The overall retail sales number is frequently used as a barometer of consumer spending and its role in the economy. Up and to the right = good. Down and to the left = bad. You get the picture.

But what exactly is included in that number? Metrics are worthless if you don’t understand what they mean.

Let’s unpack this via the July retail sales number. 

$682.8 billion

☝️That’s the July number. It’s also, more or less, the June number. 

So no change to the top-line number means consumer behavior was basically unchanged, right? Wrong.

The main reason for the lack of growth, according to The Street, was a decrease in spending on gas. People spent less on gas? Sounds like that could be bad. It’s not that people bought less gas, it’s that they paid a lot less for the gas they did buy. Gas prices hit record highs in June and have been comingdown since. 

If fuel spend dropped, then other sectors had to increase to keep the month-to-month number flat. Sounds promising.

Based on the advance (and adjusted (controls for seasonality)) calculations from the Census Bureau, the losers were cars, gas, clothes, and department stores. And the big winner was Nonstore Retailers (a lot of stuff, but looking mostly at ecomm here (Hello, Prime Day)). 

And retail sales does not take into account purchasing power. A grocery survey shows most people are paying more to get less. (It also shows that 100% of respondents grocery shop in-store.) This is leading behavior changes: fewer impulse purchases, switching to cheaper brands, and seeking deals via discounts and coupons more. While the retail sales number tells us how much people spent (on basically everything except energy and financialized products), it doesn’t tell us how much they bought. Right now that likely means: same spend, fewer goods.

I want to end with some stuff from Ben Evans, because I like the way he thinks. And because he knows more than I do on this. And because he has pretty charts. If you had the thought while reading the above, wait, gas is retail?, then you are like me. I don’t think of purchasing gas as making a retail purchase. Of course, it makes sense, because what else would it be? But when I’m wondering how the retail market is doing when looking at a client’s numbers, I really don’t care about how fuel sales are doing (I mean, I do on a macro level, but not when trying to figure out if discretionary spending is taking a hit across the board, or maybe just in their sector, or maybe just for them, for example). 

So how would it make sense to carve up this retail pie? Especially if we’re trying to focus on ecomm. Or not. This is where we come back to Ben Evans and his charts. What if instead of focusing on what a company sold, we focused on the logistics model used to sell it?

Amazon can (theoretically) ship you brake pads, a box of screws, a pair of shoes, a jar of chili crisp, a tub of AG1, and a bath mat the same way. They could even grab you lunch on the way. But they can’t ship you a case of frozen falafel without some adjustments. Same goes for a tank of gas, or the car that tank is installed in (though the car is much closer to working). So how do we categorize them under the current classification system?