How to Actually Do a Craft Beer Industry Analysis Without Wasting Three Weeks
I spent last November trying to put together a proper Craft Beer Industry Analysis for a client who wanted to break into the Northeast sour market. By the third day I realized most of the data out there was either three years stale or so aggregated it was useless. The Brewers Association releases annual reports, but those numbers flatten regional variation to the point where they basically tell you nothing about whether a specific beer style is actually moving on tap in Portland versus Providence. You have to go beyond the published reports if you want the analysis to be useful for any real decision. A decent analysis needs three layers of data. First, production volume and growth by segment —IPA, stout, sour, lager, seasonal — broken down at least by region. Second, distribution and retail mix, because how much volume a brewery pushes out the door means nothing without knowing whether it's hitting kegs, cans in grocery, or direct-to-consumer. Third, pricing and margin data, which most public sources simply do not provide. You end up triangulating from a combination of industry reports, distributor interviews, and point-of-sale data from places like Circana, formerly IRI. I started every project the same way: I pull the Brewers Association yearly statistics, then layer in state-level ABC data for production licenses and shipments. After that I check the USDA economic research service for per-capita beer consumption trends and the Census Bureau for retail trade flows. The problem is none of these datasets talk to each other. Production reports don't include on-premise versus off-premise split. State license data tells you how many breweries exist but not how much they actually sell. You need to fill those gaps yourself.
Where the Data Actually Comes From
Here is the practical breakdown of sources I use and what each one covers. The Brewers Association gives you national production volume, revenue estimates, and brewery count by type. It updates annually and is reasonably reliable for macro trends. But their regional breakdowns are broad and lag behind real market shifts by at least six months. A craft segment that looked flat nationally in their 2023 report was clearly surging in the Pacific Northwest by early 2024 based on retail scanner data. Circana and NielsenIQ provide the most detailed retail scanner data available, including SKU-level sales by channel, package type, and price band. Their craft beer data runs on a subscription that typically costs between fifteen and forty thousand dollars annually depending on scope. If you are doing one analysis for a single client, splitting that cost with two or three other people who need the same data is normal practice. I have done it twice and it cuts the price to roughly five thousand dollars per party.
State alcohol beverage control commissions publish brewery license and shipment data, but the quality varies wildly. Colorado and Oregon release detailed monthly production figures. Mississippi does not release anything useful. You have to check each state individually and build a map of what is actually available before you start collecting. Tap list aggregation services like Untappd and BeerAdvocate are not reliable for volume data. They capture check-ins and reviews, which correlate loosely with popularity but do not tell you how many barrels moved. I use them only for qualitative signal — which styles are getting attention in which cities — not for any quantitative claim.
Get the Full Details
The Process I Actually Follow
I begin by defining the scope. A national Craft Beer Industry Analysis is a different project than a regional one focused on a single state or metro area. The national version takes about forty to sixty hours of research and synthesis. A regional deep dive can be done in twenty to thirty hours if the state data is good. Next I build a spreadsheet that tracks market size by style category, growth rate year over year, and distribution channel mix. I populate it with BA data first, then layer in Circana retail numbers for the top ten metro areas by craft beer sales volume. Those metros are generally New York, Los Angeles, Chicago, San Francisco, Seattle, Denver, Portland, Austin, Boston, and Minneapolis based on recent scanner data. I ignore cities that appear on generic lists because craft penetration there is either too low to matter or dominated by macro brands that distort the picture. Then I pull state-level production data for the target region and cross-reference it with on-premise hospitality licensing data to estimate taproom and restaurant distribution. This is where most people skip steps and produce garbage. You cannot assume that because a brewery produces five thousand barrels a year, five thousand barrels were sold. Many of those barrels go to waste, give away, or sit in storage. Actual shipped volume is usually sixty to eighty percent of stated production for mid-sized craft breweries, sometimes lower during equipment downtime or formulation changes.
I always include a section on margin dynamics because it is the part everyone forgets and the part that actually determines whether a brewery survives. Ingredient cost, packaging choice, distribution markup, and retailer take rate interact in ways that are not obvious from production data alone. A sour aged in oak barrels for twelve months has a completely different cost structure than a pale ale canned on day four. Two breweries producing the same number of barrels can have profit margins that differ by twenty percentage points depending on their packaging and distribution model.
