Hallucinated facts
Plausible-sounding citations that do not exist, fabricated dates, statute numbers that are not real.
| #▼ | Company◆ | SPX Score◆ | P / D / H / O | Flags |
|---|
A public accountability index tracking which companies use your personal data to decide your price. Same product. Same moment. A different number, just for you.
It is not surge pricing. It is not time-of-day discounting. It is your price, for who you are. Location, browsing history, device, demographic inference, loyalty data, and dozens of behavioral signals feed pricing systems that quote a different number to a different shopper for the same product, in the same moment, on the same screen.
SPX scores the fifty most consequential US companies on four public dimensions and a single composite tier from Low to Critical. Every score is traceable to a documented source. Every confidence rating is derived mechanically from the source mix, not assigned by hand. The goal is the same as climate disclosure or antitrust review: a public lens on a practice the public cannot otherwise see.
Surveillance pricing is a supply chain, not a feature. Collection, identity resolution, unification, enrichment, decisioning, surface, and closed-loop feedback. SPX scores companies at every layer.
Surveillance pricing is not theoretical. Each finding below is from a regulator, court, or peer-reviewed study. None is from the company that operated the system being studied.
Composite scores split companies into four tiers. The bar below shows how today's index actually distributes — not how a normal curve says it should.
The five companies with the highest composite SPX scores. These are the ones with the deepest combination of personalization capability, data intake, documented harm, and opacity.
See all 50 companies →Average composite score by sector, weighted by the count of companies in each. Hover or tap to see member companies.
HB 895, the Protection from Predatory Pricing Act, prohibits grocers and third-party delivery services from using surveillance data to set per-customer prices and freezes posted prices for one business day. Effective October 1, 2026. Penalties up to $25,000 for repeat violations.
RealPage agrees to stop offering software that pools nonpublic competitor data to recommend rents, ends related market surveys, and accepts a court-approved monitor for ten years. The settlement is the most consequential US action against algorithmic pricing to date.
A Mass. Gaming Commission analysis finds DraftKings and FanDuel restrict customers who win regularly, while VIP rewards flow to those who consistently lose. Behavioural targeting includes loss-chasing patterns and late-night logins.
The committee sends letters to major travel and platform companies asking for documentation of revenue management algorithms, use of consumer data in pricing, testing and experimentation practices, and internal communications describing pricing tools.
Colorado moves a bill to prohibit surveillance-based discrimination in pricing and wages. After clearing the House, the bill is voted out of a Senate committee. Effective date for bills enacted without a safety clause is August 12, 2026.
Assembly bill would prohibit retailers from setting customized prices based on personally identifiable information collected through electronic surveillance technology. Civil penalties up to $12,500 per violation, three times that amount for intentional violations.
Attorney General James backs paired bills that would prohibit personalized algorithmic pricing in New York, ban electronic shelf labels in large food and drug retailers, and create enforcement mechanisms with a private right of action.
The eight intermediaries named in the 2024 FTC 6(b) order worked with 250+ client businesses across retail, grocery, and travel. Personal data including location, demographics, and browsing behavior fed pricing systems at scale.
FTC 6(b) study, House Oversight investigation, federal legislation introduced.
Maryland HB 895, the Protection from Predatory Pricing Act, signed by Gov. Wes Moore. Effective October 1, 2026. New York, California, Colorado, Illinois pending.
RealPage settled, Live Nation pending, GoodRx and Grubhub closed, Junk Fees Rule in effect.
The index ranks fifty companies on the four dimensions. Every score is traceable.
Browse the index →Personalization depth, data intake breadth, consumer harm evidence, opacity. Filter by tier, sector, or evidence confidence. Click any row to open the full company brief with sources.
| #▼ | Company◆ | SPX Score◆ | P / D / H / O | Flags |
|---|
This is analysis, not measurement. Each dimension is scored 0–100 as informed judgment. The evidence base underneath is what makes the judgment defensible. Composite weights consumer harm 30%; personalization, data intake, and opacity at 25%, 25%, and 20%.
Surveillance pricing is the use of personal data (location, browsing history, device, demographics, inferred traits) to set an individualized price for a good or service. Not surge pricing. Not time-of-day discounting. Your price, for who you are. SPX scores the fifty most consequential companies on four public dimensions. Every score is traceable to a documented source.
