218 comments

[ 2.9 ms ] story [ 270 ms ] thread
(comment deleted)
That's a fantastic looking curve, but keep in mind the absolute numbers shown. They are at 1.6 million direct queries per day. Still a ways to go to be truly competitive (3 billion per day for Google). Still, if they keep up with current growth rates, they could do some real damage.
Just one at a time is all it takes. :) Replace starfish with searches..and bend your imagination a bit.

"A man was walking along a deserted beach at sunset. As he walked he could see a young boy in the distance, as he drew nearer he noticed that the boy kept bending down, picking something up and throwing it into the water. Time and again he kept hurling things into the ocean.

As the man approached even closer, he was able to see that the boy was picking up starfish that had been washed up on the beach and, one at a time he was throwing them back into the water.

The man asked the boy what he was doing, the boy replied,"I am throwing these washed up starfish back into the ocean, or else they will die through lack of oxygen. "But", said the man, "You can't possibly save them all, there are thousands on this beach, and this must be happening on hundreds of beaches along the coast. You can't possibly make a difference." The boy looked down, frowning for a moment; then bent down to pick up another starfish, smiling as he threw it back into the sea. He replied,

"I made a huge difference to that one!""

I certainly felt some difficulty first in setting Duck Duck Go as my default search engine. Google trains you repeatedly, implicitly, over the years on how to optimise your query for Google. I often still find myself adding "!g" to my DDG query to perform the search on Google, but at least Google isn't my first point of contact any more.
if i get that right, people are doing a lot more API calls than direct searches... Which begs the questions of (i) who they are, (ii) what they're doing, and (iii) how this can be monetized?
The graph is for direct searches though. You can see API calls stable at around 9M for quite a while, and direct searches have sky-rocketed.

You're right that there are a lot more API calls still (~x6). But the gap is closing quickly.

Maybe these numbers include search requests from browser-search fields and similar. Otherwise I would be very interested in those sources, too.
I use duckduckgo as my default search engine now. If i need some more info i just use g!. Really like the bang syntax!

And its doesnt track all my stuff, and this stupid google+ (failure) is not integrated. I just want to search stuff. Google did it right in the past, now duck duck go is my uberlord.

Wow, thats a great curve. Just wanted to throw in my personal experience with DuckDuckGo, and it doesn't necessarily reflect on DuckDuckGo specifically, however:

After seeing them pop up here once or twice quite recently I thought I'd give it a go, I opened up preferences in Chrome and switched my default search over to DDG. It felt good, made a nice change but was certainly a bit "odd". I found what I wanted, most of the time albeit it with a slight drop in quality vs Google.

But my biggest gripe with them was confidence. I'd just started out at a new job as a programmer and as you can imagine, I was searching for a good few things, usually things I was struggling with and I just didn't feel sure that I was getting the best results I possibly could be. You might think thats totally crazy, but when your battling an issue as many of you know, you want more than anything to work out whats going on. I didn't have any margin for error, I wanted the best results right there that second.

Specifics aside, knowing that Google is far superior in their results makes it real difficult to use another search engine really, extra features (!so etc) or no extra features. I personally search because I need to "find" something and I usually don't know where that something is, opening up Stack Overflow isn't that much of a chore for me. Its the other bit I need help with.

Anyway, I commend their mission and hope they succeed in taking a fair slice of the market. I think he's a great entrepreneur and I wish him the best of luck. I can't imagine what it must be like to be head on with... Google.

I generally stick to Google for programming related searches and use DDG for the rest. Hopefully DDG improves to the point where I can use it 100% of the time.
Interestingly I switched to DDG because I was having to quote too many programming terms in Google. All of them sometimes, just to stop Google from autocorrecting. I still have 'g' wired as my omnibar shortcut for Google though as I don't always find what I'm looking for on DDG.
'!g' in DDG redirects you straight to Google.
If it were possible to take Bing’s search results, throw a handful of people at them and match the quality of Google’s results, Google would not be where it is today.

That is the main problem I see with Duck Duck Go. The quality of its search results depends on Bing. Duck Duck Go can do things to offer something better than Bing, but not that much better that matches Google’s quality.

the zero click content is often helpful to me, and is certainly a significant upgrade for me over basic bing results.
Result quality is only one part of the equation. Features is another, which is where DDG seems to be making a land-grab for. Users seem to be responding to this approach.
Interesting.

