• http://www.enotes.com Alex

    Great article, very insightful. I know exactly what you mean about “mouthfeel”, which is probably what dooms many new search engines as they try to compete head-on with Google by offering different features.

    The only point I would disagree with is on the name. I think that today an odd name that’s not easy to remember might actually help a firm that genuinely does have everything else nailed down. If the site starts at wolfram-alpha.com or whatever, it will limit growth at first but also make early adopters feel like they’re playing with something real, non-commercial, and different. Later they can always switch to the domain name wolf.com or walpha.com or whatever to make it more palatable to the masses.

    At this point in the game, if you can get the results and the moutfeel right, the only way you’ve got a prayer is some sort of interesting marketing, and an odd name that still works as a word might help break through.

    My 2 cents.

  • http://deliberateambiguity.typepad.com philosophygeek

    As you probably remember, I was the product manager at Powerset when we launched. Despite a lot of hype at the beginning of Powerset, we worked very hard post-Powerlabs launch (i.e. Sept ’07 and beyond) to quell the hype and focus on the technology under the hood. We were confident that we could build a really great search engine, but knew that we were years away from the final creation. With the release of our Wikipedia product, we wanted to convince the world that Powerset had created something unique, useful, and technologically advanced. Of course, we got an excellent result: the press understood what we were doing, we got users to prove our value proposition, and eventually we bought by Microsoft.

    After observing Cuill’s botched launch and Wolfram’s audacious claims, I’m surprised that no one learned from our example.

    Oddly enough, I wrote a post today about how not to analyze a search engine. Your article reminded me that even search engines themselves seems to have trouble describing why they are noteworthy.

    -Mark Johnson, Powerset/Live Search Program Manager

  • http://thenoisychannel.com/ Daniel Tunkelang

    For what it’s worth, I’m one of the folks that got to see a preview, and I was allowed to off road. Moreover, Wolfram Alpha reached out to me after I’d written a scathing blog post about their pre-launch hype entitled “A New Kind of Marketing (NKM)”–not exactly suggesting I’d be easily dazzled. I’d also say I’m “in the know”–I’m the Chief Scientist and co-founder of Endeca, and I blog a fair amount about search at The Noisy Channel.

    My verdict: the product looks interesting technically (I was pleasantly surprised, given my initial expectations), but they clearly don’t have their act together, either in terms of articulating a business model or even credibly positioning and packaging their technology. Beyond agreeing with all of the points you make in your post, I get the sense that they are tinkering and treating this like a science project rather than a production application. Perhaps they’ll find their way after they launch, but my guess is that they’ll have generated so much negative baggage by then that it will be hard for them to dig their way out. I hope I’m wrong.

    You can read more of my commentary about Wolfram Alpha here

  • http://www.ignitepr.com carmenhughes

    hi Danny,
    what a great post. I can appreciate your perspective of shaking your head at the approach that this startup has taken. Having worked with tech startups for about as long as you’ve covered search, I too, have seen some questionable approaches that startups decide to take. First of all, i must comment quite honestly about the name: absolutely terrible for a number of reasons that i won’t get into. In fact on a scale of 1-10, i’d give it a -3. Unbelievable that this company would not seek you out as one of the absolute authorities on search, which begs the question, who is driving their launch? It really sounds like they have very limited experience in how they are going about it. Also, particularly in the search space, given the false steps many have taken previously, and some rather recent, it is so important to essentially under hype and over deliver with the overall user experience; to position a search startup as the next Google-killer is basically setting the company up for failure for all the reasons that you listed, coupled with the road-kill of all the previous “Google-killer” search startups. Anyway, i loved your post and i hope it serves as a wake-up call to others.

  • http://searchengineland.com Jonathan Hochman

    Be careful, Danny. Jonathan Zittrain seems to be involved in Wolfram Alpha. He’s a careful person and technically knowledgeable. (clipped from WP: He holds a bachelor’s summa cum laude in cognitive science and artificial intelligence from Yale University.) He’s seen the product and said it’s good.

  • http://radar.oreilly.com timoreilly

    Danny -

    I haven’t been following Wolfram’s positioning of Wolfram Alpha, but if they are calling it a search engine, they are misrepresenting it. It’s not even remotely in the same category as all the wannabe search engines that have gone up against Google. It’s essentially a web front end for Mathematica, and a whole bunch of scientific data. Web search isn’t even remotely part of the product.

    It’s a search-style interface for stuff that you calculate or calculate with. It does have a natural-language style interface, but even there, it’s more of a natural language interface to mathematical equations.

