This is the first entry in our score update log. Every time we push new data or change how scores are computed, we'll publish a post like this one explaining what changed and why — in plain English, with the biggest score movements named rather than buried.

This update is a large one. Since the July 2 data push we ran a full audit of the data behind every score: how federal filings get matched to companies, how parent-company spending flows to the subsidiaries you actually shop with, how violations and subsidies get attributed, and how revenue figures are sourced. We also refined how the Political Access sub-score scales with spending, rebuilt the market-share baseline on the newest Economic Census, and added five companies. Of the 157 companies on the site before and after this update, 79 scores went up, 54 went down, and 24 were essentially unchanged. The median move was about 4 points, and 45 companies changed tier. Higher is better, as always.

Companies file lobbying disclosures under many names — formal legal names, subsidiaries, holding entities. Matching those filings to the company you actually shop with is a core part of what we do, and this update expanded that matching substantially. Sixteen companies that previously showed no federal lobbying now have their disclosures attributed — roughly $80 million over the two-year scoring window, led by Verizon (~$26M), T-Mobile (~$22M), and Dell (~$14M), with fresh Q2 2026 filings included. No company's existing lobbying record shrank. Smaller newly attributed totals moved scores too: REI's Senate disclosures show it has retained federal lobbyists continuously since 2023 (about $310K over the window, moving it 93 → 77), and Logitech, Michaels, and Meijer all picked up lobbying their records previously missed.

PAC contribution matching expanded the same way: ten companies' PAC totals grew substantially or appeared for the first time. And the audit cut in the other direction too — executive donations are matched from self-reported employer text on FEC filings, and our standard for accepting those matches is now stricter. Eight companies' executive-donation totals fell by more than half because we removed matches we couldn't stand behind.

Because Political Access is 45% of the score and companies are ranked against each other, this is the single biggest driver of movement in this update — both for the companies whose spending is now visible, and for the companies that move up when the ranking gets more honest around them.

Political spending — lobbying, PAC contributions, executive donations — is frequently reported at the holding-company level rather than under the name of the retailer you actually visit. Guitar Center is majority-owned by Ares Management; Sephora is owned by LVMH. Both holding companies file their own federal lobbying disclosures, and until this update, none of that spending flowed to the subsidiaries' scores.

The purpose of this platform is to help you steer your money away from oligarchs — and every dollar spent at a subsidiary still flows up to its parent, where it can fund lobbying and PAC spending. So the score now follows the money: Guitar Center carries Ares Management's federal lobbying ($640,000 over the window) and associated executive donations, and Sephora carries LVMH's U.S. lobbying ($480,000) and executive donations. Our ownership research now runs a holding-company political-spend check for every company with a confirmed non-public parent — which also confirms affirmative zeros: PetSmart's private-equity parent BC Partners has no reportable U.S. political spending in the window, and PetSmart's record now says so explicitly.

Two things should both be true of the Political Access sub-score: any political spending should cost points, and $10,000 of it should not cost the same as $10,000,000. Because so many companies report zero lobbying, the old ranking made the first dollar of spending cost nearly as much as the last — leaving little room to distinguish modest spenders from the heaviest ones. The refined methodology separates those two jobs: crossing from $0 into any spending carries a fixed 30-point entry cost, and spenders are then ranked against each other by magnitude. Zero-spenders and the heaviest spenders are unaffected; modest spenders — Victoria's Secret, PVH, Wayfair, Harbor Freight, Scholastic — no longer score like top-tier spenders and gained double digits. The same refinement applies to PAC contributions and executive donations.

The Market Scale amplifier compares each company's US revenue to the total size of its market, using the Census Bureau's Economic Census. Until now those market totals came from the 2017 census — a snapshot taken before the pandemic reshaped retail and before the largest e-commerce shift in history. This update moves every denominator to the 2022 Economic Census, the newest complete census published.

Two visible consequences. First, category names changed across the site: the census retired labels like "Electronic Shopping & Mail-Order Houses" and now files online sellers inside the ordinary retail categories they actually compete in. Second, several companies moved to markets that describe them more honestly: CDW, WebstaurantStore, and Ferguson are measured as the wholesalers they are; Microsoft as a software publisher; Whirlpool as an appliance manufacturer; HarperCollins as a publisher.

The audit also covered our violation and subsidy records, where attribution is genuinely hard: enforcement records are published under parent companies, subsidiaries, and near-identical corporate names. Our research process had been too conservative in one direction and too aggressive in the other, and this update tightens both.

Roughly two dozen companies' records gained enforcement actions that public records support but our process hadn't yet confirmed — Sony, Shein, Five Below, Trader Joe's, Williams-Sonoma ($3.2M), Bass Pro Shops ($5.0M), Crate & Barrel, and Razer among them. A company is no longer marked clean unless a clean record is positively confirmed.

Misattributed records were removed. Aldi was carrying Trader Joe's fines, Hanesbrands' subsidy record included rows belonging to Gildan, and similar wrong-company rows came out of several other records, including Best Buy's, Meijer's, and Foot Locker's — now positively confirmed clean. An automated contamination check now catches these before they reach a score.

Because Playing by the Rules is 30% of the score and companies are ranked against each other, completing this coverage moved companies in both directions: companies whose records grew went down, and companies with genuinely clean records rose relative to them.

Market share and buyback intensity both depend on revenue, so a bad revenue figure quietly corrupts two pillars at once. Three changes in this update:

Confidence-ranked sourcing. Private companies often have several published revenue figures in circulation, and they rarely agree. We've adopted a formal hierarchy for our research: a regulatory filing outranks a structured financial database, and a structured database outranks a news report. When sources conflict, the highest-confidence source wins — and it's the one we cite.

No number is better than a bad number. We audited the revenue figure of every one of the 162 companies on the site against that hierarchy. Four — Michaels, Samsung Electronics America, Mercari, and Lane Bryant — traced back to sources that don't meet it, and our research found no qualifying replacement in the public record. Rather than keep a figure we can't defend, those companies now score neutral on revenue-dependent factors, with an insufficient-data label on their pages. (Michaels illustrates how these changes can offset: retiring its unverifiable revenue figure raised its score, while $4.6M in newly attributed lobbying lowered it — netting out nearly flat.)

Corrected figures, exact citations. Several figures survived the audit only after correction — Aldi's stale ~2020 estimate is now the NRF 2024 group figure ($34.16B, net of Trader Joe's), Sephora's North America figure is now U.S.-only ($8.63B, per NRF), and a handful of smaller corrections landed alongside them, with dated estimates disclosed as such on the score card. Every published figure now cites the exact source it came from — the specific filing for public companies, a labeled estimate for private ones — and every citation on the site is automatically verified against its source before each release.

The Walt Disney Company, Sears Holdings, Big Lots, New Balance, and Columbia Sportswear now have fully scored profiles, bringing the total number of companies we track to 162.

Scores below are rounded and measured from the July 2 data push. Tiers: Trusted 75+, Moderate 50–74, Concerning 25–49, Avoid below 25.

All scores reflect a 24-month window ending in 2026 Q2. The next quarterly lobbying-disclosure deadline falls in October 2026; once those Q3 filings are published and ingested, the next update will incorporate fresh data and get its own entry here.