More than a billion units of grocery products a month. That is how much scanner data the UK's statistics office says its grocery datasets cover. The number comes from a country that is not the United States. The US sources gathered for this piece say much less than a reader might hope, so this article is careful about what it claims.

What the CPI is, in plain terms

The Bureau of Labor Statistics (BLS) publishes the Consumer Price Index (CPI). It is a monthly measure of how prices change for a basket of goods and services that households buy. It does not track one shopper's bill. It tracks the change in prices over time for many items, then combines them into a single index.

That design explains why collection matters. If the prices going in are stale, mismatched or incomplete, the index drifts from what people actually pay. So the real question is not only what the CPI says. It is where each price comes from, and how the same product is followed from one month to the next.

The BLS pages gathered here are mostly menus and headers. They point to sections on data sources, design and frequently asked questions. The text that would describe those sections did not come through, so this piece cannot report US collection methods.

What the available US sources do and do not explain

One BLS notice is titled "Update regarding data collection for the Consumer Price Index." Only the page furniture was captured. We cannot say what it changes or when. A 2024 Monthly Labor Review article on alternative data sources for high-tech products also appears. Its title suggests the BLS has looked at non-traditional data for some goods. The captured text does not say which sources, or how they feed the index.

Treat all of this as a pointer, not a finding. A title tells you a topic exists. It does not tell you what the agency decided.

How other statistical offices collect prices in stores and online

The detailed descriptions come from the UK Office for National Statistics (ONS) and Statistics Austria. Neither describes the US system. They show the general approach some offices take, and nothing more.

Scanner data are records created when shoppers buy goods at a supermarket checkout or online. They show what customers actually paid. ONS says personal customer information is not in the data feeds it receives. Statistics Austria gathers prices in shops, and also by email, by telephone and over the internet. It also uses automated web scrapers, sometimes called webcrawlers. These are programs that visit a web page and copy set details, such as product description, quantity, size and price, into a structured table.

The main ideas from those two offices:

  1. Scanner data capture real transaction prices, not just shelf tags.
  2. Web scrapers pull listed prices from retailer websites.
  3. Collection by visit, phone or email still appears alongside both.
A grocery receipt and a folded rent renewal notice lie side by side on a kitchen table in soft window light, the everyday prices a consumer price index tracks month to month.
A shopper's grocery receipt next to a rent renewal notice, the kinds of everyday prices a consumer price index tries to track from month to month.

Scale, automation and the matching problem

ONS says scanner data let it automate more of its price collection. It can widen what its inflation figures cover and draw on far more product prices than manual collection allows. It planned to bring grocery scanner data into its consumer price statistics from March 2026, after a parallel run of about a year. It published first indicative estimates on 29 April 2025. A parallel run means the old and new methods are calculated side by side before the switch.

Statistics Austria's CPI time series go back to 1958. The office says its price index tradition stretches back more than a century. It writes its web scrapers in-house in R and Python. The scripts run inside a framework that sends automatic monitoring emails and reports, so small website changes are noticed and handled quickly.

Bigger data is not automatically better data. A price index compares like with like. If a cereal box shrinks from 500 grams to 450 grams, the shelf price may stay the same while the real price per gram rises. If an item vanishes, someone has to decide what replaces it. Scanner feeds and scrapers still need rules for those cases. The sources here do not spell those rules out, so this piece does not either.

Open questions about US price collection

Several basic questions stay unanswered in the captured material:

  1. Which US retailers and store types are sampled?
  2. How are online prices chosen, and how often are prices collected?
  3. Does the US use scanner data, web scraping or both?
  4. What does the 2026 BLS data collection notice change?

The Eurostat document and the 2024 high-tech article could not be read either. So this article cannot say how US CPI prices are collected.

Readers who want the answer should go to the BLS directly. The data sources chapter, the CPI frequently asked questions and the data collection notice are the authoritative pages for US methods. This is general education, not advice for any one household's situation.