    Re the critique that they didn’t show it to folks like you and Battelle – that suggests that the “search engine” positioning is more in your mind than theirs. When they asked me to put together a small group to look at it, I didn’t think of you guys either: I showed it to a bunch of scientists and electrical and mechanical engineers.

    Yes, there are some data sets that are “softer” and that will appeal to the general public, but this is primarily a big win for the scientific community. And it’s a very cool advance in thinking about how to make “computability” a criterion for what kinds of data sets to bring together.

    P.S. I’m also very impressed with blekko.

    P.P.S. I also agree with Daniel Tunkelang that there isn’t a sense of business model (except to sell more Mathematica, so you can cross-search and cross-compute the public data sets in Alpha against your private data).

  • Danny Sullivan

    Thanks, Tim, much appreciated. Like I said, I count your opinion pretty highly. And if it’s not really a search engine like Google, that explains a lot. But they’ll still have expectation problems. The reviews so far have, to me, positioned it as somehow a rival to Google (even in another area) or bigger than Google. They aren’t responsible for the slant of those reviews, of course — though they can help influence them one way or the other. But I guess I figure when they launch, we’ll see lots of people who aren’t in the scientific community assuming it’s like Google (journalists among them), doing a search, perhaps coming away disappointed and feeling it was all hype.

    None of that matters, of course, if it’s not really meant to be a standalone service or have a serious business model behind it. If it attracts the audience it’s after, success!

    Anyway, I’m looking forward to seeing it in the end. More specialized search engines are great, and even some strong rivals to Google are great. I’m just hoping some of the future services learn some of the mistakes they can make in setting expectation.

  • http://searchengineland.com Jonathan Hochman

    Tim, one thing to keep in mind is that tools often turn out to be useful for different things than what people expect. e.g. Wikipedia was a tool for creating a rough draft of Nupedia. If Wolfram creates an interesting tool for organizing scientific or mathematical data, there might be interesting commercial applications. Google is horrible at organizing info like “what Yankees tickets are available for afternoon games in May against teams that aren’t in last place?” There is a large opportunity for somebody to solve that class of problems.

  • Eric Logan

    I had the opportunity to watch this past Tuesday’s Webcast demo of Wolfram Alpha.
    Stephen Wolfram was very frank about the limitation’s of the calculator in response to requests for queries involving popular culture. The presentation format was enabled with acrobat connect which I had never used previously. I walked away from the experience with the thought that it was a very open and candid discussion covering both the strengths and limitations of the system. The Screen cast did not in anyway seem to be a polished or controlled demo.Stephen stated that he was literally running a very recent beta version that was a couple of hours old. When confronted with a minor glitch which he resolved in a humorous manner. Overall, I was impressed with the presentation and also realized that my initial interests are probably less academic than Wolfram Alpha’s present focus.

  • Missmcj

    Danny, with all due respect (and I certainly have plenty for you):

    D – “Fail to brief those in the know:”

    Yes they did, they did a demo at Harvard and also a webinar for researchers. I was at the webinar, and I’ve worked in search for a while now on the IR side and SEO side. In this case I was invited because of my work in computing as a researcher.

    D – “Nor have I seen anyone who regularly writes about search having said they know about the service.”

    I wrote all about it here:
    http://www.scienceforseo.com/search-engines/wolfram-alpha-rocks/

    They are in the know on who is in search seeing they knew which researchers to invite to the demo. SEO and online marketing doesn’t really feature in this engine imho. This is why there was no real noise made in this particular community.

    As important as SEO is, the big names we respect in the industry are largely unknown in computer science. I know may SEO’s have no idea who a lot of up and coming and pioneering search researchers are. It’s nothing personal.

    D – “Controlled demos mean that a service can pick-and-choose questions where it knows it excels.”

    They didn’t in the demo – they stressed it was an Alpha and there were a few hiccups too. It’s called “Alpha” even. I don’t see how they could make it clearer.

    D – “Claim you’re the “next Google” or fail to set expectations”:

    In fact they stated that they were nothing like Google, and indeed they really are very very different. It is definitely a “new paradigm.” – I think a lot of blog posts wanting to attract traffic are mostly responsible for this erroneous view.

    D – “most of us” will be using Wolfram Alpha by default”:

    I doubt that very much. I think you’ll be using it over Wikipedia perhaps. Also, I would not be surprised if it was “as important as Google”.