This is analysis, not measurement. Each dimension is scored 0–100 as informed judgment. The evidence base underneath is what makes the judgment defensible. Composite weights consumer harm 30%; personalization, data intake, and opacity at 25%, 25%, and 20%.
How individually priced is the output? Patent filings, job postings for personalized pricing or price-optimization ML, vendor disclosures, SEC filings on dynamic pricing.
How much consumer data feeds the price? Privacy policy language, data broker contracts, app SDK inventory, loyalty program scope, biometric signals, inferred traits.
Is there documented differential pricing? FTC 6(b) inclusion, class actions, state AG inquiries, congressional letters, academic audits, investigative journalism.
How hidden is the practice from the consumer? Whether pricing factors are disclosed, whether consumers can see their price vs. a reference, arbitration clauses.
Every evidence item is tagged by source authority. Confidence is derived mechanically from that mix.
FTC orders, DOJ filings, court records, SEC filings, state AG actions, congressional letters, peer-reviewed studies.
Investigative journalism (NYT, WaPo, Reuters, ProPublica, CNBC), academic working papers, company 10-K and earnings-call disclosures.
Trade press, advocacy analysis, privacy policies, SDK inventories, industry-standard practice observations.
High: 3+ Tier-A, or 5+ total w/ 1+ A. Medium: 1+ Tier-A, or 3+ B. Low: thin file.
Surveillance pricing is not one product. It is a seven-layer supply chain that runs in the time between your tap and the number on your screen. Each layer is its own market with its own vendors, its own data, and its own incentives. SPX scores companies at every layer, because cutting any layer cuts the chain.
None of those seven appears on the receipt. The merchant you bought from did not invent the price; the price is the output of an industrial supply chain. SPX scores all seven layers because that is where regulation, journalism, and consumer pressure can actually intervene. A grocery-store ban on personalised pricing reaches Layer 06; an FTC 6(b) order reaches Layer 05; a state location-data ban reaches Layer 01. Each layer is its own opportunity for accountability.
Layer-by-layer analysis explains why some interventions work and some don't. Maryland's Protection from Predatory Pricing Act prohibits the use of surveillance data at Layer 06 inside grocery stores. The DOJ-RealPage settlement remediates Layer 05 in multifamily rentals. The FTC's Outlogic settlement remediates Layer 01 for sensitive locations only. New York's pending One Fair Price Package would reach Layers 05 and 06 across grocery and pharmacy. None of these alone reaches every layer, which is why the index tracks them all.
Not every personalized price is the result of all seven layers. A small-business retailer running an A/B price test on a single landing page is operating Layer 03 and 05 only. A telecom retention department using tenure to set a renewal offer is using a single internal database, not the full pipeline. The seven-layer framing is a worst-case taxonomy of what is technically available, not a claim that every transaction uses every layer.
Last reviewed: April 30, 2026. This map is updated when a new vendor or layer becomes consequential to the public record.
Surveillance pricing depends on a lattice of upstream vendors most consumers have never heard of. The companies on this page do not set the prices you see. They sell the inputs — identity, location, household composition, browsing behaviour, transaction history — that the price-setters consume. SPX scores most of these vendors directly; the ones that aren't in the index yet are mapped here for context.
Sources: FTC enforcement records, Gartner market guides, EPIC analysis, vendor 10-K and S-1 filings, and the Cracked Labs and ProPublica investigations on consumer-data brokers.
The layer that translates an anonymous browser, device, or hashed email into a single per-person identifier downstream systems can act on. Without identity resolution, none of the other vendors can stitch their signals together.
Approximately 2.6 billion verified global IDs. Match keys span hashed email, mobile ad ID, cookie, IP, postal address, and CRM ID. Partners directly with Acxiom, Experian, and Oracle to enrich its graph and resell to retailers, banks, and adtech.
Unified ID 2.0 hashes a logged-in email and propagates a deterministic identifier across the open web. Now adopted by Disney, Paramount, and major retail-media networks. Replaces third-party cookies for cross-site tracking and pricing decisions.
Analyses billions of transactions a year across a global consortium. Identifies returning users that wipe cookies, switch browsers, or use private mode. Sold for fraud detection but the same fingerprint identifies returning shoppers for pricing.
Formerly Iovation. Combines device intelligence with TransUnion's consumer credit file. Sold for fraud and KYC, used downstream of risk-based pricing in fintech and insurance.
