Archives For

There are many applications for machine learning, but one of the most exciting is using it to create hypothesis to test. An article in the Economist, Computer says ‘try this’, discusses many of the ways computers are now creating hypothesis to move medicine, farming and even cooking forward.

There are already many projects using machine learning to generate valuable and novel hypothesis. One project, BrainSCANr, suggests research topics for neuro-scientists by looking at millions of peer reviewed papers. Research published by Baylor College of Medicine researchers used machine-learning software to review over 150,000 papers on a technique to curb the growth of cancer (kinases). The algorithm led to seven new kinases that researchers had missed. The technique is even being used to analyze search terms in Bing and Internet Explorer to determine potentially dangerous pairings of medicine.

Slide1

What it means for you

While machine learning is already benefitting many tech and game companies, using it to help develop hypothesis for your business is invaluable. If you are a mobile game company, think of the value of having a machine learning system suggest that rather than focusing on your premium exchange rate, you should test changing the frequency of your free coin bonus. Or maybe it suggests that rather than testing your new content cadence you should test the efficacy of using IP. If you are an AirBnb, it may suggest that you test different rake options rather than how you sort options for users.

The power of machine learning for hypothesis testing is two fold

  1. It directs your resources to where your testing will have the greatest impact on lifetime value. While your test may show statistically significant results, a different test may have a much bigger impact on metrics.
  2. It analyzes other people’s findings, from online articles to consumer behavior research, and uses this information to steer your tests. Rather than re-inventing the wheel, you can build on what other people have found.

Key takeaways

  • A powerful potential application of machine learning is using it to determine the most important hypothesis for you to test
  • Machine learning can analyze different data points from your customers to suggest the most beneficial hypothesis to test
  • Machine learning can also synthesize research and other public information so you do not waste resources testing theories that have already been tested.

I hate writing about the next big thing because it is usually trite, clichéd or just hype, but I read a great piece in the Harvard Business Review about the collaborative economy (“Sharing’s Not Just for Start-Ups” by Rachel Botsman) that I wanted to share. Most of us have come across and probably used start-ups leveraging collaboration or sharing, companies such as Airbnb (where people share excess rooms with travelers) or Uber and Lyft (where people who need a ride can find a driver who is looking to earn extra funds. In and of itself, this is an exciting space with many promising early stage companies, from peer-to-peer lending (money club) to online lessons (Udemy). What I found interesting in Botsman’s article is how this opportunity can be extended to many other businesses.

There are great opportunities in the collaborative economy to create additional revenue streams (which may supplant their core business at some point) or provide channels for user growth. Botsman starts by discussing Marriott, the hotel chain, which rolled out an offering in conjunction with LiquidSpace so people and businesses could book excess conference rooms at market clearing prices. Not only did this initiative create a new revenue stream for Marriott by generating income from rooms that were sitting unused, it also helped grow the customer base. Continue Reading…