Welcome to Imprezzeo, a new image-based search company. We started this blog to give you an insight into the thoughts of the executives behind the company, and as an outlet for your feedback. My colleagues, Kurt Dressel, our VP of Sales, and Peter Chin, our CTO, will also be contributing in this space, where we will talk not only about our company and technology (as wonderful as they are), but about image search and digital photo markets in general, and other topics that touch on our reason for being here.
So what is our reason for being here?
The 'about' section of the site leads with the phrase "we use text to find text, why not images to find images" and that pretty much explains why we are here. Not using images for the sake of it but because they are more suited to the task of finding 'more like this' than text alone. What better way to search for an image than to provide an example of what you are looking for?
Imagine the art director looking for a particular image for a campaign, a picture desk editor seeking one for a publication or a parent searching for an illustration for a party invitation. Each has a picture in their mind's eye but how do they go about accurately articulating what they are after. More often than not they put a simple search term into whichever source they use and settle in for an afternoon sat in front of a computer screen, wading through page after page of results. Far better we think to search using an example, either from an initial result set, an uploaded picture or indeed your own illustration. That way, the user gets what they want quicker and the search service gets a satisfied user. I expect to revisit this simple yet apparently elusive duality in more detail in future posts.
This is not to say that Imprezzeo's agenda is to eradicate text tagging completely – there is a place for them, for example in describing an emotion in an image. Anger might 'look' very different in images tagged as such. But there is a very real risk that as the corpus of digital images grows and tagging, the currency of pick-up, becomes 'richer' (for that read 'over-tagged'), the search experience deteriorates. I caught up with photo industry analyst Dan Heller while in NYC recently and he suggests that "keyword pollution" as he calls it is directly linked to user exasperation and abandoned search sessions.
Likewise, our message to those in the business of serving up photos, for commercial gain or otherwise, is; check your search logs and the percentage of searches that result in a transaction / download (whatever constitutes a successful search outcome). And what would be the financial impact if you managed to increase that percentage by a few points, bearing in mind there is negligible variable cost here? Check also the search terms used in those abandoned sessions and assess the likelihood of you having that content, and now consider the depressing probability that you had that content the user wanted, its just that they weren't able to find it, or weren't prepared to waste time trying to find it.
Successful retailers obsess about maximizing sales from footfall. Perhaps photo search sites need to be similarly fixated on how they use leading edge technology to improve their user's search experience to then benefit from the upside that site loyalty generates.
Anyway, more on subjects like that to follow. Besides this blog, we have also joined Twitter and will be engaging with the community in many other places as well. Please stay tuned for more links and projects, and we will see you on many of your blogs and groups!
Now, I would like to turn this blog over to you. What are your thoughts on image search as it stands today and what would you like to see from image search technology?
 |
 |
|
Dinner with some of the guys in NYC |
Dermot and Dan with new hardware, post PhotoPlus Expo |