The layer that adds inferred income, household composition, life events, and trait segments to the resolved identity. This is where a ZIP code becomes an inferred income range and a wallet-share estimate.
Canonical largest US consumer-data broker. Thousands of attributes per record across hundreds of millions of US adults. Real ID feeds the LiveRamp graph; data is licensed to retailers, financial-services firms, and adtech intermediaries.
Same category as Acxiom: demographic, financial, and behavioural attributes layered on Experian's consumer credit file. ConsumerView and Mosaic segments are the productised audiences. Sells onboarding and activation services parallel to LiveRamp.
Built from acquisitions of Datalogix, BlueKai, and AddThis. Oracle announced the wind-down of its advertising business in 2024. The historic data-broker function has migrated to LiveRamp and other identity-resolution platforms.
The layer that captures where the device is. Location is one of the strongest pricing signals available and is collected through SDKs embedded in third-party mobile apps. The consumer typically does not know the SDK is there.
Subject of the FTC's January 2024 settlement — the first against a data broker over sensitive location data. Banned from selling location tied to medical facilities, religious organisations, and other sensitive sites. The Cyber Security Location dataset was publicly listed at $240,000 per year.
Pioneered point-of-interest data and now monetises mobile location via the Pilgrim SDK embedded in 500+ third-party apps. Adopted voluntary guidelines on sensitive locations after the Outlogic settlement, but remains a primary US location source.
SafeGraph supplies points-of-interest and visit data; Veraset is its data-licensing spinoff for bulk commercial and government sales. Mobility data tracks where populations of devices move at city scale.
SDK installed in 500+ apps with insights on 20 million-plus devices, per public marketing. Used for retail-store visit measurement and competitive analysis. Customers include landlords, retailers, and equity researchers.
Italy-founded location-data firm operating in the US market. Voluntarily restricted sensitive-location use after the FTC's broader 2024 enforcement.
The CDP is where first-party retailer data meets the resolved identity and the third-party append. The output is a real-time per-shopper profile that the price-setter or recommendation engine reads on every page request.
Real-time profile unification plus Einstein decisioning for predictive offers, recommendations, and price suggestions. Pricing scales as a percentage of gross merchandise value, aligning Salesforce revenue with executed pricing decisions at retailer clients.
Adobe Target runs A/B price experiments at enterprise scale. Sensei AI generates per-segment offer optimisation against Adobe Analytics signals.
Acquired by Twilio in 2020. Pipes events from web and app SDKs into a unified customer record and routes to downstream tools. Heavy adoption in mid-market e-commerce and SaaS.
Acquired by Rokt in 2025. Strength in real-time mobile event streams. AI capabilities expanded via Indicative and Vidora acquisitions.
Started in tag management; expanded into CDP with Tealium Predict ML. Strong enterprise footprint in financial services and retail.
The closed-loop layer. Same identity used to set the price is used to attribute the eventual purchase, refining the model for the next user. This is where a lot of the new growth in the surveillance-pricing economy is happening.
Kroger's wholly owned analytics arm. Loyalty-card-tied purchase data plus app, location, and inferred-trait signals. Powers Kroger Precision Marketing; the August 2024 Warren / Casey congressional letter applies to the data-monetisation infrastructure 84.51° operates.
Targets 150 million weekly shoppers across 4,600 stores plus e-commerce. Mapped to SKU-level. Sold via The Trade Desk DSP. The targeting graph is the same graph that pricing decisions read against.
Target's in-house retail media network. Built on Circle loyalty data. Targeting and attribution feed back into Circle pricing decisions.
35-plus-year intermediary between retailers, CPG brands, and shoppers. Consolidates POS, eCommerce, CRM, and media engagement into a 360° real-time shopper profile available for targeting and personalisation.
Coupon-network operator turned data broker. Decades of household-level grocery purchase history. Often paired with Inmar in grocery promotional pipelines.
Productises card-spend data into retailer-facing personalisation and pricing analytics. Named in the FTC's 2024 6(b) order on surveillance pricing intermediaries.
This is not a complete inventory of the consumer-data broker industry. It is a map of the vendors that are load-bearing for surveillance pricing specifically — the ones whose products an SPX-tier company would call into during a price decision. EPIC, Cracked Labs, the FTC's 2014 study, and the Privacy Rights Clearinghouse maintain broader inventories. The Vermont and California data-broker registries are also useful for working the upstream supply chain on a state-by-state basis.
