The Ecomm Analyst

Last spring I did something I had been putting off for two years. I opened my Shopify billing settings and added up what I was paying every month for apps. The number was higher than I expected. Then I opened the apps themselves and checked which ones I had actually used in the last 30 days.

About half had not been opened. A few I had completely forgotten installing. One was billing me $49 a month for a feature I thought was native to Shopify.

The audit took a Saturday morning. The savings ran into four figures annually.

Here’s the framework I now run on my own store and on every client account. There are three questions for every app. Have I logged into this in the last 30 days? If I uninstalled it tomorrow, would something measurable break? Is the value it produces greater than what I’m paying?

If an app fails any of those, it goes. Not “I’ll think about it.” It goes.

The hardest category is what I think of as insurance apps. You install them because something bad happened once or you read a blog post saying you should. Inventory forecasting tools, fraud detection layers, abandoned cart recovery on top of Klaviyo, these all live here. They feel responsible. They mostly aren’t doing anything you couldn’t replicate with the tools you already have.

The easiest category is what I call vanity apps. Wishlist builders, stock countdowns, recently-purchased popups. These got installed during some growth phase and never got measured. When I A/B test them off, conversion almost never moves.

The category worth keeping is what I think of as actual leverage. Klaviyo, my reviews app, my attribution platform, my support and loyalty tool. Things that touch revenue or customer experience in a measurable way and that I would rebuild from scratch tomorrow if I had to.

A clean stack is not just about saving money, though that’s a nice side effect. It’s about reducing the cognitive load of running a store. Every app is a tab you might need to open, a setting you might need to remember, a place a problem could be hiding. Cutting an app you don’t need is one of the rare maintenance tasks that makes everything else easier.

Once a quarter I redo the audit. Apps I evaluated last time get re-evaluated. Anything new gets added with an explicit 60-day review date in my calendar. If it doesn’t earn its keep by then, it’s gone.