As I mentioned on Wednesday, I attended an update from Microsoft's Powerset team in San Francisco, to gain insight into how the Redmond machine was viewing what's considered to be a mature search market - seeing just what opportunity they believed they had in a space where Google is the king and Microsoft's offerings, whatever they are called these days, usually are viewed in a role similar to court jester, fairly or not. And while the team aimed not to drown us in a well-known ritual called "Death By PowerPoint", they offered us a background presentation to take home on USB keys, chock full of statistics. We were sworn not to distribute the slides, but permitted to use any and all data from within. Maybe it's a style thing.After looking at the provided materials a few ways, it's clear that the Powerset team at Microsoft is looking at massive amounts of data, citing that search engine users report high levels of satisfaction with their current search provider, but waste an incredible amount of time honing search queries to find the perfect result. Amidst a digital content information explosion rich with links and an ever-growing number of Web sites, Microsoft believes that better tools are needed to help users find their answers more quickly and get on with their solution.
Their statistics, which cited Pew Internet, Comscore and Harris, among others, state:
- 49% of people perform Web searches each day
- Users are searching 23% more frequently per person, year over year from 2008 to 2009
- Satisfaction with search is at approximately 65%
Assuming the data is correct, and Web searchers are taking multiple queries to find an acceptable action, Microsoft found that even after users come to a decision, ostensibly by clicking a search result, almost half the time the user will end up searching for the query again, with 19% performing the exact query, and an additional 30% being a partial repeat of that query, signifying they weren't satisfied with the link they had selected.
The end result, as Microsoft lays out, is that despite users saying they are happy with their search engine experience, about half say they do not expect their queries to be successful. They cited success expectations ranging from 53% to 57% in the categories of Product Purchase, Local Activity, Specific Facts and Finding Information. While more than two thirds of users say search is frequently part of their decision making process, half they time they anticipate being disappointed.
As with any Marketing-driven presentation, Microsoft tries to summarize their to do list with three buckets:
- Opportunity: Better Results
- Opportunity: Organized Experience
- Opportunity: Powerful Decision Tools
Searching through the deck doesn't mention Kumo, or Live Search, or Powerset or MSN, or any of the many names Microsoft has tossed around. It simply says there are issues to be solved, and suggests the opportunity is out there to create something better that helps customers and delivers a higher level of satisfaction. From the discussion on Tuesday, it seemed the Powerset guys believed they had a more relevant approach to delivering better results, and possibly in a better organized way. The struggles, of course, will be in how they can search out new customers who will find a better experience with their product against well-respected competition. Can they better optimize their results to reduce the number of times you execute a query?
I would believe that much of the blame lies not with the search engines, but the links themselves downstream. Until Microsoft, Google, Yahoo! or anybody else can be held accountable for the whole content of the Web, at the end of the day, they are passing you off to a third party, who may or may not have the answer you are looking for. It's possible to continue tweaking the algorithms to and fro, and making the search experience easier, but there will always be room for error, with the exception of fact-based searches (which may explain why Powerset started with Wikipedia as its data set).
Additionally, every search engine is working with a set of users that could be improved. While it's a faux pas to blame the user, there is a very high chance, in my opinion, that users are not really sure what it is they are looking for, and their own search methodologies could be improved. It's likely the Web has reached a stage where "good enough" is "good enough", and multiple queries or keyword tweaking will be the norm for some time.