PriceMirror

A browser extension that logs the price you were shown, flags the signals that could have shaped it, and shows you what price a differently-tracked shopper sees for the same item.

Problem

On January 27, 2026, California Attorney General Rob Bonta opened an investigation into "surveillance pricing" at grocers, retailers, and hotels — the practice of setting a price per shopper based on location, browsing history, device type, and even whether an item sat in your cart overnight. The announcement cited a 2025 Consumer Reports study of Instacart that found the same grocery item priced up to 23% apart for different customers, sometimes by only a few cents, sometimes by $2.56. None of that is visible from inside a single checkout. A shopper sees one number and has no way to know if it is the number everyone gets or the number their own data produced.

Target user

Price-sensitive online shoppers, 25 to 55, who do repeat grocery delivery, book travel, or shop the same retail sites often enough to have noticed a price move between visits and wondered why. Job to be done: tell me, right now, whether this price is tied to me specifically, and show me what it looks like to someone the site knows nothing about.

MVP scope

• Local price-history log: for every product page visited, the extension timestamps and stores the price shown, entirely on-device, so a return visit surfaces "you saw $6.79 here on Tuesday, it's $7.99 now."
• Signal flags drawn from the FTC's 6(b) surveillance-pricing study: logged-in vs. logged-out state, known ad-tech/personalization cookies present, a device-type switch since the last visit, and a cart-abandonment pattern (item sat unpurchased 24+ hours before this price check).
• One-click "open clean": relaunches the same product URL in the browser's built-in guest or incognito profile so the user can compare prices side by side without manually clearing cookies.
• Opt-in price ledger: users who turn it on submit an anonymized {product URL hash, price, timestamp, zip3} record to a shared feed, so the extension can show the range other opted-in users saw for the same item in the same rough place and time window.
• Evidence export: a dated PDF pulling together a user's own price history, triggered signal flags, and any ledger spread — built to attach to a state AG complaint or a card-issuer dispute.
• MVP covers the three verticals named in the CA AG's inquiry letters: grocery delivery, hotel and travel booking, and general e-commerce.

Monetization

Freemium. Free: unlimited local price history and signal flags on any site. Paid at $4.99/mo: the opt-in price ledger (which needs server-side aggregation), the evidence-export PDF, and unlimited one-click clean-profile checks.

Why now

The CA AG's sweep is aimed squarely at the three verticals this extension targets, and it leans on a CCPA theory — pricing built on data in ways a consumer wouldn't reasonably expect — that a second state can borrow without writing new law. The FTC's own 6(b) study, ordered from firms including Mastercard, Accenture, and McKinsey, already confirmed the underlying data practice: location, browsing history, mouse movement, and abandoned carts feeding individualized prices. Regulators have named the problem; nobody shipped consumers a way to see it happening to them.

Risks & open questions

• Will people who suspect personalized pricing actually check before they buy, or only after the bill arrives — by which point the browsing session, and any evidence, is gone?
• Client-side detection can't reliably tell surveillance pricing apart from ordinary demand-based pricing (time-of-day surge, low inventory). A wrong flag on a normal price swing burns trust fast.
• The opt-in ledger needs real density per SKU and region before its price-spread numbers mean anything — a cold-start problem any crowdsourced feed has to survive.
• Telling users to check a price in a "clean" browser profile is one manual click, not automated scraping, but it sits close enough to anti-bot language in retailer terms of service to warrant a lawyer's read before launch.
• Reading prices reliably off three different site families — grocery delivery, travel booking, general retail — without a brittle per-site scraper is a bigger build than a single-vertical tool.

Next step

Run it across 10 real shopping sessions in the three target verticals and log how often a same-item, same-day price actually differs by session state before building the ledger.

Sources

• https://oag.ca.gov/news/press-releases/data-privacy-day-attorney-general-bonta-focuses-surveillance-pricing-compliance — CA AG's January 27, 2026 surveillance-pricing investigation, target industries, CCPA theory, and the Instacart/Consumer Reports price-variation finding.
• https://www.ftc.gov/news-events/news/press-releases/2025/01/ftc-surveillance-pricing-study-indicates-wide-range-personal-data-used-set-individualized-consumer — FTC 6(b) study on the data signals (location, browsing history, cart abandonment, device behavior) used to set individualized prices.