hack c'ville

• devin shields
• data scientist
• capital one labs

making money with data

• the economics of data science for entrepreneurs
  • how companies make money with Data Products
  • and when to recruit data experts for your venture

what is a data product?

• let's look at some examples

suggested playlists

• video on spotify's engineering culture

search results

• original paper on how google works

suggested products

• coursera on building recommendation engines

intuitive ui

• Fast Company article on C1's Bundle acquisition

let's talk about business

• spotify
• google
• amazon
• capital one

spotify business model

• SWOT Analysis

spotify's data feed their business

• there is a causative relationship between usage and probability of payment.
  • make it easy for users to find the right music and they'll keep paying for spotify
  • very similar to the Netflix Prize model and movie queue length
• data exhaust from user listening habits has tremendous value
  • THOUGHT EXPERIMENT: given someone's listening history, what could you price more accurately?

google business model

google's data feed their business

• google uses data exhaust from user search behavior to measure search result quality
  • the search engine with the most users gets the most data and can improve the fastest
  • yahoo patent on search results metrics

amazon business model

amazon's data feed their business

• purchase history lets amazon recommend better products to their customers
• data exhaust from websites, apps, and custom hardware let them:
  • optimize UI design and user experience
  • accurately price price new products and size new markets
• recommendation engines are not cheap

capital one business model

capital one services

• banking APIs (accounts, credit, deposits, etc.)
• user interfaces (web, mobile, etc.) that wrap these APIs
• credit decision engine (unique to banking & financial services)
• NOTE: banking and financial services have weird business models

capital one's data feed its business

• fraud reports are used to discover bad actors & data breaches
• transactions (data exhaust) predicts credit behavior in the future
• "similar" users exhibit (on average) similar credit trajectories

let's talk about you all

• what businesses has hack c'ville invested in?
• what data do you have access to?
• how could you use this data to make or improve your business?