Commentary
Sophisticated Search Technologies Could Displace Behavioral Targeting
- by Laurie Sullivan , Staff Writer @lauriesullivan, September 23, 2010
Sophisticated technologies like this one makes me wonder if SEM technologies that target ads will make behavioral targeting technologies of the past obsolete. Search has become a strategic marketing channel, and I can't help but think behavioral targeting technologies will come in second to platforms that can target based on behavior rather than those that follow consumers through the Internet collecting and sharing information.
After all, along with Google Instant, these technologies predict intent. I have seen the transition occur in other industries much older than the online advertising industry.
Although a far cry from traditional behavioral targeting technology, Adchemy's software as a service platform (SaaS) automatically tailors marketing messages in paid search ads based on intent by interpreting the consumer's online searches.
advertisement
advertisement
SEM agencies and Fortune 100 companies in telecom, insurance, consumer banking, retail and CPG continue to test Adchemy WordMap. Some notice a rise in paid search conversions by up to 400%, according to Anurag Wadehra, senior vice president of product and marketing at Adchemy.
The intelligence built into the Adchemy WordMap application will allow marketers and search engine marketing agencies to dramatically improve the performance of paid search campaigns. Adchemy WordMap extracts the underlying consumer intent or attributes such as brand, price and color. If the consumer searches for "cheap laptops," it signals the value in the keyword "cheap." Search queries create a word map that tells the technology the ads to serve up, which allows the platform to target ads based on consumer behavior.
The word map built with natural data can be stored and hosted for periods of time. All the search queries then run against the word map. If someone searches on "a white shirt for a party" the copy for the paid search ad gets constructed to serve the consumer searching on the query.
A relational database organized by topics matches the relationship with the queried words. Keyword expressions determines how intent signals get grouped together to become queries, allowing marketers to parse out keyword or query intent -- intent that might displace what we've come to know as traditional BT.
Laurie Sullivan is a writer and editor for MediaPost. You can reach Laurie at lauriesullivan@gmail.com.
Search is great at matching advertising to consumer intent as is display retargeting. However marketing is also about generating intent which is most effective through display.
Intersting technology and results for sure but I don't think this would displace behavioral targeting. BT is a bit of a loaded term but in the broadest sense is not solely leveraged for conversion activities (although that is a big part of it). BT can and should be leveraged for identification of audiences in ways that demographics fall short. Also it can be leveraged in cases where there are long purchase paths and you're looking to move audiences down funnel. Finally, search is by nature intent based, but most people aren't searching for products all day long. BT provides you a way to stay in front of the audience in a more passive setting.
Search always had a leg up on display as it is pull rather than push, the user is telling you want he wants and based on the results optimizing his queries to better improve the results. Now there certainly is value in dynamic creative targeted to the search queries which is something that display has been working on for some time and utilize regularly with re-targeting.
That said, all of this is fine and dandy however we need to consider 2 things, 1 what is the cost benefit relationship here of utilizing the technology, i know display has been having a tough time justifying the added costs of multiple layers of data and technology, how does this tech size up, and 2 as with display we run into the same issue of multiple users on the same machine and how do we avoid building a misaligned wordmap or a misaligned BT profile.
It would be interesting to see a large sample size and some empirical data rather than a vague claim about a 400% lift....
BT based on semantic content, user provided or URL related is different from what the author proposes from Adchemy. This is apples vs oranges.
I would only add that search words are a weak predictor for any individual over time to create word maps because of the non-conditional probability that one word used at one time is correlated with another word at another time.
Search queries have very low Bayesian conditional probabilities across the time domain.
One could say that someone interested in music or artists or movies may make more queries in this domain as opposed to finance or travel or other categories.
Given that assumption, the correlations would not be enough to create useful word maps simply from key words in searches.
If you can remember what you have used Google for in the last month (easy to do with Google) you would be surprised how broad the word map would be.
I think that Adchemy is a good start but not at all equivalent to what BT can provide. The problem with all these solutions is that they are not mature enough to create dynamic ontologies that are useful for the individual or can be aggregated into market segments that advertising can use.
Search technologies such as AdChemy are unlikely to supplant real BT efforts given that BT is push marketing and search is pull marketing. The one drawback of search is that you cannot search for something you don't know exists. BT creates the awareness that drives people to search because it identify audiences. Search on the other hand waits for the audiences to identify themselves.
While AdChemy may have some influence on re-targeting efforts, it is unlikely to have much impact at all on true BT.