If you've spent time on the analytics side of the beverage alcohol industry, you've probably come across BW166. Their monthly and quarterly reports are useful because they do the annoying part: pull TTB production volumes, BEA spending data, and other federal tables into one readable view of the US market. They're widely cited, and they're paywalled.
Fair enough. If you need a finished report for an executive deck, buy the finished report. But if you want to do your own analysis, most of the raw material is public. This is the map I would start from.
TTB production reports
The Alcohol and Tobacco Tax and Trade Bureau publishes monthly and quarterly statistical reports for beer, wine, and distilled spirits. These are the federal records for how much alcohol was bottled, packaged, taxed, or removed from bond.
- Where: ttb.gov/statistics
- Format: PDF and Excel, posted monthly with a 1–2 month lag
- What's in it: production volumes by commodity, bottling and packaging totals, withdrawals subject to tax, and similar figures broken out by category
- What it doesn't tell you: dollar values, retail prices, or anything below the national aggregate
We mirror the cleaned, machine-readable version of these reports at colacloud.us/ttb-data-pulse as CC0-licensed open data, refreshed daily.
BEA Personal Consumption Expenditures (PCE)
The Bureau of Economic Analysis publishes how much US consumers actually spend on alcohol, broken out by spirits, wine, and beer, in both nominal dollars and real (inflation-adjusted) dollars. They also publish quantity indexes and price indexes, which are the bridge between "people spent more" and "people bought more."
- Where: BEA iTable, NIPA tables 2.4.5U (annual expenditures), 2.4.4U (price indexes), 2.4.3U (quantity indexes)
- Format: web table, XLSX export, JSON API
- What's in it: nominal and real consumer spending on alcoholic beverages, refreshed quarterly with annual revisions
- What it doesn't tell you: anything sub-national, anything brand-level, anything below the spirits/wine/beer split
BEA has a free API. Once you have the right lines, the spending-side charts are not especially mysterious.
Census Monthly Retail Trade Survey (MRTS)
The US Census Bureau tracks monthly retail sales for "Beer, Wine, and Liquor Stores" as NAICS 4453. In plain English: how much money standalone alcohol retailers rang up.
- Where: census.gov/retail/mrts
- Format: Excel and CSV, monthly with ~6 week lag
- What's in it: monthly retail sales for beer/wine/liquor stores, both seasonally adjusted and unadjusted
- What it doesn't tell you: anything about grocery, on-premise, or e-commerce alcohol sales — only the stand-alone retail channel
FRED (Federal Reserve Economic Data)
The St. Louis Fed aggregates and serves a huge number of federal time series, including the BEA and Census series above. It is not a separate source so much as a better front door to the same source material.
- Where: fred.stlouisfed.org
- Useful series: CPI alcoholic beverages (CUSR0000SAF116), beer/wine/liquor store retail sales (MRTSSM4453USS), PCE for alcoholic beverages off-premise (DAFFRC1Q027SBEA)
- Why use it: free API, consistent format, easy to chart and pull into Python or R
TTB Public COLA Registry
Production volumes and consumer spending tell you about the market in aggregate. The COLA registry tells you about the market at the product level — every new product introduction is in there months before it reaches retail. This is the dataset BW166 doesn't really cover. Their angle is volumes and dollars; this is catalog data.
- Where: TTB Public COLA Registry (raw), COLA Cloud (cleaned, with images and AI extraction)
- Format: web search at TTB, REST API + SDK + bulk Snowflake share at COLA Cloud
- What's in it: 2.9 million label approvals back to 2005, every approved alcohol product, with brand, class/type, ABV, container size, country of origin, and the actual label images
- What it doesn't tell you: how much of any of these products was actually produced, sold, or consumed
See what a COLA is and how it was used to document tariff engineering on European wine for what this looks like in practice.
Where paid sources still earn their keep
The federal sources above cover production, consumer spending, and product introductions. They do not cover everything. The expensive data is expensive for a reason:
- Retailer scanner data — brand-level, sub-channel sales at grocery, convenience, and liquor stores.
- Distributor depletion data — what's actually moving out of distributor warehouses to retailers.
- On-premise data — bar and restaurant menus, pricing, by-the-glass volumes.
- International coverage — the federal sources are US-only.
If your question is "is brand X gaining share at grocery in Texas," that's not a federal-data question, and no amount of TTB or BEA wrangling will answer it. Paid analytics providers earn their keep there. For a lot of market sizing, category trend, and public benchmark work, federal data plus an afternoon of integration gets you surprisingly far.
How to actually do this
If I were building a lightweight internal market view from scratch, I would start here:
- Pull BEA PCE and Census MRTS via FRED's free API for the macro view.
- Pull TTB statistical reports for production volumes (or use our cleaned mirror).
- Pull COLA registry data via the COLA Cloud API or Snowflake share for product-level introductions and catalog growth.
- Put all three on the same monthly or quarterly grain. That gets you most of the public-data backbone behind a typical market report.
The narrative layer in paid reports is real work. I'm not hand-waving that away. The point is narrower: the underlying public data can sit in your warehouse instead of in a PDF attachment, and that makes it much easier to join to everything else you know.
If you're trying to do this and want a hand on the COLA side specifically, reach out — we'll point you at the right endpoints and data shapes.