Entrevista publicada no IDG International em 12 de novembro de 2010
Duckduckgo could be just another search engine which displays results from other sites, such as Google´s, Bing´s etc. Instead, it concentrates its effort in ranking those results in a different, user friendly and plural way.
As a result, queries about “apple” bring up first a variety of clusters with different meanings of the noun. By selecting which result your query is related to, users are taken to the specific page about the theme they´re search for.
DDG´s UI is very simple to use and, since it has been developed using PERL, the engine is also pretty quick. Using DG you do not have to change pages. The SERP extends itself when users get to its end. But the speed can also be result of a low usage which leaves the bandwith available for a, supposed, low usage.
IDGNOW! Brazil contacted Gabriel Weinberg, an internet startup entrepreneur, to know about this search engine with a rather strange name.
By the way, you do not have to type in all ten letters, DDG is accessible using only dukgo.com on your browsers´ URL bar.
IDGNOW!: At this point, what do you think DDG is really good at?
Gabriel Weinberg: In terms of queries, it’s really good at nouns, i.e. people, places and things. The reason for that is for those types of queries there is usually good Zero-click Info including a topic summary, often an image, and many times related topic and category links for discovery.
More generally, I think we have a compelling UI with less clutter and spam. I think that’s why we’ve gotten traction and have a growing user base who uses it as their primary search engine.
IDGNOW!: In terms of crawlpower how is DDG going on? how wide is its crawl data base?
GB: We’re a hybrid search engine, which means we do our own crawling and indexing but also rely on third party crawling and indexing. I started out before those third party APIs existed, and so was doing everything myself.
But when they came out and after talking to users extensively, I decided to drop most of that and concentrate on our value-ads, which I see as Zero-click Info, less spam and higher relevancy. At this point, we mainly crawl for those purposes only. So while we hit about 100M domains a month, it’s mostly for detecting and removing spam. Most of our deep results are provided by Bing, Boss, and EntireWeb, which we re-rank and merge with our Zero-click info indexes (that also include some links).
IDGNOW!: Your Serps (search engine results page) are very clean and bring up close to zero information on links and snippets, on the other hand, by clicking on similar sites, you´re not taken to another SERP like Google does. Are you still looking for equilibrium in order to balance DDG SERPs for better user´s experience? how is it going so far?
GB: Absolutely, I’m constantly looking for the right balance here. Like Rich from Blekko, I don’t think there is a one-size-fits-all search UI. I’ve tried to concentrate on making results pages more readable and understandable by introduce interface elements like cleaner serps and Zero-click Info.
IDGNOW!: Do you think that this wiki-alike approach to search engines is the very future o SE´s?
GB: I think that there is certainly an increasing place for these sources in SERPs because they are easier to read, provide high(er) relevancy, and have very little spam within them.
IDGNOW!: How did you manage to keep (financially) until now? Who´s supporting DDG?
GB: It’s been self-funded to date. I was lucky enough to sell my previous company.