Using data-driven decision making to reduce risks

Normally this weekend I'd be planting tomatoes and other warm veggies in the garden.

But this week we had a record-setting snow storm for the area. Not just for how much, but how late it was.

My DIY greenhouse took some damage but most of the plants seem fine with just a few chilled leaves.

Not for the neighbors who planted early though. I've seen some lost plants and some that were hastily covered up (and probably will have to stay covered for a week).

I avoided all of this because I use data to help decide when to plant.

This weekend is the 10% last frost date. That means in the historic records, there's only been frost after this date in 10% of the years. Pretty safe.

If I planted last month, say March 14th, then the data says there would be a 70% chance for a frost. Pretty unsafe.

I use 10% because I'd rather not have to rush out in a cold night to cover up plants and risk. I reduced my risk in exchange for less growing time.

Even then, there's a chance of a freak event like what happened this week (10% isn't 1%).

This is an example of what data-driven decision making means. Not just collecting data, but analyzing parts of it and making informed decisions.

You could use Repeat Customer Insights to automatically analyze a bunch of your Shopify data to find how your customers behave.

With multiple models and over 150 customer segments applied automatically, it gives your store the analytics power of the big stores but without requiring a data scientist on staff to read the tea leaves.

Eric Davis