Welcome to WebmasterWorld Guest from 126.96.36.199
Forum Moderators: open
"...today announced that it acquired Applied Semantics, a Santa Monica, Calif.-based producer of software applications for the online advertising, domain name and enterprise information management markets. Applied
Semantics' products and engineering team will strengthen Google's search and advertising programs, including its fast-growing content-targeted advertising offering."
They are interested in contextual ad technology.
Classic example of a nich company that has built world class applications that is bought to decrease google r&d and developmental costs of similar technology while preventing competitors from acquiring such, in a marketplace google is about to dominate. And with the continue purchase of best of breed technology and content property purchases and agreements, they are certainly well on their way. The captain of this google ship has only one thing on his mind at this point and google has never been further from its utopian search roots.
A key application ... is Applied Semantics’ AdSense product that enables web publishers to understand the key themes on web pages in order to deliver ... advertisements
"Oingo is a little simpler in that you type in a search word, and Oingo returns several possible meanings for the one word. Type in turkey, for example. You will get several checkboxes allowing you to further clarify what kind of turkey you meant. You can chose turkey as in poultry, turkey as in dud, Turkey as in place (several places in fact) and more. If none of Oingo's attempts at context are relevant, you can choose to search again for just the plain text of the word with no additional relevance."
As you can see, AppSem can not only determine page content, but can place the page content into one of several meaningful categories despite keywords which might fit into one of several meanings. (And bear in mind I wrote the above in October 1999. I'm sure they've improved and expanded the technology since then.)
I can see where this would be useful in building an ad distribution network, but I'm also thinking of the problems AOL is having, the resources they might not have to put into DMOZ, and who I would want if I found myself suddenly wanting to build a usable Web directory from scratch....
Nice post. Lost you a bit in the last paragraph though. Maybe you can expound.
Heini. Thanks. Google indeed is a leader in determining what a page is about from a search standpoint. Now that they have turned the corner in the Ad serving world, a different set of problems present themselves, such as RBuzz suggests. For example, in the search world, RBuzz's excellent example of turkey is not a matter of contextual interpretation. It is a matter of the ALGO. In the search world, mistakes are okay, because the owness is on the searchers (and site owners) not Google to refine the search in RBuzz's example. In other words, what is the searcher really looking for turkey recipes, hunting, the country etc? and Google has no real resposibility in the search world to figure that out. In fact, they can't.
In the ad world, though, everything changes and with googles recent moves in the contextual advertising arena of its site, comes a significant set of responsibilities. The major one is not serving an ad for turkey recipes on a content site dealing with the country of turkey. That would make both the publisher and the user angry. Google doesn't want that, and in the case of the recent acquisition, Google is acquiring what they consider to be the pre-eminent mind share and application infrasture in this space. Remember, for Google, the concept of ADwords is still quite new and not without its problems. We can't expect that Google can mobilize in the contextual advertising space in the position of a leader without acquisitions. That would be far too expensive and unreasonable.
Google is acquiring what they consider to be the pre-eminent mind share and application infrasture in this space. Remember, for Google, the concept of ADwords is still quite new and not without its problems. We can't expect that Google can mobilize in the contextual advertising space in the position of a leader without acquisitions. That would be far too expensive and unreasonable.
Actually, I'd say Google considers themselves to be the preeminent mind share in the contextual space. After all, they rolled out a much larger system than Applied Semantics did and without acquiring AS at the time. Nor did they suggest that they were having any problems in doing this.
The AS DomainPark program is far larger than its contexual ad program, and that's the real gem Google has picked up. All new traffic, a system that Google has nothing like and traffic that will be taken from Overture.
That's not to downplay that Google will pick up the AS contextual ads technology and make use of whatever benefits that offers -- as well as the people and limited partners involved with the AdSense contextual program. But buy AS to solve the "turkey" problem? Hey, Google was adamant that this wasn't a problem with its existing contextual ads technology. AS tech may help refine this, but it's already supposed to be a core compentency of Google's contextual program.
What Google cannot do well -- or at least has never shown this -- is autocategorization of content. AS technology takes paid links, then puts them all into nice, neat subject categories that forms the heart of its money-making DomainPark program.
That categorization technology may now get some other uses, such as to help people refine queries when doing Google web searches (look for cars, get suggestions to also search for trucks, used cars, sports utility vehicles).
"Applied Semantics' products are based on its patented CIRCA technology, which understands, organizes, and extracts knowledge from websites and information repositories in a way that mimics human thought and enables more effective information retrieval. A key application of the CIRCA technology is Applied Semantics' AdSense product that enables web publishers to understand the key themes on web pages to deliver highly relevant and targeted advertisements."
Danny, slice it or dice it as you may. This is about serving better ads. Regardless, of the manner in which it is ultimately achieved. Of which, I expect some surpises and ingenuity from Google. Additionally, Google might not of admited weakness in the arena of contextual advetising, but what makes you think they had to. You ever buy a kewyord phrase on ADwords without a bracket  only to find you ads showing up in a manner (for search phrases) you never imagined. Do you think BLOGs was the pot of gold at the end of the rainbow or the tip of the iceberg? Morevover, Google itself suggested that this move was people (human resource intentiligence) oriented. They are looking for a jumpstart to their application development needs in the space wherein they inherited, not earned, the designate as the "largest performance-based search advertising program" in the world.