A Real Problem I Ran Into
During that November project I hit a wall with the New England pale IPA segment. The national data showed flat or slightly declining growth, but my client was convinced the style was still hot in Massachusetts and Connecticut based on what they saw at local taprooms. The scanner data confirmed the decline nationally but showed a thirty percent increase in dollar sales in those two states over eighteen months. The volume was dropping because prices had risen sharply. Breweries were charging more per unit while selling fewer units, and the national aggregate smoothed that out completely. I resolved it by pulling distributor invoices for three mid-size retailers in the target markets and comparing per-unit cost to retail price across twelve-month intervals. That gave me actual margin data instead of estimates. The workaround took me about eight hours of phone calls and email requests to distributors who were not excited about sharing pricing information. Most said no. Two said yes and provided partial data. I used their numbers and extrapolated conservatively with a confidence interval noted in the final report. Being transparent about uncertainty is better than pretending the data is more complete than it actually is.
Common Mistakes That Ruin These Analyses
Using brewery count as a proxy for market health is the most frequent error I see. More breweries does not mean a healthier market. It often means the market is fragmented and margins are compressed. The average craft brewery in the United States produces between five hundred and fifteen hundred barrels annually, and roughly forty percent of those close within five years. Adding another twenty breweries to a saturated metro area usually reduces per-brewery volume rather than expanding the total pie. Another mistake is treating craft beer as a single segment. The sub-segments move in opposite directions at different times. Hazy IPA peaked in retail scanner data around 2019 and declined steadily through 2022. sessionable lagers and light lagers have been growing since 2021, especially in on-premise channels. Sour and wild ale production has expanded but remains a small share of total volume, roughly eight to ten percent nationally. If your analysis treats all craft beer as one category you will miss the actual dynamics that matter for strategic decisions. A third mistake is relying solely on published reports without checking the methodology. The Brewers Association defines craft beer using specific criteria around ownership, volume caps, and traditional brewing. Some breweries that qualify as craft under that definition operate more like macro satellite facilities with centralized large-scale brewing. Their data skews the craft narrative in ways that are hard to detect unless you look at individual brewery profiles.
What This Approach Cannot Do
No amount of analysis can reliably predict consumer behavior shifts during economic downturns. During periods of inflation and reduced discretionary spending, premium and super-premium craft segments contract faster than value segments, but the timing and magnitude are unpredictable. I have seen analysts project steady double-digit growth for craft IPAs in markets that then experienced sharp declines when consumers traded down. The data available at the time of the projection did not contain enough leading indicators to flag the shift. Craft Beer Industry Analysis also cannot account for regulatory changes that happen unexpectedly. A new state allowing direct-to-consumer shipping, a sudden tax increase on alcohol by volume, or a pandemic-related restaurant closure can invalidate months of work in weeks. The best analyses include scenario planning with alternative assumptions rather than presenting a single forecast as if it were certain. If you do not have access to paid scanner data, the analysis will be thinner. Free sources can get you to a reasonably accurate macro picture, but you will lack the granularity needed for market entry decisions at the metro or SKU level. In those cases I recommend partnering with a local distributor or consultant who already has the retail data, rather than trying to reconstruct it from scratch. It saves roughly fifteen to twenty hours of effort and usually produces a more accurate result.
Practical Output Structure
A usable analysis should include a market size section with current production and sales volume by style and region, a competitive landscape covering the top fifteen players by volume and share, a distribution analysis showing on-premise versus off-premise split and key retail partners, a pricing and margin overview based on available data, and a trend section with forward projections that include confidence ranges. Any claim beyond current and historical data should be explicitly labeled as a projection with the assumptions stated. I usually deliver the final document as a twenty-five to forty-page report with an executive summary no longer than three pages. The appendices contain the raw data tables and source citations so anyone reviewing the work can verify the numbers. That appendix alone takes about half the total project time, which is why people who skip it later get caught when someone asks where a specific figure came from. There is no shortcut around doing the work carefully. The data is messy, the sources disagree, and the market changes faster than any report can capture. But a disciplined approach that acknowledges those limits produces something actually useful for someone deciding where to brew, what to brew, and whether to enter a market at all.
