Expectations are an often-overlooked factor in an internet startup's prospects for success or failure.

Hype, in one sense, is the irrational raising of expectations to a point beyond what is realistic and/or healthy.

While hype can, in the short-term, be very beneficial to startups, creating favorable funding options and lucrative exit opportunities; in the long-term, hype is rarely sustainable and often detrimental.

After all, when it becomes clear that inflated and exaggerated expectations about a company will never be fulfilled, the reaction is almost always harsh and swift.

In the world of internet startups, hype is most frequently related to the magnitude and longevity of general trends and to the beliefs that specific, rapidly-growing startups will mature into valuable, substantial businesses.

There are, however, certain startups that inspire hype before they even get off the ground and prove that they're "the real deal."

Perhaps the most appropriate case study of this in the world of Web 2.0 is that of search startup Powerset, which aims to revolutionize search through the use of natural language processing (NLP).

NLP is designed to, in theory, provide users with the ability to search more effectively using "natural language".

For instance, a search for "Is Drama 2.0 a good blogger or the best blogger?" would return results that lead to the information that provides the answer - a page explaining "He is neither because he doesn't exist."

Formed in 2006, Powerset (covered here) and its NLP search engine technology, which is based on technology licensed from PARC, has been hyped by some as a "Google killer."

Back in November 2006, Bambi Francisco, formerly of MarketWatch, stated "there actually is a lot of substance behind the intrigue" and detailed her experience when testing Powerset:

"After sitting down with the three founders over a few hours to test their search engine and compare it to the industry's big guns, I think Powerset has a great shot at improving the overall search experience.

"Indeed, searching with Powerset was a far richer and more liberating experience than what you get with the rivals."

Of course, her test was limited to The New York Times and Wikipedia websites since, at the time, Powerset's index was far more limited than Google's.

Fast forward to 2008. After years of growing expectations and buzz, more than $12m in funding and an $80m valuation according to Henry Blodget's inane SAI25 index, Powerset launched its public beta.

And yet the NLP search startup that caused Bambi to "shudder" when thinking "about the consequences of such technological advances" only searches Wikipedia and Freebase, an open database service.

As pointed out by Kevin Heisler at SearchEngineWatch, Powerset has essentially launched "Piggybackipedia."

In other words, regardless of whether or not Powerset's technology is the best thing since sliced bread, Powerset already finds itself, for all practical purposes, nowhere near capable of competing with Google in the real-world.

Unfortunately, even when it comes to gauging the prospects for Powerset's technology given where it's at now, quite a few of those who have taken the time to try Powerset and compare the search results from its ultra-limited index with Google's have found little to extol.

Larry Dignan at ZDNet wasn't entirely impressed and concluded:

"Bottom line: Powerset’s interface is nice and the search could be applied to multiple markets. However, the benchmark remains Google. The average bear isn’t going to ponder the semantic Web and its importance. We simpletons will want results. The real game for the semantic Web will be delivering those results."

Josh Catone at ReadWriteWeb wasn't impressed either. While he admitted that his test was "quick and informal," he makes a common sense observation that is worth noting:

"But our first snap impressions are that Powerset doesn't do a markedly better job of finding answers than Google for most queries. Some might argue that we didn't play to Powerset's strengths and frame our queries properly, or search for things obscure enough to notice any differentiation.

"But the promise of natural language search is that people don't have to learn how to search -- they can just ask questions as they normally would. We also can't expect that everything they're going to look for will be obscure and hard to find via traditional search engines -- more often than not, they probably won't be."

Of course, the fact that Powerset looks to be, at best, an interesting work-in-progress does not necessarily mean that the hype it has created won't contribute to a successful early exit for its founders and investors. Indeed, there are rumors that Microsoft has taken an interest in the startup.

But rumored acquisition interest notwithstanding, Powerset demonstrates the fact that managing expectations is extremely important.

For many, being left underwhelmed after years of Powerset hype has done the company few favors with the people who matter most - the consumers who Powerset hopes will find more value using their service than the services of dominant competitors such as Google.

Unfortunately, my observations over the years have led me to conclude that when it comes to internet startups, following the old adage, "underpromise and overdeliver" is still far better than overpromising and underdelivering.

While this is not to say that startups should intentionally lower expectations and launch less-than-compelling services, it is important to consider that managing expectations and preventing premature hype reduces the unnecessary risks that occur when you make a poor first impression that doesn't live up to the expectations you've set or allowed to be set for you.

If Powerset fails to woo an eager acquirer, I suspect it may learn this lesson the hard way.

Drama 2.0

Published 14 May, 2008 by Drama 2.0

237 more posts from this author

You might be interested in

Comments (2)



Is this really all that different to Ask? They've been using natural language search for years and their market share has never been into double figures. I would say the real Google killer, should such a thing exist, would be semantic search where user feedback becomes the major ranking mechanism as opposed to website feedback as is the case now.

over 10 years ago


Cyndy Aleo-Carreira

I knew you wouldn't drink the Kool-Aid! A year and a half later, and they still aren't any farther ahead than they were? And they index so much slower than indexing the entire web is completely out of scope at the present time. But no, there is no bubble!

over 10 years ago

Save or Cancel

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.