The Ecomm Analyst

Ask most operators how their store is doing and you’ll get a single number back. “We did $480K last month, up 12% year over year.” That’s the number on the dashboard, the number in the investor update, the number that drives the “good month” or “bad month” conversation in standup.

It’s also basically useless on its own.

Total revenue is an aggregate that mixes two completely different businesses. There’s the business of acquiring new customers, which is expensive, slow, and the only way you actually grow. And there’s the business of selling more to people who’ve already bought from you, which is cheap, fast, and mostly a function of how well you set things up two years ago. Lumping them together and calling it “revenue” obscures the only signal that actually tells you whether the business is healthy.

Here’s what I mean. Last fall I was looking at a brand whose top-line numbers had been steady for six months. Same revenue, plus or minus, every month. Nothing alarming on the dashboard. When we split it out, new customer revenue had dropped 38% over that period and returning customer revenue had climbed almost the exact amount to mask it. They looked stable. They were quietly bleeding out the front door while their email program kept the back door full.

You only catch that if you’re looking at the split.

The split also changes how you evaluate channels. Paid social and Google have very different new-vs-returning mixes. Email is almost entirely returning. Affiliate and influencer skew new. SMS is overwhelmingly returning. If you grade all of them on blended ROAS, you’ll consistently overpay for retention channels and underpay for the acquisition channels you actually need to grow. The first time I separated new customer ROAS from total ROAS by channel for a brand, the rankings completely flipped. Channels we’d been planning to cut were actually our best new customer engines, and channels we loved were just harvesting people email would have closed anyway.

A few specific numbers I look at, in order of how often I check them:

New customer revenue, weekly. Tracked alone, not blended. This is the closest thing to a real-time growth signal a store has. If this number is dropping, nothing else on the dashboard matters.

New customer ROAS by channel. Filtered to first-time buyers only. This is the number that tells you what to fund. Total ROAS lies to you because it’s flattered by repeat purchases.

Returning customer revenue per active customer. Are your existing customers buying more, less, or the same as a year ago? This tells you whether your retention engine is improving or quietly degrading.

Repeat purchase rate by cohort. Customers acquired this quarter, what percentage came back? Compared to last quarter’s cohort at the same age. This catches retention problems six months before they show up in revenue.

Doing this in spreadsheets is technically possible and practically miserable. Most platforms don’t expose customer-level attribution natively, so you end up reconciling Shopify orders against Meta ad data against Klaviyo flows and trying to figure out which customers were new at the moment of which click. I used to do this in Sheets. I don’t recommend it.

These days I let ThoughtMetric do the heavy lifting. It tracks customers as customers, not as anonymous conversion events, so the new-versus-returning split is built into the underlying data instead of being something you derive after the fact. Whatever tool you pick, the important thing is that you stop treating total revenue as the headline number and start treating it as two businesses, because that’s what it actually is.