“A new spam blocker called Defensio is attempting to improve on the standard set by Automattic’s Akismet, the default blocker for WordPress” – reports TechCrunch.
But they are planning to do more than improve… they want to replace Akismet. From Defensio’s FAQ:
- Is Defensio a replacement for Akismet?
Yes… - Will Defensio outperform Akismet?
-We can’t make any promises (sorry), but our early testing suggests that Defensio’s performance is very, very good. As with any adaptive filter, it will only improve with time. - My accuracy is not satisfactory. Why?
-If you have recently installed Defensio, then try to be patient. The filter’s learning algorithms take some time (usually no more than a couple weeks) to become very effective… - Does Defensio work in conjunction with other spam filtering plugins?
Not really. We highly recommend that you use Defensio as your stand alone spam filtering solution.
And therein lies the rub. Other plugins, like Spam Karma 2 may also outperform Akismet at times, but they don’t replace it; instead, they work together quite well. A status change in either takes effect in the other. Now, I don’t know how good Defensio will be, but it takes a huge leap of faith to give up the proven standard (i.e. Akismet) and replace it with something that may need weeks of learning, as we’re warned by the FAQ. They lost me as an early user right away.
The bigger question here, for software startups: should you start challenging the market leader right at launch, or ride with it, grow as part of the “ecosystem”, then perhaps declare you’re better and no longer need them after you got traction?
Related posts: Weblog Tools Collection…
The big question for me is why can’t they make the “learning” a centralized thing? It’s not like people spam differently on different sites… they automate the spam.
I think you should not announce that you are better than the competition or the current standard in the beginning or after. Just do what you do better and people will switch. They will not switch because you claim you are better but they will switch because either someone they trust said it was better or they experienced that it was better. Claiming that you are better than the competition lacks credibility, raises doubt and is in the end a very weak and method of convincing people.
James: To answer your question, “Somebody’s ham is somebody’s spam”. In order to make Defensio effective, it learns and adapts from each user’s behavior. We feel that it’s a much more robust approach.
Another reason that we favor individualized vs collaborative learning is that it would be -very- easy for spammers to confuse/mess up the filter if all the learning was centralized. However, we do have some degree of learning collaboration going on under the hood.
Keep in mind that your performance won’t start at 0%! Typically, people get an accuracy rate of about 95% when they first install Defensio, and it goes up to around 99.5% after we gathered information from your traffic. The length of the training period greatly depends on how much ham/spam you get. If you get a lot, it is very quick (couple of days?).
Hope this helps!
Have you had the chance implement it?
Better still – sidestep the competition by offering something unique. Building a better mousetrap ain’t the way to go.