Comments for New Ask.com UI Gives 20%+ More User Satisfaction

New Ask.com UI Gives 20%+ More User Satisfaction

2007-06-08T21:25:16Z

This week Ask.com released a new user interface, which they nicknamed Ask3D. The vertical layout scheme looks very unusual for a general purpose search engine. However it's aesthetically pleasant. But as well as the visual enhancements, the new vertical scheme has other wisdom behind it. Let me explain with this graph:

]]> In the above examples, I queried my name "Emre Sokullu" on Google and Ask respectively. The results: Google gives me 5 results in the first page, Ask has 7. Assuming that most users don't even scroll down, my satisfaction probability is 40% higher at Ask. You might think that people do scroll down, but past usability experiments have shown that most do not. Or even if they do, they do it very carelessly. Indeed users don't just give up at the first page if they don't feel satisfied, they also are likely to give up at the first eye shot if they don't feel satisfied - this is a natural consequence of increasing user expectations.

But there are a few important parameters here. First of all, Emre Sokullu is a typical long tail query example. As the query approaches the head of the tail, Google and Ask return the same amount of results. But most of the time, while Ask returns 6, Google has 5. Consequently, the user satisfaction is approximately 20% higher. Also note that another important parameter here is a user's screen resolution. The average mainstream screen resolution worldwide is 1024x768.

Overall, a small difference of ~ 80 pixels means a lot. Ask has basically increased the user satisfaction probability by a huge 20%, with just a UI change. This is a very good example of where user interface innovation in search may provide tangible improvements in the search experience.

The F-shape theory

The heatmaps above represent the way users read different web sites. The interesting point here is the commonality of the F-shape reading pattern during the eye tracking studies. More about this interesting work is provided on Jabob Nielsen's site.

This suggests that the current Ask interface has no conflict with the F shape pattern. The extra margin at the left side gets reflexively omitted by users, as shown in the first and second examples above.

Other Enhancements

The enhancements at Ask are not limited to UI. Additionally, in popular queries, the right panel can give you a shortcut definition from Wikipedia and other relevant results from images, videos, blogs and shopping channels. Therefore the chances of getting a relevant result at the first shot increases.

Customization

Last but not least, Ask has taken some courageous customization steps as well. It allows you to have your favourite desktop image in the background each time you visit the site. This is a feature that complements the current social networking and mobile phone customization trends. Just as you enjoy having your own background image on your cell phone, Windows desktop and MySpace page, you can now enjoy having your own background on your search engine page.

All of these fancy enhancements come at a price though - the pages get more and more bloated compared to the simple Google and Yahoo interfaces. Ask's logic is to leverage the advantages of high speed broadband connections and expand in the US market mainly - where page download differences at that level are not even noticeable.

Conclusion

A few months ago, myself and our search czar Charles Knight had a discussion around the topic of UI innovations. My point then was that user interface innovations are not that important - the important thing is to make technological advances. I still advocate this idea and think a new search engine won't ramp up significantly just because it brings a few UI innovations. However it is certainly a good idea to combine technological shifts with user interface innovations; or in some sense, feed technological breakthroughs with smart UI magic. And this is what Ask.com has done.

Comment from Sam on 2007-06-08

2007-06-08T21:56:25Z

Yahoo gives more results than Google does as well, while MS gives slightly less. It is not clear to me what would drive adoption of another search engine at this point. I'm pretty sure that it is not the number of results on the page though.

Comment from David on 2007-06-08

2007-06-08T22:14:34Z

I have enjoyed many Read/WriteWeb articles, but this reads like a poor press release from Ask. Nothing in this article justifies the title of "New Ask.com UI Gives 20%+ More User Satisfaction" ... "Overall, a small difference of ~ 80 pixels means a lot. Ask has basically increased the user satisfaction probability by a huge 20%, with just a UI change." ... c'mon, 80 pixels does not mean a lot.

I'm looking for a "Bury" link on this page, but I guess this comment will have to suffice.

Comment from JabbyPanda on 2007-06-08

2007-06-08T22:46:08Z

Unfortunately for Ask.com, search results at Google are sooo much more relevant in niche searches, I've just undergone through.

More relevant search results == more usability for the end user.

Comment from Gary King on 2007-06-08

2007-06-09T00:11:42Z

Sorry to break it to you, but relevancy != number of results returned. Just sayin'.

Comment from Emre Sokullu on 2007-06-08

2007-06-09T00:43:11Z

Gary and David,

Of course relevancy and number of results are not equal but your chances of getting a relevant results is higher when you see more results at the first look. Do the simple probability math and you'll see.

Press release? Go PRLeap and see what a press releases is. I've read no Ask press release so far but I think Ask haven't published a press release with this "data" yet; this is my own finding. And yes, personally I liked the new interface a lot, found it very useful.

Comment from Terrence Wood on 2007-06-08

2007-06-09T01:11:05Z

Could it be possible that you get 2 extra results (from a single search) because 2 of the results have single line descriptions? 20% "extra results" is completely margin of error territory, all things being equal.

My own non-scientific test returned one less result, and one more advert top-left in the Ask UI. That irk's me, but more inportantly, I would investigated into the reduced input area on the SERP - there appears tp be 20% less space for me to refine my search.