I've also been demo'ing DDG the last week. As an ahem experienced developer, I've found that most of the time, the results I get from DDG are pretty close to what I'd get with Google, minus Google's interface.

I've built up enough confidence to feel that DDG is going to give me what I need.

I do wish it were a little faster though. Speed is a feature, and I'm sure it's hard to compete with google.

Simple results are better than Google - including the summary of the term grepped from wiki. Complex or rare queries aren't as good.

I use it at home - it's quicker to find out when the new South Park episode is out. But it's less use at work for finding the solution to that weird problem

Ad confidence; my default search engine is Google's "Feeling Lucky". Many repetitive searches (wiki, imdb, documentation) are so accurate I can skip the step of clicking on results. But most of the time I use keyword searches anyway. With google, images, wolframalpha, tineye (because google image search is still blocked using user agent sniffing), torrentz, public transit, and a few more (including stuff like open a subreddit, so not really a search).
I find that I am slowly losing confidence in Google. What I often do is input searches in the form: "general term", "specific term", then click on a link and immediately do a find for the specific term I searched for. In the last few months, the specific term I searched for often does not exist on the linked site at all, which I find quite annoying.

I tried bing a couple of times, but the results are even worse.

Maybe Google verbatim mode will fix the problem? I've set up a quick search recently, but have yet to start using it.

Would love to see some example queries to debug if you have them handy. (You can check your search history https://www.google.com/history/ to jog your memory if you have it turned on.)
Here's one: python list popleft

Third result is: http://docs.python.org/library/stdtypes.html

6th result is: http://stackoverflow.com/questions/4426663/how-do-i-remove-t...

Neither one of those contains "popleft".

I don't know - if I had to guess what the query 'python list popleft' meant, it would be 'how to pop the leftmost element from a list', even though the answer does not contain the word 'popleft'.
Try putting quotes around popleft in your query. That works in this case to get rid of the results that don't specifically include the word.
In this case it does, but very often doesn't. It seems to me that this happens mostly with rapidly changing websites, as if the cache showed a somewhat later state than when it was indexed.

And this has been an annoyance since years.

I know what you mean.

These days, I've been searching for things about Flask with Heroku. So I tend to search for something like "python flask heroku sqlite", and often times, I get results that are not specific to Flask. (NB: I don't know if this example works but you get the idea)

I hear a lot of people say that google search results are worse than they used to be - personally I have a hard time telling. What's interesting, though, is that nobody is doing search better, which suggests that it a very tough problem to solve, and even with all the resources Google has, they aren't doing it as good as it could be.

So my question is - what is google doing wrong?

Has their search algorithm become worse?

Has the web become more complex?

Have people just become lazy and feel like they should never have to dig beyond the first 10 results to find what they're looking for?

My take. Disclaimer: I'm a fan for a while now (lesson to startups: If you reply to feedback quickly and friendly, as Gabriel did in Aug 2010 when I submitted a tiny useless bit of feedback to the site, you'll be hard to remove from my list of 'great things').

That said, I do understand the confidence issue every now and then. For me it seems to arise mostly when DDG only shows half a page or less of results. Those are usually good. But I'm used to all the bullshit Google puts above and below that and sometimes have the feeling of missing out.

A quick check every now and then by now convinced me that I'm just conditioned to expect quantity. Which needs to be corrected.

You can use the DDG '!g' prefix for that.

For queries which I'd like to double check with Google, I resubmit it to duckduckgo with '!g' prefix. It redirects to Google.

It's a smart move by DDG to help people resist the temptation to make Goog their default search engine.

I removed Goog from my list of Firefox search engines a few weeks ago.

I have DDG set as my default, but I admit that that's more or less despite the quality of the search results.

When I'm searching for "discovery" purposes, such as when trying to diagnose an error message, I add a !g. When I want an image, !gi. If I want to look up some library, I use !clojure, !python, !php, etc. The only time I let it fall to DDG is when I'm looking for some specific thing that I don't know the URL for, like "American Airlines", or something that I know has a lot of results of about equivalent quality, like "Pulled Pork Recipe."

I find I'm effectively using DDG as a search routing service. I really enjoy that, and I miss it when I use a computer that's not set up with chrome and ddg as the search default, but I feel like they're missing out on some ad revenue here.

> Specifics aside, knowing that Google is far superior in their results makes it real difficult to use another search engine really, extra features (!so etc) or no extra features.

Just so you know, you can use keyword searches to get this same functionality directly in your browser for any search engine of your choice.

