All commodity volume is the dollar sales of everything a store sells, across every category, used as a weight. That is the whole concept. ACV distribution takes that weight and reports the share of a market’s total retail dollars sitting inside the stores that carry your brand.
Which is why 70% ACV distribution and 70% of stores are two different numbers, and the gap between them is bigger than most brand teams assume. A product in half the doors in a market can hold a quarter of the market’s ACV or three quarters of it. Same door count. Same product. The only variable is which half.
We’re gmware, a software and data engineering firm in Austin, TX, with delivery centers in Bangalore and Mohali, India. We also build Shield Suite, our retail-intelligence platform for beverage-alcohol brands across 60,000+ storefronts. A lot of our week goes on reconciling a weighted number out of a syndicated file against an unweighted account list out of a distributor feed, then explaining why the two are never going to agree.
One boundary first, because these get mixed up constantly. ACV is about what you weighted by. On-shelf availability is about what you divided by. Coverage and availability are separate metrics with separate failure modes, and this post is only the first one.
| Term | What it is in one line |
|---|---|
| All commodity volume (ACV) | A store’s total dollar sales across every category it sells |
| %ACV distribution | Share of a market’s store dollars sitting in stores that carry you |
| Numeric distribution | Percentage of doors carrying you, every door counted once |
| PCV | The same weighting run on category dollars instead of all-commodity dollars |
| Sales per point of distribution | Sales divided by %ACV distribution |
What the number is actually made of
Start with the raw ingredient. FMI’s food industry glossary defines ACV as “the annual dollar sales of everything sold in a store” and says the measure is rarely used by itself, which is correct. MASB’s Universal Marketing Dictionary makes the all-category part explicit: the dollars that go into ACV cover “the entire store inventory sales, rather than sales for a specific category of products”, and that is where the phrase comes from.
So a supermarket’s ACV includes its produce, its detergent and its rotisserie chicken. None of that has anything to do with how much tequila the store moves. ACV is a store-size yardstick, nothing more, and the only reason a brand ever looks at it is that it becomes the weight in a distribution measure.
The peer-reviewed literature states that weighting mechanic cleanly. Hirche, Völckner, Trinh and Göbl, in a 2024 Journal of Marketing Analytics paper built on NielsenIQ scanner data covering 1,682 brands and 12,049 stores, write that under weighted-distribution metrics “Stores carrying a brand are weighted according to the sales revenue they generate”. They also draw the line between ACV and its sibling metric PCV. ACV is weighted by total store sales; PCV is weighted only by sales of the category in question. For mainstream packaged goods the two usually land in a similar place. For a category with a narrow retail footprint they can drift apart, which is one reason a spirits brand should know which of the two its report is built on.
Then the period, which is where the published definitions genuinely conflict and where most glossary pages pick a lane and move on. FMI and MASB both describe ACV as annual. But the same MASB entry runs its worked example in weekly ACV per store, and the studies that actually compute ACV-weighted distribution from scanner data work per week and then average. The 2024 paper calculates weekly averages for the observation period and states that “the weekly averages for the year are calculated by dividing the sum of all weekly brand PCV values by 52”. A companion 2023 study of US wine retailing, using the same class of data across 3,524 SKUs and 4,218 stores, says the same thing: “The metrics for distribution and market share are calculated as weekly averages for the year.”
The formulation that survives both readings: ACV is a total-store, all-category dollar-sales figure for the reporting period, quoted annualised by convention, while ACV-weighted distribution is normally computed per period and averaged. Which means a single %ACV number always has a period attached to it. Practically, ask what the period is before you compare your number to last year’s or to a competitor’s.
Two formulas, side by side
Everything in this post comes out of the difference between these two lines.
Numeric distribution (%) = 100 × outlets carrying brand ÷ total outlets (Marketing Accountability Standards Board formula)
ACV distribution (%) = total sales of stores carrying brand ÷ total sales of all outlets × 100 (Marketing Accountability Standards Board formula)
One divides counts by counts. The other divides dollars by dollars. FMI puts the consequence in one plain sentence, describing %ACV distribution as “comprable [sic] to percent of stores selling, but bigger stores get heavier weight”. Circana’s public dictionary, in the provider’s own wording, calls ACV a figure “used as a weight to value store importance”, which matches the independent sources. The definition gets treated as proprietary. It isn’t.
