Where Receipt Panels Fall Apart in the Real World

Let’s say you’re trying to understand market share for a brand of battery-powered zero-turn mowers (ZTRs) sold at a specific retailer. You want to know how your product is performing:

• At a specific retailer 
• For a specific 42-inch deck size
• Within a certain price band
• In Q2, post-promotional period

In a receipt panel, each of those variables shrinks your sample. And as they compound, the data evaporates. By the time you’ve layered in brand × retailer × deck size × power type × timeframe, you’re often looking at a handful of receipts or none at all. The result? Asterisks. Missing data. Or worse, a false sense of confidence through precision.

Let’s walk through a real example using receipt data. This data shows all receipts for a major retailer in the US for March 2025. You’ll see how the sample thins as you add more variables. Bottom-mount fridges represent 0.0011% of the total receipt sample.  Even quarterly, this would total 228 receipts.

“This real data example illustrates how receipt volumes quickly dwindle as you drill into more granular cuts of the data. That’s exactly why we built a mixed modeling approach that adjusts by category and product type. In many cases, it’s not possible to measure durable goods performance at the level our customers need based on receipts alone.”
— David Fernie, VP of Data Science at OpenBrand 

Now imagine applying that same methodology to categories like:

• Pesticides and seed treatments (seasonal, often regional, mostly in-store)
• Small appliances (broad SKU variety, often gift-driven, with sharp regional sales patterns)
• Outdoor grills (highly seasonal, high-ticket, and influenced by regional trends)
• Power washers (niche but competitive, often driven by specs and seasonal demand)

These aren’t edge cases. In a receipt-based model, effectively every durables category becomes an edge case.

Receipt-based panels alone simply don’t have the depth or structure to measure them with the fidelity you need to make confident decisions.

Why Modeled Market Share Works Better

Durables are different. Which is why we built a different model—one designed for long purchase cycles, low-frequency transactions, and complex category structures.

1. Optical Web Collection with AI Processing, Not Scraping

Our proprietary optical web collection mimics human browsing behavior. It captures cleaner, more complete product and pricing data, without relying on brittle scraping templates that break every time a webpage updates, delivering the most accurate, dependable, and robust web collection data.

2. Multi-Source Modeling for Confidence and Coverage

Our modeled market share is built on a robust foundation of diverse, high-fidelity data sources:

• Transaction Data: Accounting for more than $150 billion in transactions
• Optical Web Collection: Spanning hundreds of retailers and over hundred of thousands of unique part numbers
• Consumer Tracking Survey Data: Hundreds of thousands of annual responses, demographically and geographically balanced to the U.S. Census
• Clickstream Records: Based on tens of millions of real-world digital journeys—tracking search, consideration, and engagement
• In-Store Collection: Monthly audits across dozens of metro markets, across top national and regional retailers tracking inventory, pricing, and placement

Many durable categories, especially low-incidence or long-tail products, are sparsely represented or completely missing from receipt panels. That’s why we deploy a category- and product-type-aware mixed modeling approach. Our system dynamically adjusts the weight of each data source based on the availability and strength of signal, ensuring accuracy even when receipt data is limited or absent.

This enables us to deliver SKU-level market share with statistical confidence, even in the parts of the market where traditional methods fail to produce usable insights.

3. Designed for Depth, Not Just Direction

With OpenBrand, you can analyze:

• Retailer and brand share trends 
• Estimated gross units and dollars by brand, retailer & SKU
• Mix shifts between product attribution (like battery vs gas), or premium and entry-level SKUs
• Performance by spec, like deck width, fuel type, or horsepower
• Data is completely unmasked

Built for Durables, Not Wall Street

Here’s the truth: most receipt-based systems were initially built for hedge funds and then were repurposed for commercial use. They prioritize fast release company-level insights over accuracy and product depth, and often serve up directionally valid, but shallow data. Meeting the needs of manufacturers and retailers is the core of our business, and has been for over 20+years. 

At OpenBrand, we’ve built a market measurement model that’s:

• Grounded in industry-specific taxonomies
• Backed by a team that knows the products and the retailers
• Able to handle the long tail of durable goods, not just high-volume SKUs

This isn’t market share guessing. This is precision data for decision-makers to ensure better outcomes.