Last reviewed: April 30, 2026. New vendors and category shifts are added as the public record warrants.
Surveillance pricing is now an active enforcement area at the federal level, an active legislative area in at least a dozen states, and an active litigation area in housing and live events. This page is a snapshot of where regulation, legislation, and litigation actually sit. It is not legal advice. Dates are public-record. If a status has shifted since the last review, the AI Accountability log will note the correction.
July 23, 2024 — present. The Federal Trade Commission issued compulsory orders to eight intermediary firms (Mastercard, JPMorgan Chase, Accenture, McKinsey, PROS, Revionics, Bloomreach, Task Software) requiring disclosure of how consumer data feeds individualized pricing systems sold to retailers. The January 2025 staff perspective found these intermediaries worked with 250+ client businesses across retail, grocery, hospitality, and travel.
In April 2026 testimony before Congress, FTC leadership confirmed staff work on surveillance pricing continues and the agency is assessing whether additional disclosures may be required when pricing is highly personalized or driven by consumer data.
March 5, 2026. The House Oversight Committee formally launched an investigation into the use of surveillance pricing. Letters were sent to major travel and platform companies requesting documentation of revenue management algorithms, use of consumer data in pricing, testing and experimentation practices, and internal communications describing pricing tools and outcomes.
Representatives Greg Casar and Rashida Tlaib introduced legislation that would ban companies from using AI or personal data to set individualized prices or wages. The bill specifically prohibits airlines from adjusting prices after detecting a search related to sensitive personal circumstances such as a family death.
S8616 / A9396 “One Fair Price Package.” Backed publicly by Attorney General Letitia James on March 16, 2026. The package would prohibit personalized algorithmic pricing in New York, ban electronic shelf labels in large food and drug retailers, and create enforcement mechanisms with a private right of action.
AB 2564 (Feb 20, 2026). Would prohibit retailers from engaging in surveillance pricing — defined as setting customized prices based on personally identifiable information collected through electronic surveillance technology, including data acquired from third parties. Civil penalties up to $12,500 per violation, three times that amount for intentional violations. The California Attorney General launched a separate sector inquiry on January 27, 2026 covering grocery, hotel, and large retail surveillance pricing.
HB 25-1264. Prohibits surveillance-based discrimination against consumers and workers through automated decision systems that use surveillance data to inform individualized prices or wages. Passed the House; voted out of a Senate committee in April 2026. Effective date for bills enacted without a safety clause is August 12, 2026.
Protection from Predatory Pricing Act. Signed into law. Provides a baseline state-level prohibition on surveillance-driven price discrimination and a foundation for downstream consumer-protection enforcement.
More than 40 related bills covering surveillance, algorithmic, and personalized pricing have been introduced across more than two dozen states. Active jurisdictions include Illinois, New Jersey, and several others. Trade associations track this as the fastest-growing state-level consumer-protection front in 2026.
August 23, 2024 (complaint) → November 24, 2025 (proposed settlement). The Department of Justice and a coalition of state attorneys general alleged that RealPage's algorithmic rent-pricing software pooled nonpublic competitor data to coordinate rents across landlords. The proposed settlement requires RealPage to stop offering software that uses nonpublic competitively sensitive data shared among landlords, prohibits market surveys used to gather nonpublic competitive intelligence, requires retraining of models on compliant datasets within 180 days, and subjects the company to a court-approved monitor.
RealPage did not admit liability. The settlement must still be approved by a judge.
May 23, 2024. DOJ and 30 state attorneys general filed an antitrust complaint against Live Nation/Ticketmaster alleging that the company maintains monopoly power partly through pricing practices. The complaint is the highest-profile antitrust action in any direct consumer pricing market.
February 1, 2023. First-ever enforcement action under the FTC's Health Breach Notification Rule. GoodRx paid a $1.5M penalty for sharing health-related information with advertising platforms in ways that informed personalized offers. Established that health-adjacent surveillance flows can be reached under existing federal authority.
December 2024. $25M settlement covering deceptive pricing and undisclosed delivery fee practices. Settlement remedies imposed disclosure requirements that target the fee-stack opacity at the heart of the platform's consumer-facing pricing.