    D – “And chances are, it’ll fail to do so”:

    Why? If you don’t have enough information about it as you claim in this post, then how can you say that? You can probably assume so and that wouldn’t be out of order considering the plethora of rubbish engines that have littered the internet recently. I do think however that good engines are often left by the wayside when they are in fact very cool indeed, look at Duck Duck Go for example:

    http://duckduckgo.com/

    I have “mouthfeel” for Wolfram Alpha, and you know what? I think that nobody could as jaded as I am, however jaded you think you are. I get emails all week about new ideas, algorithms, research systems, beta and so on. I welcome those emails because I’m open to anything and Have learnt not to judge before I’ve seen it for myself and had a think about it too. It’s rare for me to be excited and I am.

    D – “And how do those measure against the types of searches a typical search engine handles?”

    I think that you will be very surprised at the format you’ll be seeing. I doubt very much you will be using it for navigational queries at all. There is no SERPS. It’s very different to what a lot of people expect and have speculated about. This is NOT a typical search engine.

    D – “…in a niche that’s separate from what the existing search engines do”.

    You’re absolutely right.

    D – “…it’s more that if I don’t see anyone who regularly covers search talking about a service, that’s usually not a good sign.”

    !!!!!!!! I cover search all the time and so does Daniel Tunklang !!!!!!! – Get with the program :)

  • http://searchengineland.com/ Danny Sullivan

    Marie-Claire, I cover search. Period. That includes search marketing, but it is by no means just search marketing that defines my work. So I guess I don’t feel I fall into some search marketing ghetto. I’ve covered the size of search engines, relevancy of search engine, patterns on how people search, search user interface advances. For years.

    I do think Wolfram appears to have done outreach to more to research folks. And I totally get that they probably are thinking within the traditional information research area. But this is a product they say is explicitly aimed at consumer searchers — as I’ve outlined in my postscript. Sorry, I just haven’t seen them having done outreach to those who cover things for those consumers.

    Moreover, that suggests to me that they’re going to have some wake-up calls about just how the actual “average” searcher might make use of their service, if they’ve not looked outside the traditional IR channels. It’s a classic mistake that’s happened time and again with search engine launches — there’s a great research idea that doesn’t pan out so well when it hits the commercial world.

    As for stating they aren’t like Google, been there, seen that. Seriously, they are far from the first service to say “we’re not like Google.” That’s also a marketing tactic.

    The bigger point is the suggestion of bringing in a new paradigm — that this is somehow going to radically change how we search. They do seem to suggest that, and I think you’re seeing that suggestion rubbing off on the review that are coming out. And all I can tell you is that for a mass audience, those types of shifts just don’t happen. Maybe Wolfram really will do it, but that type of claim along with other things just sets off the red flags.

    And yes, plenty of good search engines do get left by the wayside (looked at Duck Duck Go, and sorry, I wouldn’t count it as one). One reason they get left by the wayside is that they come forward without good business plans, without good thought on how to reach their audiences, without a good search product period along with other issues. Hence the point of this post. It gets tiring seeing the core of good products shoot themselves in the foot by not being prepared.

    Remember Applied Semantics? They did some cool things — and in the end, that dropped by the wayside and turned into AdSense and powering paid links on parked domains (Google had its own contextual product before buying Applied Semantics, but after they got it, they made use of the AdSense name). Several other seemingly cool consumer search products ended up shifting to the enterprise or paid advertising areas because the product turned out not to be as cool or as useful to as many people as they expect.

    So how can I say chances are that Wolfram won’t massively change how we search. Because chances are, it won’t, based on my experience in covering the space. I think I’ve said that maybe I’ll be wrong. It would be pretty cool to be wrong. I’d like to see some major paradigm shifts. Like you, I’m also open to new ideas. But that doesn’t mean that I ignore my experience, either.

    Anyway, I appreciate your comment and your review. Everyone who has seen Wolfram in action comes away impressed. Here’s to hoping I will be, too.

  • Missmcj

    You are most welcome – It’s good to have reviews from different stand points, they all merge somewhere and allow for a more realistic view. Duck Duck Go is a brilliant engine for bloggers in my mind. I think a mistake we are all guilty of is to compare everything to Google. I am a big Google fan, but I think there are a lot of new directions that search is taking, and I see us using other engines as well in the future. Wolfram is but an example.

    I hope you like Wolfram too, I look forward to your review.

  • http://thenoisychannel.com/ Daniel Tunkelang

    The business development guy from Wolfram Alpha who talked to me seemed to realize the danger of overhyping the technology, while Wolfram himself clearly doesn’t. And Wolfram doesn’t exactly have a track record of underhyping his accomplishments. Perhaps this will be like New Kind of Science: popular with a critical mass of devotees, while lambasted by establishment critics–despite concessions that there’s at least a kernel of value buried in there.