It may sound trite when you hear the examples given by companies like AS when they explain how important "disambiguation" is. But in time, linguistic technologies will make today's keyword-centric technologies look a bit dumb.
I'd link to my two recent efforts to explain this stuff, but as Tara pointed out, the TOS here don't permit it, and I don't feel like cutting and pasting. In any case, good for Google. They admitted they don't know it all.
One thing to watch for down the road: a much better "search term suggestion tool" - at least for advertisers who spend enough with Google.
Wordtracker and the like may also be on the way out as more sophisticated tech of this nature is housed within Google. You have the benefit of lateral thinking married with the real value of real time keyphrase search behavior data from the world's biggest search destination.
I'm sure Steve Harmon's group who led the second round of investment in Applied Semantics are very pleased. Oingo raised $1 million in their first round and had a *very* tough time raising the second round but eventually found backing to the tune of $5.4 million in Oct. 2000. Their investors evidently forced them to move much faster on "applying" the technology to moneymaking schemes like DomainSense. Google, the new sheriff in town, also seems to have more interest in the CIRCA technology's bottom line merits than in its potential to steam up the lab with interesting new ideas about search... at least for the time being.
But longer term, here's what you guys have been missing about this story. As semantic technology begins to make a stronger contribution to web search, SEO as formerly practiced becomes a quaint anachronism. You cannot "optimize" nearly as easy for ideas as you can for keywords. For example, you could keyword-optimize a title really well for a phrase, and find yourself ranking lower than some site which doesn't even contain those exact words anywhere! Let's say that on semantic grounds, those words are counted as matching the search query 81% as well as an exact match could be. At present, the lack of those words on even a high quality site would literally bring it to "0%" on that phrase, so the site wouldn't rank on that phrase no matter how good or topical the site was. Now (and by now I mean a couple-three years from now), clearly reputable sites on a given topic may begin beating out lower-quality "cleverly keyword-optimized" sites even on the keywords they've carefully optimized for and even if those exact keywords don't appear on the higher-quality site.
Applied Semantics aren't the only group doing this, of course. It's the wave of the future. Google will have plenty of cash, and could apply a second group like them, if they felt it was important enough.
Could it also be that Mike Grehan will someday have to take back his claim that search engines do not consider "themes"?
Stay tuned... for at least a couple of years... for the answer.
Categorize newswires to route them to the relevant reporters or editors.
Categorize, keyword and summarize content in an editorial system so that editors do not have to spend time writing metacontent.
Categorize syndication and archive content for more effective information sharing and retrieval.
I'm beginning to wonder if there's a syndicated news service in the offing similar in concept to Froogle.
It seems to me, however, that the ideas surrounding search, news, and the like fail to consider the current condition of Google operating model. Search has been a market share mechanism for G. A costly one. Ever wonder why a company with 75% market share can be so far behind its competition from a revenue standpoint? This is now changing and changing rapidly and search and news isn't the reason. ADwords revenues are exploding and Google has much ground to develop relationship wise, human resource wise, and infrastructure wise.
I disagree that it will take a few years to see this fruits of this aquisition as Andrew asserts. That is the google of the past, not the pre-IPO, revenue centric, largest ad program in world Google of 2003. Investing and reinvesting in search through acquisition that is not associated with revenue generation is not what is happening. One must presume at this stage that this acquisition is not about R&D, rather it is about the immediaty applicability of applied technology to G's current operating model to support significant revenue generate- at a quarter over quarter clip the likes that we have not seen in the Internet arena. To support such significant revenue growth, they must have the technology and human capital to support their ideas for serving ads which are vast.
If infact this aquisition is search related to the extent to which is being suggested in previous posts, then one must believe that the new quality of search of the future has revenue attached by way of ad serving. If search doesn't make substantive money for G, who has not tried to be a Y or AOL from a community standpoint, why in the world would they continue to invest substantially in it without the prospect of serving really relevant ads? I believe the captains of the G ship are becoming infatuated with the prospects of revenue. A shift with far reaching consequences in how we percieve G's business moves and motives.
[edited by: Chicago at 2:47 pm (utc) on April 24, 2003]
An interesting tit-bit from the article
Applied draws about a half billion unique impressions a month across its domain-name network, with a click-through rate of 50 percent on many of the listings. The company collects fees on the commercial links and splits the money among itself, the registrar and the commercial ad network
Thats a lot of traffic and a very good clickthro ... i didnt imagined parked domains will generate this much traffic ...
With an average PPC of 10 cents (very conservative) this is $2.5 million in revenue per month!
As far as potential applications go, contextual advertising, and anything remotely related to natural language processing intertwined with themes of data mining and info retrieval is a very hot field indeed. Even though many of the applications can be considered intrusive and sometimes even parasitic to ones surfing habits, in the end, you are better able to target content and increase conversions significantly. Good acquisition, but I'm not convinced Google bought out the best company in the niche.