December 2024. Final rule on undisclosed fees in live-event ticketing and short-term lodging. While not a surveillance-pricing rule per se, it directly constrains a major opacity vector that surveillance pricing operates inside.
Last reviewed: April 29, 2026. If you spot a status change, file a correction and it will be added to the AI Accountability log.
SPX is built only on public sources. Every score is traceable to one or more documented references on the company’s detail page. The list below aggregates every distinct source cited across the index, grouped by source tier. 60 unique sources are in active use as of April 2026.
If you have a primary or reputable source we have missed, the index is designed to be updated. Email Anna and your evidence will be added to the next release.
FTC orders, DOJ filings, court records, SEC filings, state AG actions, congressional letters, peer-reviewed studies.
Investigative journalism, academic working papers, company 10-K and earnings-call disclosures.
Trade press, advocacy analysis, privacy policies, SDK inventories, industry-standard practice observations.
Every score, label, and source citation in this index was reviewed by a human before it shipped. The errors logged below were caught during that review. They are kept in public for the same reason climate scientists publish their model uncertainty: the only way to trust a number is to know how it was made.
If you find a number on this index that looks wrong, that is the experience this page is supposed to make possible. Tell us, and you will be added to the log.
An “error” in this log is anything an AI assistant produced that would have shipped wrong if a human hadn’t caught it. That includes:
Plausible-sounding citations that do not exist, fabricated dates, statute numbers that are not real.
An evidence item credited to the wrong outlet, a regulatory action assigned to the wrong agency or year.
A dimension score that does not match what the cited evidence supports, or a tier assignment that ignores the evidence mix.
A status that was true at the time of writing but has since been overtaken by enforcement action, settlement, or company disclosure.
A trade-press or advocacy item miscoded as Tier A, lifting confidence higher than the public record warrants.
Companies dropped from a filter without explanation, deep links that point to the wrong record.
This log starts with the build of v0.4. Pre-launch errors caught during prototype iteration are not retained — the policy of “write it down before fixing” starts now.
The first build loaded company data via fetch(‘./data/companies.json’). On the public site this was fine; previewing the file directly from disk in Chrome triggered a CORS block and the table rendered the empty-state error. Fixed by inlining the dataset as a <script type="application/json"> block inside the page so file:// review works the same as the deployed site.
The handoff prototype shipped with a warmer cream background (#f5efe3) and a slightly cooler forest green (#0f3b2d) than DCSI uses. Reviewing the live DCSI CSS surfaced the canonical values: stone #F4F2ED, forest #1B3A2D, bronze #A8885A. SPX now matches those exactly. Reason the prototype was off: the handoff doc claimed brand-correctness without verifying against the live sister tool.
v0.3 masthead linked to Home, DCSI, SPX, and Newsletter. Cross-tool links belong in the footer; the masthead is for sections of the current tool. Replaced with The Index, Methodology, Data Sources, AI Accountability, About.
Surveillance pricing scores fifty companies on whether your data is setting your price. Every score is traceable to a public source. No corporate funding.
The Surveillance Pricing Index answers a question that should be simple: when you pay a price online, in an app, or at a shelf, is that price the same one your neighbor sees? SPX scores the fifty most consequential companies on four public dimensions — personalization depth, data intake breadth, consumer harm evidence, and opacity — and combines them into a single tier from Low to Critical. Sister tool to the Data Center Stress Index.
Each company is reviewed across four dimensions, scored 0–100 based on public record. The composite score weights consumer harm 30%, personalization 25%, data intake 25%, and opacity 20%. Every score has an evidence base attached. Every evidence item is tagged Tier A (primary, e.g. FTC order, DOJ filing, peer-reviewed study), Tier B (reputable, e.g. major investigative journalism, 10-K disclosure), or Tier C (secondary, e.g. trade press, advocacy analysis, privacy policy). Confidence is then derived mechanically from the tier mix, not assigned by hand.
SPX does not measure the prices a given consumer paid. It does not generate legal findings. It does not name individuals at the companies it scores. It informs judgment. It does not replace it.
SPX is a project by Anna R. Dudley. It is a sister tool to the Data Center Stress Index. The entire tool is built using publicly available data and is published under a Creative Commons Attribution license. The purpose is to put the same kind of accountability lens on surveillance pricing that is already routine in environmental disclosure and antitrust.
For questions, speaking inquiries, or evidence submissions, contact Anna R. Dudley. Power moves before policy does.