DailyPerfect: Latest News Aggregator to Attempt Personalization
It's been nearly 4 years since news aggregation site Techmeme (or tech.memeorandum, as it was called back then) launched to the world. Since then it's grown to be the leading aggregator of tech news in the blogosphere. There have been no shortage of pretenders to the throne over the years, particularly from startups hoping to crack the elusive "personalization" nut. What could be better than a personalized, automatically filtered page of news for you to peruse over your coffee each day? However Techmeme founder, Gabe Rivera, has been consistently skeptical of personalized news over the years, claiming that it's too hard a problem.
Well, let's welcome the latest startup to try for a personalized news service: DailyPerfect. This app has been built on the company's "predictive personalization technology" and claims to predict what news a user will want to see simply by analyzing the person's name.
DailyPerfect hails from Estonia and is an incubator project of investment company Ambient Sound Investments (ASI), who we interviewed earlier this week on ReadWriteStart. DailyPerfect uses behavioral targeting to try and predict a user's interests, through what the company says is "an automated semantic analysis of publicly available information on the web." The company is also releasing an API.
Does it Work?
When you first enter the site, you're asked to enter your name into a textbox. Then you sit back and wait for the personalized news to come rolling in, based on your 'digital footprint.'
The topics that DailyPerfect thought I would like initially were a motley bunch. Some were correct, like 'web 2.0' and 'alternative music.' Some were broad enough to have little chance of not being correct, such as 'History' and 'Fiction.' But there were also some perplexing topics presented to me: for example 'Mining' and 'Benelux countries' (Belgium, Luxembourg or The Netherlands). However the site offers the familiar thumbs up or down beside each option, so you can train the system. The thumbs also apply to individual stories.
There are also options to follow people and websites, which is useful in this age of Twitter and blogging. With websites, you can import your OPML file of websites you subscribe to in your RSS Reader of choice. I entered my Google Reader OPML file, however it only seemed to include a random selection of my feeds.
The site is well designed and the stories were fairly relevant to me. However we can safely say that it's no Techmeme challenger. For one thing it doesn't bind the same story from different sources together, which may be Techmeme's enduring killer feature. Anyone can scan Techmeme and quickly find out what the trending stories are, and what sources either originated it or are the most popular links.
DailyPerfect, on the other hand, appears to select just one source for each story - and it's a mystery how that is done. I saw a few links each to Telegraph, Reuters, and Macworld; along with links to a smattering of blogs, including one ReadWriteWeb story. There was even a Techmeme link in there.
Conclusion: Not Perfect, Maybe Useful
I'm unconvinced by the claims of personalization, semantic analysis and other technical fandangery that DailyPerfect made in its PR. Many new web apps make these same claims, but the proof is in the pudding - and as of now I don't see anything particularly special about the content served up by DailyPerfect.
I can't honestly see myself continuing to use DailyPerfect. It's likely to join the long list of web apps I've tried once and then never came back to. Admittedly, that's probably because I'm an information hound that looks for (and needs, for my work) context in my daily news fix. DailyPerfect may well suit casual news readers who don't require a wide choice of news, but simply a well-picked selection of stories. The question is whether those types of readers want an automated solution like DailyPerfect (other options include the well-established Topix, or a site like PopURLs), or whether they want the human curation touch that aggregator news sites like Huffington Post and CNET offer.
News aggregation and filtering is a crowded field, and DailyPerfect is going to need to do more than throw around words like "personalization," "semantic," and "predictive" if they're to survive and thrive.
Share
Facebook
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
addendum: popurls gives you a small block of personalized news that are based on your reading habits.
Hi Richard,
Thanks for taking the time to review our site. I would like to make a few comments - first, it's important to point out the the dailyperfect news site is intended to be a showcase for our predictive targeting technology, and is not the core asset/product of the company. Our business model is predicated on offering our personalization IP as an API-driven service to publishers, ad networks, ecommerce providers and direct marketers. As with all targeting, there should be an expectation that the 'black box' will not give 100% useful results for 100% of all users (as in your case with the news site). However, we have found in our beta tests that for most users, our algorithms have significant predictive value, as measured by news relevancy and the ability to predict ecommerce purchase intentions. In other words, for the majority of users (and our B2B advertising and marketing partners), our technology is very useful.
Best,
Louis Kanganis, CEO - dailyperfect.com
Richard,
It is interesting to see the growing interest in news aggregation. I think it talks to the growing demand/interest in services that cut through the noise and provide people with easier and quicker access to the news they need to know. The same could be said for e-mail and social media activity, which are other elements of the digital tsunami facing many people.
Mark
I was a bit skeptical when first hearing this, but have backed off from that some. Although for personal use I may not use it that much, it could be a good indicator of finding out about different types of trends. Some of the topics I received weren't very relevant while others were relevant, but just broad.
Richard,
Future success in news publishing is largely dependent on the ability to provide content to news audiences that is relevant to them. Personalization technology should not work in isolation from other forms of news discovery, such as editorial or community-based. Rather, all forms of news discovery should work in concert with one another. Publishers should be in tune to the expectations of their audience and calibrate the formula to deliver against those expectations.
Lisa
"The topics that DailyPerfect thought I would like initially were a motley bunch. Some were correct, like 'web 2.0' and 'alternative music.' Some were broad enough to have little chance of not being correct, such as 'History' and 'Fiction.'"
Sounds like Kreskin's old act.
Actually, it's a bold and interesting idea: Take all the little tidbits of personal info we sprinkle around the web and make something personal and useful with it.
It's not perfect yet, but it's a very interesting direction.
Actually, it's a söve bold and interesting idea: Take all the little tidbits of personal söve info we sprinkle around söve the web and make something personal söve and useful with it.
It is interesting to see the growing interest in news aggregation.Thanks a lot.
addendum: popurls gives you a small block of personalized news that are based on your reading habits.