July 18, 2026
Checkout lane chaos
Funny item co-occurrences in 3.2M Instacart orders
Instacart’s weird cart hunt sparked laughs, eye-rolls, and a mini privacy side-eye
TLDR: A writer sifted through 3.2 million Instacart orders looking for bizarre grocery item pairings, but the comments quickly turned into a fight over whether the results were hilarious or just making fun of shoppers. The biggest takeaway: people had stronger opinions about the jokes than the data.
A deep dive into 3.2 million old Instacart orders was supposed to answer one deliciously nosy question: what are the strangest things people buy together? The results coughed up odd couples like bananas with fancy mixed nuts, avocados with baby cucumbers, and the now-infamous vibe of “Oreos and lube” from the community’s own riffing. And honestly? The comment section had way more personality than the grocery math.
That’s where the real fireworks started. Some readers were howling at the chaos, tossing out their favorite cursed pairings like “kale and an enema,” “parsley and condoms,” and “adult diapers and baby food” as if the grocery aisle had turned into open-mic night. But others were not amused at all. One blunt response basically stole the show: “Which one of these was supposed to be funny?” Another called the whole thing awkward and a little mean, comparing it to the time Netflix publicly clowned viewers for bingeing bad Christmas movies. Ouch.
Meanwhile, one commenter went full grocery-data historian with a passionate banana manifesto, explaining that bananas are bizarrely hard to predict and basically the main character of supermarket shopping. In other words: yes, bananas are everywhere, and yes, that may be why they keep showing up in these oddball pairings. A smaller crowd was impressed by the behind-the-scenes categorizing trick used to group products, but the mood of the room was clear: people came for weird carts, stayed for the debate over whether we should be laughing at strangers’ shopping habits at all.
Key Points
- •The article analyzes Instacart’s open dataset of 3,214,874 grocery orders to find unusual product combinations.
- •The dataset includes about 10 products per order on average, 49,688 unique products, and 134 aisles.
- •Ranking combinations by least-common occurrence was not useful because roughly 97% of possible product pairs never appear and about 22 million observed pairs appear exactly once.
- •The author switched to using lift, defined as observed frequency divided by expected frequency, to identify unexpectedly rare co-occurrences.
- •Examples of low-lift results include Banana with Extra Fancy Unsalted Mixed Nuts for pairs and Banana, Bag of Organic Bananas, and Soda for triples.