27 Oct 2013 Visual Search and Microstock

Microstock prices are so low that agencies compete more on the speed of search than on price.  Most serious buyers are happy to pay higher prices at a website with the functionality that helps them find the image they’re seeking quicker.

Visual search is one of the key tools agencies use to reduce the time customers take to find an image.  The difference is evident when you look at which agencies have visual search tools and which don’t.

Microstock Agencies with Visual Search

The most simple visual search technique is searching by color. iStockphoto, Shutterstock and Fotolia all have search-by-color functions.

Negative space – empty space in the image where copy text can be placed – is also available at iStock. They’re not alone in having this feature, though they have sought to be the only ones by registering a patent, though not without challenge.

There’s also some attempts to group similar shots, though usually only by keywords and data mined from customer behaviour. Shutterstock’s “Instant” search does this quite well, and many agencies display similar images on image detail pages based on the same metrics: keywords and past customer behaviour.

What Other Techniques Exist

These are some proven visual search technologies that are not (yet) present at any microstock agency:

  • Search by multiple colors
  • Arrange images in space by topic
  • Upload a sample image to find similars (Crestock‘s parent company Masterfile has this last function)

Enter Pixolution

Prof. Dr. Kai Uwe BarthelGerman company Pixolution is currently offering the technology for all these visual search techniques and more. It’s headed by Dr Kai-Uwe Barthel, a professor for Media & Computing at the University of Applied Sciences (HTW) in Berlin. He’s the visual search specialist that understands the stock photography industry, and has already done quite a bit of work within it.  He’s also speaking at Microstock Expo next month.

Pixolution’s latest offering is called Fusion, and as you may have guessed from the name, it fuses visual and keyword searches for better controlled results.  For a chosen image it will use the visual features, plus keywords, to find more accurate matches.  You can change the relative influence of keywords versus visual with a slider, giving you greater control.

Pixolution's Fusion search - keyword weighting Pixolution's Fusion search - even weighting Pixolution's Fusion search - visual weighting

Pixolution’s Fusion service, and other services, integrate with Apache Solr, the open-source search solution that forms the basis of many microstock agencies search functions.

What’s the Opportunity for Microstock Agencies?

While obviously not the only distinction, visual search is consistently present in the big-and-growing microstock agencies and consistently not-present in the less successful ones.  According to Dr Kai Barthel, it’s also becoming less of a distinction between traditional agencies and microstock agencies, who have historically been ahead in technology.

Visual search is clearly a critical element of success for microstock agencies.  The top agencies have already invested, although none appear to be pushing too hard of late.  Shutterstock’s ‘labs’ search experiments, for example, haven’t seen much love for some time.  The functionality available from visual search solution providers such as Pixolution is more advanced that anything at even the best microstock agencies, and so those services represent a clear opportunity for any of the top agencies to gain more competitive advantage.

The not-so-top agencies need to catch up.  Visual search is not simple, and not something that can be built in-house in a short amount of time.  So search solution providers present a clear opportunity for newer and smaller agencies to better compete.

As search is such a critical part of customer choice in microstock, it’s surprising more agencies aren’t pushing ahead more aggressively with visual search.

No Comments

Post A Comment