a network could undertake specific actions on the site’s behalf, whenever necessary.

• What about a discussion forum with hundreds of thousands of topics where moderators have tools that scan and determine potential flashpoints within seconds, allowing moderators to use “just in time,” targeted, control,

• Or, visualize a community site for kids, where all profanity is blocked automatically before being read by children.

Call for testers

Mollom is about to launch a new add-on product in private beta at Mollom. That new product is effectively a "hosted moderation interface". Our goals are to:

  • Provide an optimized and intelligent moderation interface -- sort and bulk moderate comments by spam score, profanity content and more.
  • Make it easier to moderate multiple websites -- moderate all your sites from a single, unified moderation interface.
  • Make it easier to support moderation teams -- create moderation teams, define their workflows and track the performance of individual team members.
  • Provide moderation as a service -- seamlessly outsource the moderation of your site fully or partially to a dedicated team.

The product is a work in progress, but we'll soon be accepting a limited number of private beta test users. If you're interested in being an early beta tester, sign up here.

How does Mollom stand out?

When talking to companies considering implementing Mollom, I often get asked the same questions. Two of the most common are: How is Mollom different from its competitors? How does Mollom stand out? We realize that these are really important questions to ask, and I'm always happy when I can provide an answer. They give me the opportunity to elaborate on Mollom without my rattling being considered "evil sales talk."

Basically, Mollom has two main competitors: Akismet and Defensio. The first is owned by Automattic; Defensio is owned by Websense Inc. We’ll not dive into the technical details, but keep things simple.

The major difference, in my opinion, is that Mollom considers posts as being non-binary. Content is not black or white; there are grey zones that apply. If Mollom is not completely sure that a submitted comment is spam (or completely sure that it's good content), the service runs an additional test by issuing a CAPTCHA challenge. The poster then has to – in a way – identify himself by solving the CAPTCHA and thereby proving to Mollom that she is not a spam bot, but a real person, trying to post a legitimate comment.

Mollom has its own CAPTCHAs, both text and audio based, and updates these regularly to always be one step ahead of the spam bots.

A second difference is the user registration protection that Mollom offers. Avoiding allowing spammers to gain accounts on your site is clearly a first step in order for spam protection to work more efficiently.

Last, but not least, and to me - as a non-technical person - the most convincing argument: Mollom is constantly improving its products and services and is adding new features as we speak. We already support plug-ins for numerous platforms (Drupal, Joomla!, WordPress and many others), we already use reputation tracking, we already have profanity filtering and language detection, … but there is much, much more to come!

We’ll be announcing new releases with some exciting features pretty soon now and we’ll definitely keep you posted!

Mollom's highly-scalable backend infrastructure reviewed

High Scalability, a site dedicated to describing and cataloging successful, highly-scalable, websites, has recently posted a detailed review of Mollom's highly-scalable backend infrastructure.

As a web service protecting close to 40,000 websites, and handling about 100 requests per second, Mollom certainly qualifies as a site with much traffic. Currently, Mollom finds and prevents about half a million spam posts per day, and does a single spam check quickly, with low latency. Every feature we add to Mollom is carefully analyzed for its impact on speed. Personally, I can't wait to see Mollom grow even larger so that our scalability challenges become even more interesting.

The article, written from a series of interviews with Ben and Johan from the Mollom team and then interspersed with the author's own experiences with installing and using the service for himself, is incredibly detailed. It accurately describes much of Mollom's history, the challenges that we faced early on, and how those challenges were overcome. All in all, it provides a great look behind the scenes of the Mollom service. Thanks for the profile, HighScalability!