"Specifics aside, knowing that Google is far superior in their results makes it real difficult to use another search engine"

Except that Google is far from delivering superior results. In fact, many of their results are full of spam links and garbage.

One of the main reasons I use DDG over Google is that DDG often delivers better results.

Another major reason is DDG's promises of no-tracking and privacy (which I hope can be verified some day). That alone sets them far above and beyond Google, which is spyware incarnate.

Google is now basically in the exact same position that AltaVista was in when Google came along and ate their lunch. Only Google is much more dominant and wealthier than AltaVista ever was (not to mention their legendary ability to attract and keep talent), so they'll no doubt hang on much longer. But DDG and other search engines are still primed to take a bite at those parts of the business where Google has drop the ball -- like privacy, literal search, and search quality.

People care about privacy and not being tracked. I predict Google's arms race with Facebook to profit more from user's data will augment this growth. And/or dampen that arms race.

That's why I search with DuckDuckGo by default (previously Scroogle).

With everyone's concern with privacy, I'm surprised Ask Eraser isn't mentioned more frequently (http://sp.ask.com/en/docs/about/askeraser.shtml).
I used to use Ask until they made an announcement that they were leaving the search business and staying in the toolbar sector. I'm surprised they're still up.
That's why I switched from Google, the !bang feature is something I could not live without now.
Incidentally, I switched my default search to ddg 2 days ago. Use !g a lot but also ddg native frequently. Starting to get a feel for what type of search will benefit more from DDGs extras.
pg's analysis was spot on, then [1]. I recently realized DuckDuckGo would be a hit when a hacker friend who does not read HN started using DDG as his primary search engine.

To me, this means it is out there. I switched too, at least for a while, and now I use both google and DDG, constantly having the impression that DDG's results are getting better and better.

[1] http://paulgraham.com/ambitious.html

I remember specifically thinking of DDG when PG mentioned search engines in that keynote. I've used DDG since 2010 and I would never switch back.
Although I don't use DDG, I too thought of them specifically when PG mentioned it.

Let's see if PG turns out to be right on this one.

DDG used to be pushed heavily on reddit. Not in a bad way.
!g is the google killer. I switched to DDG, but kept drifting back to see if Google was doing better. I find myself using !g less and less, as I get used to the result format.

At the moment, DDG reminds me of Firefox (or was it WaterSquirrel then?) back-in-the-day. Its something that "feels" better than what is the standard - and it's something that I go on and on about to other nerds, when I'm drunk ;)

> (or was it WaterSquirrel then?)

Phoenix -> FireBird -> FireFox :)

He's probably thinking of "IceWeasel" which was a rebranded firefox a linux distro (Fedora maybe?) was shipping due to some sort of copyright/licensing quirk.
Debian actually. It's now GNU IceCat though. The reason they changed the name is because Debian backports some security features to Firefox, and Mozilla denies the use of the Firefox branding to unofficial builds of the software.
(comment deleted)
I hate this (!g) though.

If I want Google results, it's a heck of a lot faster, easier, and natural to just Google it instead of adding some kind of search operator BS to my query. I used DDG for a while (a long time ago, admittedly) and basically had to !g every query I ran in order to get acceptable results. In the end, I gave up because it was just adding typing and thinking time to search.

I suppose one could argue that the interface lacks some of the annoyances of Google, but I feel like quality and speed are the most desirable traits of a search engine (in that order). If we're just using somebody else's results anyway then why should we accept a slowdown?

I don't keep statistics, but my perception is I have to click on one of the 'Try Xxx/Yyy' links maybe 1 in 10 times.

I don't mind since it saves me from having to go into "privacy defensive" mode to interact with the big G.

I like the !bang operators because I can get results from Google for specialty searches like maps (!m) and images (!i) right from the address bar in Chrome, but still keep DDG as my primary search engine. Google's results are slightly better, but it hasn't been a big deal for me. I still end up finding what I am looking for.
Yes, the prefix is easier if you search from the firefox search box (ctrl-k).

Also I frequently find nowadays that Goog and DDG provide very very very similar results for most queries for which I dissatisfied enough with DDG to try Goog.

I get into using DDG sometimes, then run into a batch of wanting to use !g and just switch the engine in FF and forget to come back until I read some story on HN ;)

It kind of irks me that DDG is still asking if I meant [similar sounding children's book to my business with 3 different letters] when !g stopped doing that months ago. It also has my psycho competitor's defamatory website ranking #8 when this page should have crappy SEO with google.

