No P&L has a line called catalog. The money leaks anyway.

Field notes from following four leaks to their line items and finding the same cause under all of them.

We have never seen a P&L with a line called catalog. Not once, in 28 projects. The money still leaks; it just signs the guest book under four other names.

We know because we kept following it. Every audit starts with a symptom that has a respectable owner: conversion is soft, the site feels slow, returns are up, payroll grew. Four symptoms, four departments, four meetings that never mention product data.

For a long time we walked into those meetings with the same wrong answer as everyone else. Conversion is soft, so look at the ads. That answer burns a quarter and finds nothing (we ran that particular loop more than once).

The alias called marketing

Google commissioned Deloitte to measure what raw speed does to a retail funnel: 37 brands, over 30 million user sessions. A 0.1 second improvement in mobile site speed lifted retail conversion by 8.4 percent.

Not a redesign. A tenth of a second.

Average order value moved 9.2 percent in the same direction. Meanwhile, when conversion goes soft, everyone stares at the ad account, and the timer keeps eating quietly under the floorboards.

The alias called tools

So where does slowness come from? Mostly from code nobody remembers choosing. The 2024 Web Almanac found third-party code on 92 percent of the pages it crawled, and a Shopify store wears that number as apps.

Shopify grades storefront performance where the money moves: product pages carry 40 percent of the score, collection pages 43. That puts 83 percent of the grade on the two surfaces where the catalog lives. In Shopify's own worked example, installing a single app drops a storefront five points.

The app subscription is small and sits visibly on a statement. The weight it adds to the product page is invisible and gets billed, through leak one, to marketing.

The alias called logistics

Returns get booked as shipping, restocking, warehouse hours. NRF and Happy Returns expected American consumers to send back $849.9 billion in merchandise in 2025.

Online, the return rate runs at 19.3 percent, roughly one order in five turning around at the door. And in apparel the reason is rarely a mystery: Coresight and Alvanon found that nearly 70 percent of shoppers returning clothes bought online cite size and fit.

Size charts. Measurements. Fabric behavior. That is catalog data. A return caused by a missing size chart gets filed under logistics, and the catalog walks away clean.

The alias called payroll

The fourth leak is a person. Picture the Tuesday: someone capable is copying specs from a supplier PDF into metafields, row by row, while the catalog grows faster than they type. When the store adds a brand, the fix is another hire, and the line that grows is payroll.

That is the tell. Labor scales with catalog size; infrastructure doesn't. Our own pipelines push around 5,000 products a day, and enrichment prices at $0.50-$1.50 per SKU, because a system's cost per unit falls as volume rises. A keyboard's doesn't.

One cause, four names

For years the market has treated these as four separate businesses: performance consultants, app auditors, returns software, virtual assistants. We assumed the same split. Then we noticed that every fix, in every one of the four meetings, started in the same table: the product record.

Structure the data and the page gets lighter. Lighter pages convert. Complete the fit data and the return never ships. Automate the enrichment and payroll stops tracking SKU count.

A cost with no line item never loses a budget argument. It is never in the room.

What the leak map is worth

Run the catalog as one system instead of four symptoms and the levers compound: the structured data that speeds the page is the same data that prevents the return and retires the copy-paste Tuesday. Across our client work we have documented conversion lifts of up to 45 percent. That figure is the ceiling, not the average, but the direction has never reversed on us.

What we keep taped to the wall

• The catalog has no line item, so audit it through its aliases: marketing, tools, logistics, payroll.
• Speed is a catalog metric. Most of Shopify's performance grade is set on product and collection pages.
• A return is a data-quality report you pay return shipping on.
• If cost per SKU doesn't fall as the catalog grows, you're renting labor, not running infrastructure.

We still haven't seen a P&L with a catalog line, and we've stopped expecting one. We just knock on the four doors it hides behind.

Follow the aliases.

If you'd rather have your four leaks measured than described, OKART's Infrastructure Stress-Test maps them across your store in 48-72 hours, read-only.

Sources

1. Milliseconds make millions: Google-commissioned Deloitte study of 37 brands and 30+ million sessions; a 0.1s mobile speed improvement lifted retail conversions 8.4% and average order value 9.2%.
2. Consumers Expected to Return Nearly $850 Billion in Merchandise in 2025: NRF and Happy Returns; $849.9 billion in returns, 15.8% of annual sales, 19.3% online return rate.
3. Sizing intelligence is strategic priority as brands prepare for AI-driven commerce: coverage of the Coresight/Alvanon report; nearly 70% of online clothing returns cite size and fit.
4. Storefront performance: Shopify's app performance scoring, weighted 17% home, 40% product, 43% collection, with a worked example of one app costing five points.
5. Third Parties, Web Almanac 2024: HTTP Archive data; 92% of crawled pages include at least one third party.