Weighted Distribution: A Métrica que Separa Presença de Potencial de Venda

The case that makes the mistake obvious

In its 2024 annual results, Celsius reported three facts in the same sentence: total points of distribution rose 37%, ACV moved above 98.7%, and category share kept climbing in U.S. energy. The combination matters because it shows a discipline many brands still lack. The company does not report presence only as quantity. It reports the economic quality of that presence.

That detail matters more than it looks. Growth in distribution can mean two very different things. One path opens many small doors. The other wins the accounts that concentrate category turnover. On a dashboard, both can look like expansion. At the register, the distance between them is huge.

Weighted distribution exists to separate those readings. The metric asks what share of the market is covered by the stores where the brand truly sells. Instead of treating every outlet as equal, it gives greater importance to the outlets that hold more category sales.

That is why the topic matters well beyond trade marketing. It changes the brand’s sales ceiling, the efficiency of media and the interpretation of almost every meeting about growth. If the brand is absent from the accounts that pull the category, the rest of the plan works on a narrow base.

Numeric distribution answers one question. %ACV answers another.

The simplest way to read distribution is numeric distribution. If a product is present in 300 of the 1,000 stores in the measured market, numeric distribution is 30%. The measure is useful because it shows breadth. It gives a clear view of how many doors were opened and helps track territorial expansion, commercial execution and operational coverage.

The problem begins when that metric becomes the whole story. Numeric distribution makes all stores look equivalent. A very low-traffic store counts the same as a chain that carries a very large share of category volume. To answer where the brand is present, it works well. To answer how much market opportunity is covered, it works poorly when used by itself.

Ashok Charan states the distinction plainly: numeric distribution is the percentage of stores handling the product; weighted distribution is the percentage of stores handling the product weighted by category store sales. Put another way, weighted distribution equals the share of category sales generated by the brand’s handlers.

That change in lens looks small, yet it changes the conclusion. Two brands can show the same numeric distribution and still hold very different commercial potential. The brand that entered stronger accounts covers a much larger share of the real market even with the same number of stores in the spreadsheet.

What ACV actually measures

NielsenIQ defines ACV as the total dollar sales of a retailer within a specific market, across everything the store sells. Rather than isolating your category, it measures the outlet’s economic size. When a brand calculates %ACV, it takes the ACV of the stores where the product is sold and compares it with the total ACV of the measured universe.

Robin Simon describes %ACV as “percent of stores selling,” but with extra weight given to larger stores. Sally Martin makes the same bridge when she explains that the value of distribution can be weighted by the retailer’s economic size instead of a raw door count.

The result is a view of coverage that is much closer to the effective market. Instead of saying “we are in 6 of 10 chains,” the team can say “we cover 72% of the measured market’s buying power.” The second sentence changes the quality of the discussion.

There is also an important methodological detail. In Nielsen and Circana practice, a product counts as distributed when it scans. That avoids calling something “distribution” when it was only authorized, or when a shelf tag existed but the item sat out of stock. Simon is very clear on this point: authorization is one thing; scanner-captured distribution is another.

Two simple examples show why the metric changes the story

The Repsly example is useful because it breaks a common mistake. The fictional brand Solar Sparkling Water sells in two out of three retailers in its territory. Numeric distribution: 67%. Yet Solar is absent from Store 3, which alone represents US$ 50 million in ACV. The stores carrying Solar add up to US$ 50 million and the market totals US$ 100 million, so weighted distribution falls to 50%. By door count the brand looks broadly present. By economic coverage it reaches only half of the opportunity.

Robin Simon presents the reverse case. If the brand scans in Fun Foods and Smart Stores, it is present in 80% of the stores in Anytown USA, yet it covers 90% of ACV. Here the weighted number rises because the served chains carry most of the category turnover.

Sally Martin offers another version of the same lesson: a product present in two out of three stores lands at 67% numeric distribution but reaches 78% weighted distribution because those two stores are the largest ones. The three examples travel different paths and land on the same conclusion: doors never tell the whole story.

That is where practical use gets much better. When numeric distribution and %ACV sit side by side, the team can see whether the brand is growing laterally into smaller outlets or vertically into the accounts that matter more.

When the gap between the two metrics becomes wider

The gap tends to widen in categories where retail is concentrated. In those markets, a small portion of accounts controls a large portion of turnover. A brand can then show modest numeric distribution and still cover a large commercial share, or the opposite: it can appear in many small stores and remain outside the center of the category.

This pattern shows up often in beverages, snacks, personal care and categories where large national chains live beside long tails of independents. For expanding brands, the temptation is to celebrate any new door. Sound allocation of commercial effort depends on knowing whether those doors raise %ACV or simply inflate the store count.

In less concentrated categories, the difference may be smaller, yet the logic remains useful. Even when retail is more fragmented, some banners, clusters and formats still carry more basket value, more occasions or more repeat purchase. Weighted distribution translates that weight into an operational metric.

The sales equation starts with distribution and then moves to velocity

NielsenIQ defines velocity as sales per distribution point. An item with US$ 1 of velocity and 100 distribution points would sell US$ 100. The elegance of that definition is its order. Velocity measures performance inside the coverage already won. Before velocity comes coverage itself.

