During the Web 2.0 Summit, I got a chance
to sit down with the team at Ask.com and find out more
about their search engine. This was straight after a Summit session entitled ‘Disruption
Opportunity: Beating Google at Their Own Game’ – in which Ask CEO Jim Lanzone and Senior
VP of the Online Services Group at Microsoft (and ex-Ask CEO) Steve Berkowitz discussed
with John Battelle how they are competing with Google. R/WW’s coverage of that
session is
here.
Letting the stats do the talking…
Whenever I talk to or meet Ask.com people, I always get the feeling they are a little
pissed off at the lack of attention they get from blogs. To compensate, out come the
stats to prove how big they are. For example, they often make a point of saying that
Ask.com is the 5th biggestsearch engine in the
US – behind Google, Yahoo, MSN and AOL. Also, Jim mentioned during the Summit session
that Ask is the 7th
biggest web property in the US – ahead of the likes of Amazon, NY Times and
Apple.
So there’s no doubting that Ask.com is a significant player in the Web business – and
I agree they don’t get their due for that. But what about the actual product – the
Ask.com search engine. How does it stack up? I spoke to the team and took the search
engine for a test drive to find out…
What really differentiates Ask from Google?
I asked this of the Ask.com team, in our hallway meeting at the Summit. They told me
that Ask’s technology “looks at the Web differently”. Whereas Google’s PageRank ranks its
search results by popularity, Ask has something it calls “ExpertRank”. Essentially this
is an automated search algorithm (like Google has), but on top of that results are
ordered using topic communities and the editorial functions that create ‘Smart Answers’.
While the ExpertRank formula is a secret sauce that Ask.com won’t divulge, they do say
that the top results in searches are determined by expertise – and not popularity. As it
states on their help
pages:
“Identifying topics (also known as “clusters”), the experts on those topics, and the
popularity of millions of pages amongst those experts — at the exact moment your search
query is conducted — requires many additional calculations that other search engines do
not perform.”
Smart Answers
I was curious to know how ‘smart answers’ are determined. For a start, they don’t pop
up on every search result – for example a search for “richard macmanus” displays my
primary RSS feed at the top of the page, instead of a smart answer. The Ask team told me
that smart answers are editorially done – and is a reminder of their natural language
past. If you recall, Ask Jeeves (as it used to be known before the butler was fired, er I
mean de-commissioned) started out as a search engine where you could ask a natural
language question – e.g. “what the heck is web 2.0?” – and get back a good answer. Smart
Answers is an extension of that philosophy of providing a natural language answer to a
user’s search query. It does this by a combination of automated data mining and human
editorial. But the human editors don’t physically write the answers, I was told – rather
they act as aggregators and filters.
I was told that currently over 20% of Ask’s entire search terms – and hundreds of
categories – have a Smart Answer.
Comparison of Ask with Google
If you compare Ask.com to Google, there are immediately some noticeable differences.
An obvious one is that Ask.com puts its advertisements within the main content pane,
instead of in a separate right-hand pane like Google does. So when you do a general
search in Ask, the right-hand pane is sometimes occupied by advanced search options. Also
Ask often has their ‘smart answers’ (see above) at the top of the main pane. The effect
of all this is to give the user more immediately useful information – and drill down
options – on the first page of results. This is what Jim Lanzone meant at Web 2.0 Summit
when he
said that “Ask.com enables users to do more, faster.”
Below are a couple of screenshots, showing a search on “new zealand” in Ask, followed
by the same query in Google:
Other Features & Conclusion
Ask.com also says it does social search, but rather than rely on user tagging – which
they say is only popular in niche tech circles (e.g. del.icio.us and Flickr) – Ask.com
lets its algorithm do the work. It does this by breaking terms down into groups and
presenting the results to the user. If you do a
search on gardening for example, you’ll see how it is broken down into multiple
categories.
Ask.com also has the usual search portal (circa 2006) features – mobile, maps, news,
blogs, binoculars (page preview), etc. There are subtle differences in all of those
features when compared to Google, Yahoo and MSN. But ultimately I have to ask (pardon the
pun): is Ask.com ‘next generation’ enough to challenge the big 3 plus AOL?
I do like the concept of ExpertRank and their willingness to get as much useful info
on the first page of search results as possible. It seems like an innovative approach and
certainly differentiates Ask from Google.
But when it comes down to it, the results I see aren’t sufficiently different
to make me want to change. I suspect a lot of Google’s 50%+ market share of users feels
the same. Ask.com is still a successful business though, even if they don’t manage to
make great inroads into the market. I’m sure Ask is not crying into its milk about being
5th.