Wolfram Alpha and Beta PR
May 12, 2009
At the airport this morning, I flipped through a beta wave of publicity for the Alpha search system, actually the soon-to-be-released search system. I found
Christopher Dawson’s write up a useful example of beta PR. The article “More Details Emerge on Wolfram Alpha” here representative. For me, the most interesting comment in Mr. Dawson’s story was:
Google has also attempted to add semantic search capabilities (and I’m sure will get there sooner than later; they’re Google, after all), but so far, this doesn’t give you much.
The view seems to me to be that Google is not in the semantic search race. Wolfram Alpha, with its demo and previews, is.
My research for my three Google studies suggests otherwise. You can scan information about my Google studies here. I can’t easily summarize the research I have conducted over the last six or seven years. Making the situation more tricky is the fact that some of my work has been published by BearStearns’ and IDC as client-only reports.
Nevertheless, the notion that a demo makes Wolfram Alpha ahead of Google strikes me as incorrect.
The interesting question that i have been thinking about is, “Why are observers so keen on finding an alternative to Google?” What surprised me was the high expectations for Cuil.com by former Googler Anna Patterson and her team. Cuil.com has improved, but I have picked up hints that the GOOG has not been far from the Cuil.com project, particularly with regard to some tests on message collections.
What’s happening is a bit of cover your tail combined with wishful thinking. The pundits saw Google as Web search and ads for a decade. Now that anyone with a willingness to look at its mobile, shopping, maps, and other services can see that Google has been a platform and is now sufficiently diverse to make the Web and ads crowd look, well, anachronistic.
Enter a new search system.
The pundits claim that it is a Google killer without setting forth much in the way of a yardstick by which one can measure the progress of Google death. Here are some examples:
- Cost of infrastructure, to date, to grow, over five years
- Number of users of sophisticated search outputs. (Remember only about five percent of search users take advantage of advanced search features)
- Number of documents processed by time unit, including transformation, parsing, and indexing
- Business model (ads are okay but will advertisers pay Google scale cash flows to reach a sophisticated service)
These four points need some consideration. But when speculating about “to be” products and services, one has the advantage of working with modest evidence.
Stephen Arnold, May 12, 2009
Comments
8 Responses to “Wolfram Alpha and Beta PR”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
I dont think your evaluation of Alpha is fair.
Although your reasonings of why it cant be google killer seem alright, you seem to discount for the fact that it is a new kind of search—Computation on massive amounts of data.
At this point, Alpha seems limited to carefully curated datasets. Perhaps, some innovation in creating new datasets by understanding the structure of Web pages would make it super pervasive.
I think your write up is balanced: there are HUGE hurdles for WA to cross, even if it had a better service. Just building the infrastructure for a highly-available search service is an extremely expensive undertaking, especially without a secured revenue stream. The press always wants to write a story about a Google-killer without having a clear understanding of the complexities that would entail.
I wish Wolfram | Alpha the best and I think they’ve been making some smart moves but it’s just going to be a technology demo, not a finished product. Nonetheless, I can’t wait to play with it =)
Govind Kapra’
How can I be unfair to a demo?
Stephen Arnold, May 15, 2009
I’m watching this closely from afar (…) but it seems that WA is jumping the gun as far as “release” goes. I can’t find a beta, and his/their main websites smell like Public Relations BS. I love the idea, but worried we’re getting a lemon. Getting a computer to learn to say “mommy” to me is amazing, though…so I might be coming from the wrong direction.
[…] Wolfram Alpha and Beta PR […]
[…] Wolfram Alpha and Beta PR […]
[…] Beyond Search > “Wolfram Alpha and Beta PR“ […]
[…] Wolfram Alpha and Beta PR […]