Where is the proof that people don't scroll? Research suggests users scroll long web pages.

And what does eyetracking have to do with it? Eyetracking finds reading english readers tend to read from left to right and start at the top, no surprises there, but surely gaze is driven by design and purpose.

Comment from mario on 2007-06-09

2007-06-09T11:56:42Z

As for result relavancy: GoogleBot and Google UI understand HTTP content negotiation, Ask.com does not. And that's just one out of hundreds of optimizations Google aquired over the years..

Ask.com is another me-too search engine. No technical advantage over Google, and AFAIC not any better user-treatment-wise (tracking cookies until 2036, puuuhlease).

Comment from Charles Knight on 2007-06-09

2007-06-09T12:50:16Z

Charles Knight here: Emre is correct, there is a classic chicken/egg debate over UIs. Do you want a killer engine/crawler/index/algorithm and then a nice UI on top;
(cake and icing?)

or is better to have an awesome interface, with perhaps no crawler at all! I know Alts that have built a cool UI on top of Google's index, *and then* they got to work on their own Index - in that order.

I think Emre would choose the former, and I would choose the latter. Why? Look at www.kartOO.com, www.KoolTorch.com,
and yes, www.msdewey.com (is that Emre gagging?).

Now, go to www.SpaceTime.com and www.EveryScape.com and see if you can put 2 + 2 together (if the light bulb goes off, email me at Charles@ReadWriteWeb.com).

Google has the world's plainest UI on top of the world's greatest engine, and they seem to have done well. :-) The
alts seem to have done the opposite; all 1,000 that we track on www.AltSearchEngines.com barely make up 5% of the search market pie. What's up with that?

Later! Excellent post, Emre, as usual.

Comment from Rishi on 2007-06-09

2007-06-09T23:39:31Z

"Unfortunately for Ask.com, search results at Google are sooo much more relevant in niche searches, I've just undergone through."

This is largely due to the Development of Google Co-Op, which allows Google to take the refinements made on their personal search engines and integrate them into Google's main search engine.

http://www.google.com/coop/

Comment from Emre Sokullu on 2007-06-10

2007-06-10T07:33:25Z

5% ? Are you sure Charles? This a very big number; I guess you count local search engines in, otherwise I don`t think long tail search engine can make so much success - remember Don Dodge`s article - this means $5 B valuation floating in the air :)

Comment from Mathew Patterson on 2007-06-10

2007-06-10T22:51:17Z

It's a big leap to equate two extra results with an automatic increase in user satisfaction - what if they are completely irrelevant results?

What if Google's first result is the one I want and Ask doesn't show it until number 7? I totally agree that it is an interesting idea, and may actually be a user satisfaction improvement, but 20%? - nobody knows yet.

Comment from Kyle on 2007-06-10

2007-06-11T01:59:11Z

The title of this post is misleading. I just did a Google/Ask comparison of a non long-tail search such as Daibetes, where Google easily bests Ask with more results above the fold at 1024x768(thanks in part to Ask allowing more sponsored links at the top). That's including Ask's equilavent of the OneBox.

Furthermore, it's all subjective. To me personally, Ask's new interface is a mess of information heirarchy, where they seemingly couldn't stop throwing more and more information on the page.

My guess is that these improvements will prove popular with some of Ask's current audience, as well as information retrieval geeks, and maybe grab them some new users. However, I think they're stand to lose more novice/average users with the new interface.

Either way, I am/was a big fan Ask until this current redesign (although, I still love the new homepage rebranding). Would have loved to be a fly on the wall in the user research done up to this launch.

-Kyle

Comment from Emre Sokullu on 2007-06-11

2007-06-11T15:12:36Z

Matthew and Kyle,

I agree with your points. What I say here is: at Ask you have 6 results in a single page, at others you have 5 - therefore ***mathematically speaking*** your probability of getting a relevant result is 20% higher - very simple. The other parameters such as relevancy, speed etc are all very subjective. For me, Ask works well and I'm satisfied with the results I get, so the relevancy is good for me. The new interface seemed cluttered to me too at the first try (see my comments to RWW's AskX posts) but now I like it, this is a matter of adaptation, just give it a chance, you'll like it - it always takes time to adopt a new thing but it's worth if it's worth :)

Comment from daniel read on 2007-06-11

2007-06-12T02:11:37Z

We have definitely seen a big jump in user satisfaction from Ask3D, so Emre is on the right lines here. It comes from both results being higher on the page & the new content that users are now exposed to.

If readers are interested, I would be happy to expand on what research told us about the ask3d design & how it fundementally changes the search user experience.

Daniel Read, Ask.com.

Comment from Charles Knight on 2007-06-12

2007-06-12T13:49:21Z

Barely 5% market share for the Top 100 Alternative Search Engines?

I'll stand by my estimate; it looks like it's actually 4.47%

http://searchtheweb2.blogspot.com/2007/02/rankings-top-100-alternative-search.html

Comment from Emre Sokullu on 2007-06-29

2007-06-30T04:19:43Z

Hi Daniel, I've just seen your message, I'd be interested in seeing the research docs you mention; we may share it with RWW community, I guess.