Counting doors versus weighting dollars
Which half of the doors you landed in
MASB publishes the cleanest available demonstration that store count and ACV are not interchangeable, and it is worth walking through slowly.
Picture a city with two hardware retailers. One runs five big-box locations. The other runs five small locations of the mom-and-pop kind. Total market: ten stores, and by MASB’s numbers, $20MM of weekly market ACV, of which the big-box chain accounts for $15MM and the small chain $5MM.
Now put a tool set into the small chain only. It sits in five of the ten stores, so it has 50% of the doors. On ACV it holds a quarter of the market, because that is the share of retail dollars its five stores represent. Put the same tool set into the big-box chain only and it still holds 50% of the doors, but now it reaches “75% total market ACV”. MASB’s own conclusion: “the percentage of stores a product sells within is less relevant than a product’s share of the store ACV value.”
Same 50% of doors, three times the ACV
We flagged that inconsistency rather than quietly fixing it because it illustrates the working problem. ACV arithmetic is easy to state and easy to get subtly wrong, and every downstream number inherits the error.
What three-tier distribution does to the number
The glossary pages ranking for this term stop at the formula. For a beverage-alcohol brand there’s a fact behind the formula that matters more: you don’t choose which half of the doors you land in. Your distributor’s account coverage and the chain programs you win choose it for you. %ACV is the metric that prices the outcome.
Three consequences follow, none of which appear in the definition.
It moves in steps, not slopes. In a market where large-format grocery or mass can sell your category, those doors carry most of the ACV weight, so a single chain authorization can shift your %ACV several points in one reset cycle while a full quarter of field effort in independents barely registers. That makes %ACV a poor weekly execution metric and a good quarterly coverage metric. Teams that put it on a monthly dashboard end up reading noise, then reading a step change as a trend.
Your channel ceiling is set before you start. Which retail channels may sell wine or spirits at all differs by state, and in control states the state itself sits in the wholesale or retail tier. So the store universe underneath a national %ACV figure is not one uniform population, it is a stack of state-specific ones with different large-format participation. A national number can conceal two markets moving in opposite directions. Our control-state data guide covers what is measurable where.
The number you can see and the number you’re graded on are computed differently. Your internal view of coverage comes from depletion reporting, which is an account list: unweighted, and covering every account the distributor ships. The %ACV figure comes from a syndicated panel: weighted by store dollars, and covering only retailers who contribute data. Neither is convertible into the other without a store-level ACV weight for every account on your list, which is a thing you do not have. When a sales lead says coverage is up and a category manager says %ACV is flat, both are usually right.
There is a fourth effect that shows up as a data-quality bug rather than a reporting one. FMI is explicit that “The item must scan at least once during the period to be counted as in distribution” and that authorization is not distribution. Under three tiers, an account can be sold in by a distributor, receive the cases, leave them in the back room, and never ring one up. Your depletion file shows the account. The %ACV number does not. That is not a measurement failure. It is the metric working exactly as specified, on a definition of “distributed” that means observed selling.
Four ways a %ACV number gets misread
Summing it. FMI writes it in capitals: “Percent ACV Distribution is NOT additive across products, markets, or periods!” Three states at 60% ACV each do not give you 180, and they do not reliably give you 60 either, because each state’s figure has its own denominator. A national figure has to be computed against a national store universe from the start. Adding %ACV across your own SKUs produces a different metric entirely, total distribution points, which is a legitimate measure in its own right and is not a percentage of anything. Label it accordingly or somebody will read 340 as an impossible share.
Treating it as depth. A store enters your ACV base on its first SKU. The 2024 paper is blunt about the consequence: “Stocking or delisting additional SKUs of a brand does not change the brand’s ACV or PCV.” A chain cutting you from six facings to two shows up nowhere in this number. If assortment depth is the thing you are managing, %ACV is the wrong instrument, and the same paper notes that brand-level breadth metrics saturate: because one SKU is enough, “large brands tend to reach high distribution levels with low variability of measured distribution across brands.”