I've been using DDG for quite a few months now and in general it works well. However I still find myself adding that "!g" at times when they find no results, and seeing what I was looking for at Google. I am also a bit disappointed with the "I'm feeling ducky"-results which are rarely what a person is likely to be looking for as soon as you get more complex than one word.

But that's the negative parts, and otherwise the positive outweighs those by far. I love the bang-syntax and I found myself using their information-box directly on the search page very often, mostly for converting currency.

I think better results will come with time and increasing usage, and I wish them good luck in the future.

It's interesting to see how many requests are coming from the API, offhand I can't think of any uses that I've come into contact with. Does anyone know who is using the API?
DuckDuckGo is definitely pulling users from Google. By mimicking Facebook, Google has lost its mojo. Stepped right into the "social" honeypot. No news. The more interesting thing is the fantastic feeling more and more people have: that Facebook and Google will take eachother down the spiral.

As for the Google part, right now it's probably just an alarmist prediction, but it's going to be fun to watch un/fold. Especially if DuckDuckGo and other engines continue to get better while practicing their principles instead of a crude mixture of addictive search quality, addictedness to bucks and PR along the lines of "Don't be evil".

[1998]

Google is definitely pulling users from Yahoo. By mimicking Alta Vista, Yahoo has lost its mojo. Stepped right into the "portal" honeypot. No news. The more interesting thing is the fantastic feeling more and more people have: that Alta Vista and Yahoo will take eachother down the spiral.

As for the Yahoo part, right now it's probably just an alarmist prediction, but it's going to be fun to watch un/fold. Especially if Google and other engines continue to get better while practicing their principles instead of a crude mixture of addictive search quality, addictedness to bucks and PR along the lines of "Yahhooooo!".

Haha. Reminds of the "The social graph is neither" article [1].

"Right now the social networking sites occupy a similar position to CompuServe, Prodigy, or AOL in the mid 90's. At that time each company was trying to figure out how to become a mass-market gateway to the Internet."

[1] http://blog.pinboard.in/2011/11/the_social_graph_is_neither/

One difference in this story is....

Google was vertically integrated from the start: They sent out their own bots to scan the internet, they had their own reverse index, and created their own algorithm.

I'm not sure about DDGs backend---but if people use the g! tag, its almost as if DDG is simple chrome for Google.

Maybe that'll be the future: Google will become the maker of engines, and let everyone else make the car.

As far as I know, DDG uses both data from other search engines (Google, Bing) as well as its own crawling bots. If they only used their own crawlers, with the current size of the internet it would take a long time for them to get a good amount of data.
I swapped to DDG on all of my browsers about a month ago and I've been really pleased. Like others, I find myself occasionally using !g, particularly for programming queries. Things like !so, !mdn and !msdn actually help out a lot in some of these cases, which makes me think that as I learn the bangs I'll rely on Google less and less.
Nice. If nothing else this proves that there is unmet demand in search.
It was mentioned before, but I think it's more a kind of "different" and unpersonalized search. Like proposed by Paul Graham: http://paulgraham.com/ambitious.html
I agree. My point is simply that this is strong evidence that Google's 'one size fits all' approach leaves room for other entrants.
I've gotta say, Gabriel's tenacity is both admirable and inspirational. I remember long ago in its early days I tried it and shrugged it off as just another search engine, destined for failure after a few months of obscurity and being largely unknown.

It's now my default browser, and has been as of about 2 weeks ago. Gabriel has been very keen on picking up where other search site are slacking.

That's awesome! Just goes to show that search is still an area for growth.
as a happy DDG user I have to say congrats and I absolutely love the zero click box, a feature that often provides superior search quality than Google.
I've been using blekko.com and found that I prefer it to ddg. Like others have, I too go back to Google for programming related searches at times.
That's consistent with our Google Analytics. Visits from DDG are up to 58 last week, up from 15 the first week in January with a steady climb. There was a dip last week over week for some reason.

Compared to Google's 38K visits for last week, they have a long way to go to make a dent. We perform better in DDG results, as well.

Do they publish revenue? I wonder what the infrastructure costs are and how they scale in comparison to the number of request.
I think that most of their results come from Bing if I remember correctly. So the scaling doesn't have to be as heavy as it might seem. But I may be wrong.
As with most new search startups, people will give it a chance. Hence the increase in traffic. I tried it a couple of times but, as most people do, I then checked Google just to make sure I'd really covered what is considered the full spectrum of search.