“Shoppers cannot buy something that’s not in the store!” Robin Simon, CPG Data Tip Sheet

The logic looks obvious, yet it disappears from many growth conversations. When a brand undershoots the sales target, the immediate reflex is often to blame packaging, promotion, price or communication. In many cases the bottleneck appears earlier: the brand has not yet entered the accounts that would give real scale to its velocity.

This also explains why velocity alone can mislead. A young brand may have excellent sell-through in a handful of strong accounts and still remain small because %ACV is low. Another brand may have only average velocity but strong economic coverage and therefore stand in a better place to scale through execution, pricing or visibility.

When distribution and velocity are read together, diagnosis improves a lot:

What academic research adds to the conversation

The point goes beyond teaching examples. Wilbur and Farris studied a census of more than 79,000 SKUs across 37 consumer packaged goods categories totaling US$ 55 billion in annual revenue. In 86% of the categories, the relationship between market share and distribution at SKU level was increasing and convex. In practical terms, gains in distribution and gains in share tend to reinforce each other, and the reinforcement often accelerates as presence improves.

The authors also show that the same relationship appears inside the portfolios of leading brands. Low share and low distribution do not hurt only small brands competing against leaders. They also hurt weak SKUs inside strong brands. That matters a lot for assortment decisions, line extensions and portfolio priorities.

Reibstein and Farris had already described, in 1995, a generalized convex relationship between retail distribution and unit market share. In management language, distribution can work as a compass for expected share. When a brand falls far below the curve, proposition, price or execution may need attention. When it sits above the curve, the brand may be showing strength worth defending.

Martin Hirche and coauthors extended the logic in 2021 by showing that machine learning models can classify 83% of under- and overperforming SKUs within the distribution and market share relationship. The value is direct: the curve stops being only a descriptive picture and becomes a decision aid for portfolio management.

Antonis Michis adds another layer by studying weighted distribution and sales with Dutch laundry detergent data. His model shows that the relationship can be nonlinear and even nonmonotonic in some contexts. In practice, the marginal gain from each extra point of distribution varies. Winning one major chain can matter much more than opening many weak doors. There can also be stretches where coverage grows faster than the brand’s ability to convert that coverage into sales.

Why scanning matters so much in the definition

A recurring mistake in commercial teams is to treat authorization as completed distribution. The buyer approved the listing, the code exists, the theoretical planogram is ready, and the team starts behaving as if coverage were already secured. Yet an empty shelf, poor replenishment, a shelf tag without stock or a store that never received the item does not turn into sales.

Robin Simon answers a common question on this topic by noting that it would be unworkable to audit tens of thousands of stores all the time just to confirm shelf presence. The scanner becomes an operational proof that the product was truly available for purchase. The method can understate slow-moving items in some periods, but it avoids an even bigger mistake: celebrating distribution that never reached the register.

For management, that changes the conversation with brokers, trade teams and supply. If listing progressed and scanning never appeared, the bottleneck sits in execution. If scanning exists but %ACV remains low, the bottleneck sits in economic coverage. If %ACV rises and velocity softens, the bottleneck may sit in assortment, price or visibility.

What to do with this reading in practice

1. Prioritize accounts with category logic

When negotiating expansion, the first filter should not be store count alone. The central issue is the account’s weight in the category. In some cases, a retailer that looks smaller in total revenue can matter much more to your category than a larger account with weaker occasion fit.

2. Read distribution by SKU as well as by brand

The Wilbur and Farris paper is a strong warning. Inside a powerful brand, weak line extensions can absorb commercial energy without building proportional share. Reading %ACV by SKU helps separate items that deserve more defense from items that mostly consume team attention.

3. Read distribution together with share of shelf

Entering the right account solves only part of the problem. Inside that account, the brand still needs facings, replenishment, visibility and an assortment that fits the buying occasion. High economic coverage with weak shelf presence limits the return on distribution.

4. Cross coverage with Category Entry Points

A chain may hold high ACV and still be a weak outlet for a specific SKU if its dominant buying occasions do not match the item’s role. Mature distribution reading links account, occasion, store format and mission of purchase.

5. Avoid isolated readings of demand signals

Even upper-funnel signals become partial when physical buying is difficult. That is why indicators such as Share of Search stay incomplete when they are read without enough physical availability.

A simple framework for channel reviews

If the team wants less noise, six fixed questions already improve most distribution reviews.

That set solves many expensive mistakes. It blocks empty celebration of capillarity and creates a cleaner conversation across sales, trade, marketing and supply.

What the metric teaches about brand growth

The most mature reading of brand growth ends in mental availability and physical availability. Weighted distribution sits at the center of the second half of that equation. It shows whether the physical presence already won is happening where the market truly buys.

That matters because many brand strategies fail in the transition between desire and access. Communication lands. Memory grows. Interest appears. Yet the brand is still absent from the heaviest accounts or reaches them with an incomplete assortment. In that situation, the brand creates demand that competing retailers or competing products capture.

Weighted distribution corrects that blindness. It forces the company to treat market access as a brand topic and a sales topic at the same time. A growing brand must be easy to remember and easy to buy. Counting doors without weighing those doors leaves half of the job in the dark.

That is why the right question is rarely “how many stores are we in?” The question that moves results is “how much of the market that matters is already covered by our real presence?” That is the moment when weighted distribution stops being a trade spreadsheet and becomes a strategy tool.