Assuming the denominator is fixed. “Total sales of all outlets” reads like a constant. It is a modelling choice. The 2023 wine study states its own choice in a footnote, limiting the universe to “all sampled stores that have sold any wine SKUs in the year of investigation”, which is a defensible decision that also changes every percentage it produces. In beverage alcohol the equivalent questions are live ones. Do convenience and drug stores that sell beer belong in the universe for a spirits brand? Are non-participating retailers projected in or excluded? We could not retrieve a published methodology document from either major syndicated provider, so we are not going to tell you what any specific vendor does by default. Ask your provider, in writing, and keep the answer next to the number.
Reading it as reach into your category. ACV is all-commodity by construction, so a store with enormous total sales and a thin category set contributes its full weight regardless. That is the intended behaviour, and it is why PCV exists as an alternative. Where your category’s footprint inside a retailer differs sharply from that retailer’s overall size, the two metrics tell different stories and only one of them is about your shopper. The channel version of this hole, and what it does to brands whose volume sits in small doors, is in our independent liquor store coverage post.
Two divisors both called velocity
This is where the ACV caveats stop being academic, because velocity is computed from %ACV and inherits every one of them.
MASB defines sales velocity as the rate at which dollars or units pass through in a given time, basically sales divided by distribution, and gives the headline formula as “Sales per point of distribution ($) = Sales ($) ÷ % ACV distribution”. Look at the divisor. It is a percentage, so the output is dollars per ACV point. Not dollars per store.
Field-sales practice uses the same vocabulary for a different ratio. Rate of sale is commonly expressed as units per store per week, dividing a unit count by a store count and a week count. That is common usage rather than a standards-body definition, and we have not found a trade-association source for it, so treat it as convention. It is also the number a chain buyer is most likely to have in mind when they say your rate of sale is weak.
Same word, three divisors
The deeper problem is that the breadth-based divisor may not be sharp enough for the job people give it. The 2024 paper says so directly: “Common distribution-velocity models use distribution-breadth metrics. However, distribution-breadth metrics lack the variability needed to meaningfully differentiate competing brands.” That is a peer-reviewed statement that the standard divisor is inadequate, which is a stronger claim than we expected to be able to make when we started sourcing this.
The literature also restates a long-standing finding that ought to change how you read a velocity comparison: the relationship between distribution breadth and market share is convex, a double-jeopardy pattern in which “high-share brands tend to sell more per point of distribution”. If that holds in your category, then a small brand posting lower sales per point than a market leader is not necessarily executing worse. It is sitting where the curve says it should sit. Setting a velocity target by benchmarking against the biggest brand on the shelf builds the wrong expectation into somebody’s plan.
What we’d recommend
Publish %ACV distribution and a plain door count together, always, with the period and the store universe printed next to both. Two numbers and two lines of boilerplate. That single habit resolves most of the arguments this metric causes, because the argument is nearly always that two people are holding different denominators and neither has said so out loud.
Use %ACV for the questions it answers well. Pricing what a chain authorization is worth before you chase it. Reporting coverage to a board in dollar-weighted terms, or comparing this quarter to the same quarter last year on the same universe. As a quarterly coverage measure it’s genuinely good.
Don’t use it to manage execution. It can’t see facings, it can’t see shelf position, it saturates once you’re broad, and it enters a store on one SKU and then goes quiet. For anything week-to-week you need observation at the store, with a date on it, which is a different kind of data with a different cost structure. Our syndicated data guide walks through where the line between the two sits.
Below roughly a few million dollars in revenue, weighted distribution is usually premature, especially if sales sit in one or two states. Count your doors. Keep the account master clean, and spend the money on sales instead. The weighted view starts paying for itself once you’re multi-market and the question shifts from how many accounts to which accounts.
At that point what you want sitting next to a %ACV number is the store-level detail underneath it. Which specific doors. What a competitor is doing in those same doors, and what changed since last week. That’s what our Competitive Intelligence module produces across 60,000+ storefronts. The less glamorous half of the job, getting your own store and product masters into a state where any weighted number can be trusted, is ordinary data analytics work.
Bring us a %ACV figure and the store list you believe sits behind it, and we’ll give you a straight answer on which of the two is lying. Usually it’s the store list. We work with beverage-alcohol brands whose weighted coverage looks fine right up to the week a single chain resets.