Using ML in business raises many considerations outside of the normal content of ML textbooks.

A recent post from the Ali Wytsma at Shopify, attempts to encode some of their learnings about applying this technology. Some tips:

• Should we use ML? Often it is not what is needed.
• Keep it simple. If you do use ML, use the most simple methods you can. EG a linear model with a small number of features.
• Measure before optimising. Really understand the business context and how it relates to predictive accuracy before you start optimising.
• Have a plan to iterate. Set up infrastructure to allow iterations early in a project.
• Consider model stability. Keep an eye on things because models stop working over time.

Overall, a nice set of practical guidelines from Shopify.