Tattletech Hot Seat Lucien Burm, Kimengi
Last Friday, Tattletech went to the office opening of Kimengi where the CEO Lucien Burm was hosting an informal chat and cocktails on what Kimengi will bring to the blog sphere.
Tattletech: What does recommendation technology mean?
Lucien Burm: In our view, recommendation technology must deliver a technically working virtual model for how people give and receive recommendations, how they handle them and try to improve that proces by using the internet.
The ‘engine’ we are developing creates something of a dynamic blueprint for how masses of people share and interact on suggestions with people they know or encounter. The big advantage of the virtual world is, that we will be able to also leverage on suggestions, opinions and interests of people you don’t know, and there are many of them. So people’s virtual components (avatars) can create far more interactions with each other (and not limited by time) than our real-life parts can. We are going to use this fact to create better recommendations.
In short, recommendation technology needs to become much more personal, more realtime with more detail for more content. These factors are still a bit opposed to each other, due to the current state of technology. It is a real challenge to combine all four factors.
At Kimengi we have defined 5 requirements for a next-gen recommendation service. It takes little to much blog space to handle in full here, but in short they are: automation, adaption, induction, inspiration and transparency. We will say at least something about the last one in the answer to your last question.
All our ideas and models for recommendation technology are based on the following question: If your brain could handle all the information now available online in single moment, what would you want to do next? We basically try to help you get as close as possible to that level.
TT: Would you consider this technology a step towards semantic web?
LB: It is more like the other way around. The semantic web could help us a step further. We are not trying to build the semantic web, but better understanding of content and context will help greatly to determine people’s real interests. And that is the business we are in. We use semantic web technologies as an input to recommendation engines, but its the core of that engine that we are interested in. The semantic web gives us a better understanding of content and a better response to requests, which is a true step forward in the web. But the flood of information does not diminish because of it. We need a different kind of technology to help us with that. We hope our technology provides a shift the growing need for information filtering.
TT: What are, in your eyes, the most compelling features blogs can implement in order to attract more readers and followers?
LB: Well, it is all in the content of course. If you write something interesting, you’ll become interesting and people will visit your blog. The speed of your content travelling through on- and offline networks determine how many people you reach and how fast. We believe there are basically two types of networks: One type is more social driven and the other type more content driven. With more social driven we mean working with emails, social bookmarking, social networking and microblogging. Of these networks email may be the slowest and microblogging the fastest to spread your content. With more content driven networks we mean search engines, trackbacks, comments, syndication, referrals, etc.
As a strategy we think that your content could trickle down from microblogging through social networks and bookmarks to the level of content driven traffic generators. So first, you spread your content out, next you let it register deep. That could create a sustainable stream of traffic for a while. We see some blogs playing that strategy quite well. We don’t think there’s one way or feature, but how you play a combination of instruments.
There are tools available for your blog that use both types of networks. They all work if you put enough effort in it, though sometimes we think blogs are a bit crowded with tools.
With f»dforward we combine these social and content networks into one system, a recommendation network, that should create instant spreading of your content to the right people and blogs. There is no intermediary channel, like portals, search engines or bookmark sites, it works instantly from blog to blog and from reader to reader.
We once said: Content may be king, but users make the democracy. (Tattletech note: we love this philosophy) Blog well, and your readers will bring you forward. With f»dforward, we try to fuel that.
TT: Does recommendation technology offering differ when implemented by bloggers as opposed to using it for corporate blogs?
LB: Yes, it probably will, because the reasons for blogging tend to differ. Though big blogs make money as any other company, the relation to products is quite indirect while at corporate blogs the relation is more direct. Still, sometimes it can be useful to be directed to product blogs. This is where the readers come in. I can imagine that if a product blog is bluntly promoting his product instead of providing some interesting company or product insight, readers will move away from it. A good recommendation service will be able to use that immediately and adjust the recommendation for this subject, whether it be a good product blog or bad personal blog.
Still, we wouldn’t be surprised if we will offer a different kind of widget for company blogs, because they could be interested in different services than personal blogs, e.g. integration with the company website as well or options to provide product information.
TT: We are curious to know, does Kimengi have a special meaning?
LB: Yes it does, though we sometimes joke that it was the only word left which was still available as a domain. We started experimenting on the subject of recommendation more 13 years ago. At that time, people were talking about matching instead of recommendation, but it is basically the same of course. At that time we already found a direction in which we would like to develop and we came up with Kimengi as a name for it. Now that we are finally here as a company, we just thought it cool to keep the name.
All we can say is that the ‘engi’ part is from ‘engine’. You will have to figure out the rest for yourselves. It is our best kept secret. it is our pagerank, our coca-cola recipe 
Maybe we should release the full name when we release the engine, which by the way is not yet implemented in f»dforward. The current network works on currently available recommendation technologies with some twists to make it ready for our own engine.
After registering we discovered that in fact there are people with Kimengi as a last name too. So, maybe we should us a disclaimer: “Any resemblance to real persons, living or dead is purely coincidental.” 
TT: Following the huge upheaval after Facebook tried to change its terms of service, how does Kimengi handle user privacy, do you collect any data on users or users profiles?
LB: Privacy or rather ‘privacy control and transparency” are at the roots of f»dforward. We serve both blogs and their readers by being glue and a firewall at the same time. We want to provide readers with content they like and blogs readers they can serve their content to. This is why we the use of f»dforward must be anonymous. That starts with never having to register. We don’t have logins or passwords. We use your blog or homepage as an ID that you can use to administer your account on the network.
Next, we never store anything that you don’t want us to. As a standard, f»dforward registers complete anonymous sessions between opening and closing your browser and only in our network of course. If you like, we can user more sessions to provide you with better recommendations, but that is up to you. In the help part of the widget, anyone can change these kind of settings to their wishes. The help page also provides full insight into all the information we registered during your session.
We would like to win the trust of our users, by giving them complete control of their data whether you delete it, change it or enhance it. You know what we know. By gaining that trust we also hope to provide future services in dataportability as a trusted source in social intermediary. Thanks for having us on your blog, Jennifer!
– SM and JLH