    Of course, selling actual, usable technology isn’t the same as selling a coffee-table book. Wolfram may be satisfied with a cult following; but, if his goal is to make his technology useful and practical, he’ll have to play better with others. That is, unless he really thinks he can create a compelling stand-alone application. At this point, I plan to wait until the public release–which is supposed to be next month–to see how it looks without any training wheels.

  • marbaehr

    I agree that Wolfram Alpha is likely to be highly based on Mathematica and I think it’s going to herald a new category of search. (I have not seen a demo so am guessing.) Curated data is key #1. If you look at what’s already been curated in Mathematica (linguistic, economics, geographical, mathematical, financial, chemical, biological) you can get a sense of what is already there. It is not comprehensive but it is lots. There is a knowledge hierarchy and ontology in it. The Mathematica language allows all types of data to be intermixed, manipulated and presented in many, many ways. So this indicates a new way to pull data out. Mathematica’s own search capability is very robust and so gives a clue to how answer selection and presentation might work. And, their Demonstrations Project already has almost 5000 demos, which gives an indication of how user participation might amplify what is already curated. Wolfram Alpha will not be a Google killer but it may stand nearby to enhance. And it may not yet have a business model but neither does Twitter. I look forward to seeing how I might prep and curate my own data to add it to this exciting repository. And, maybe Stephen Wolfram is not so bad a marketer… just look at all the free press he’s getting right here.

  • http://www.marketingwithpaul.com pamoore1

    Interesting post. I have never heard of Wolfram, but I will have to check it out. My default search is Google. It appears nicely as my homepage when I start Firefox. I know Wolfram has got its work cut out for it…

  • Simone S

    Has anyone checked out http://www.sweetsearch.com? It seems to have avoided the PRish pitfalls of Wolfram outlined above.

  • http://www.searchcowboys.com Bas van den Beld

    I think you hit the bullet here Danny. A major issue in my opinion with the launch of these search engines, and also with other start ups I might say, is the lack of reality. Reality in different forms: where do you stand in between the already existing competitors AND who do we need to address, and how.

    I’ve noticed that there seems to be some kind of ‘arrogance’ among these ‘start ups’. They feel their product is best, which is fine, but then they don’t look at the world around them. Getting the support of those that matter is one of the key points in getting the success you want.

    Besides that, aren’t we all looking for ‘the hype’ too much? I get a lot of press releases at Searchcowboys and most of them are from startups who think they have created the new Google. Realism is not always a part of what they are doing.

    I’m actually talking about a part of this subject at A4UExpo Amsterdam on Wednesday, thanks for the last minute input!

  • http://brianratzker.com brianratzker

    Some of the best inventions were made from unintended applications. You enter your question like a search engine but Wolfram Alpha is clearly not a typical search engine. It is a database of knowledge and the ability to intelligently answer questions. The technology may be applied to commercial searches in the future but it may done by another company.

    They are all about the technology of this new search engine. I agree they should have contacted prominent figures in the search engine industry if they wanted to create something for profit but I believe this project is less for profit and more about creating something revolutionary. Who knows, they may partner with Google or another company to commercialize the technology.

  • Andrew Goodman

    Hi Danny,

    I’ll agree with a couple of premises:

    (1) Blekko looks interesting (and as always good mouthfeel) – can’t wait to see the latest developments;

    (2) Dr. Wolfram is evidently poor at working with PR people, and they, not so good at working with him.

    I’m not sure that from this it follows that the product is not worth considering for its purpose. It could be a great tool.

    Wolfram isn’t the only company in the world that pays a lot of attention to you when they’re looking for regurgitated press release coverage, and then goes away when you’re actually looking for them. That issue is annoying across the board.

    Regardless, my thirst for useful, new, non-overhyped tools in this business is so strong, I’ll look seriously at something that is useful, new, and slightly overhyped but is at least coming from a scientifically credible organization with a history of shipping product.

  • Scrapsflippy

    I “asked” Wolfram Alpha, “Why did you guys pick such a stupid name for your search engine?” and it returned, “Wolfram|Alpha isn’t sure what to do with your input.”

    I asked Google, “Why did those Wolfram Alpha guys pick such a stupid name for their search engine,” and this article was the first search result.

    I don’t work in the “search industry.” I’m just an end-user consumer who found this stark contrast so funny, I jumped through the hoops necessary to register so that I could post it here as a comment.