Image Search 4.0 and Digital Marketing

Any search engine does image search by using the text input in the search query form to search for the text around the image, the keyword tags, the listing in the webpage and the keywords in the direct links. However, there is no search engine (that I know of) currently exists, in which you can use an image as a search query.Nevertheless, I think eventually we would have some search engine that can give us this image string search functionality. ( May be Google is doing R&D on this.) For example, if I have an image of some famous monument that I dont know the name and details of, I should be able to search for the details of the image and the details inside the image, through a search engine, by simply copy pasting the image in the search query box, and clicking search. So if I copy paste a photo of Eiffel Tower, and use this photo as my search query, then I must get my search results rank ordered with pages containing the exact same photo as I used in the search query, AND, pages that has all the relevant information details about Eiffel Tower. The idea is to find all the places in the web that has the same photo, and then find more information about the contents and subjects in the photo.


Such image string search 4.0 capabilities can have number of applications in digital marketing. Imagine this. Harley Davidson sells 50000 bikes in 2009 Q1. A proud owner of a new Harley Davidson bike takes photos of his road trip and uploads on his social networking profile. This is a great advertising tool for Harley Davidson, only that this photo has no Call-to-Action for the viewers of the photo. Those who are really interested will google the bike model name (often there is no precise information available), and try to get more information, but there are so many hurdles, including user inertia, and often these linkless photos dont generate leads.


Now imagine this. Using image string search 4.0, we can now search for all photos in the web that feature that Harley Davidson bike. Firstly, the user himself can search directly using the image and get more info. But importantly, now the digital marketing manager at Harley Davidson, can build an application that will use a particular Harley Davidson model photo as search string, and search for all pages, blogs, and social media user profile. Users can subscribe to this application, and allow the advertisers to directly supply a link on the user generated photos. So the moment I upload a photo of Harley Davidson in my facebook album, the application would crawl, and if it detects the same photo of the bike model supplied by the Harley Davidson ad campaign manager, then the application would supply a link to the Harley Davidson bike page. The owner of the photo can earn revenue under the ppc model. 


For the sake of brevity, I have outlined such a simplistic and direct model. In reality, there are a number of challenges before we even perfect face recognition in photo search. I have always wanted this image string search functionality in search engines. But I know that unless there are some perceived commercial applications, or strategic gain, there are no incentives to develop this kind of complex technology. Thus I thought of an commercial application that I outlined above. Will such an advanced image search technology ever be possible? Hmm...Why not?

Comments