Leveraging Unstructured Data To Build New Products

unstructured-data

Interactions on the web are increasingly distributed and new applications must be built with a decentralized model in mind.

As web apps become increasingly ubiquitous, Apple’s expression, “there’s an app for that!” rings truer than ever. In such an environment, first mover advantage becomes increasingly important as early traction can compound quickly via network effects.

A while back Stowe Boyd wrote about building listening devices, i.e. eavesdropping tools, to find and act on data posted to our various interaction streams.

It struck me (and one of my brainstorming partners, Mike) that there is an opportunity for more sites to launch with integrated, personalized content already culled or scraped from the open web.

For example, a bot could easily scour the web and auto-aggregate content on your behalf. It could suck in your data, organize it, make it look prettier, and then alert you that a custom feed has already been created for you — all you do is claim it.

One company I consider an early pioneer in this strategy was Zoominfo. Zoominfo began scraping the web and assembling publicly available data into profiles similar to what you’d find on Linkedin. I recall Googling myself way back when and discovering that I had a profile already populated on Zoominfo. It was actually quite compelling because much of the information there was out of date it made me want to go in and update it (and I did) signing up for service along the way.

Another app that effectively leverages auto-population is Facebook’s photo tagging app. While now it’s difficult to imagine anyone not being on Facebook, the genius of the photos in the early days was to allow others to tag you. If you were not a Facebook member, you’d get an email alerting you that you’d been tagged. Once curiosity got the better of you, you’d sign-up and already have numerous photos awaiting your arrival. That’s sticky.

Note: With a few friends I’m working on a little side project. We’re hoping to use this eavesdropping technique. Assuming we do, perhaps we will make the code available for others.

LinkedIn
Twitter
Email

You might also enjoy:

Thanks for stopping by!

If you’d like to receive occasional updates and new writings from me sign-up below and never miss an update.