The Next Wave of Startups Will Focus on “Dirty Data”

“Why now?” is the gold standard question of deciding whether to invest in a company.

But these past few years, I’m starting to feel like there’s a lot of things that should have been built 10 years ago but nobody got around to it. Some of it is just difficult to build.

One huge barrier I see is data. Not big data, but clean data. One thing that everyone wants is something that compares grocery prices of different places. You want chicken, just type chicken and it shows the closest one.

Or something that evaluates land value in an area. You want to start a high end cafe, just pop open the site, and it’ll tell you where the demand for cafes is the highest.

These are restricted by data. Collecting data is difficult. You can buy data, but it may be dirty. Sometimes the data is just lying around for free, but difficult to clean.

“Dirty data” startups are also very resistant to being disrupted by AI/ML. AI learns with the data you give it. But give it the wrong data and it goes the wrong way. http://thiswaifudoesnotexist.net is a good example of what happens with bad data.

I’d say the near future of startups is either building companies of ‘dirty data’ moats and/or cleaning up and selling off the clean data to larger data-driven companies.