How Retailers Improve Store Assortments for Local Shoppers

The assortment lesson I would share comes from our own product: more options is not the same as more sales. At Simply Noted we sell handwritten notes, and early on we offered a long list of card styles and layouts because we assumed customers wanted choice. When we looked at what people actually ordered, a small number of designs drove most of the volume, and the long tail slowed down decisions and added complexity to our production. So we trimmed the options and put the proven ones up front, then tested new styles in small batches before giving them permanent space. Customers decided faster, and operations got simpler because the robots and our team handled fewer variations. For retail teams, the same principle applies to shelf space. Look at what each store really sells rather than what you expect, give extra room to the few items that turn quickly in that location, and let slow movers earn their space through a short trial rather than sitting there by default. Preferences do vary by store, but the discipline is the same everywhere: let the sell-through data decide, and revisit it regularly. Rick Elmore, Founder/CEO, Simply Noted (simplynoted.com)

The cleanest way to decide shelf space is to stop arguing about preference and start measuring revenue per unit of space. Treat a shelf slot the way we treat a screen position online: it is a scarce resource, and each slot should be judged by the revenue it produces in that location, not by how it sells in your best store.

We learned the cost of the opposite approach in a B2B shop. We added more options to the cart, volume tiers plus a shipping calculator, on the theory that more choice serves more buyers. Revenue per user fell 52.58% (control 9.70 EUR, variant 4.60 EUR, 1,770 users, 100 days). More options did not read as more service, it read as work. A shelf works the same way: breadth can look fresh and still slow the shopper down.

The decision that taught us this: strip the tail, give the space to the proven movers, and test the replacement in a few stores before a chain-wide rollout. Fix placement before you cut a product, because a slow mover that never got a fair slot was not tested, only hidden.

Local variation is real, but keep the metric constant across stores. Revenue per unit of space is comparable; raw units sold is not.

1. Treat every store as if it were in a completely different neighborhood. Sell-through and margin per sq ft by location are good starting points. But then you have to make decisions that the data cannot make for you. I.e. who shops in that store, what time of day do they shop, what do they reach for first? Be willing to make decisions based on incomplete information. The store located near offices for young people and the store located near family condos will likely have completely different looking stores even though they are selling the same brand. Listen to your floor staff. They get asked "do you have this in a larger size?" long before it appears in a report somewhere. Use the data to make decisions on shelf space based on proven demand at that store location. Make sure to leave a small deliberate section of the shelf to test new items and don't just reward the items that are doing well yesterday.

2. A "shelf" in our case for CheapForexVPS.com is the plans (and corresponding locations - servers) we are promoting. The behavior of our customers (traders from Asia, Europe, USA etc.) is that of typical shoppers in different stores. What matters to them is which data center is closest to their Forex broker. I used to promote all our plans more or less equally at all locations. Then there were 2 or 3 "tiers" of plans which for whatever reason did not get much promotion and thus sat on the shelf like unsold stock in the back of the store. Eventually I regrouped the offering by region and promoted the plans and corresponding locations that already got the most sales in that region. The "slow" ones got "folded in" to a few better offered alternatives. The end result was that customers no longer hesitated to complete the order, there were fewer questions to support about which plan to choose for a given scenario, and conversions went up. In general - fewer better placed choices on the shelf are almost always better than a crowded one.

Retail groups determine shelf space through clustering at the store level rather than based on a national average. In my various projects, I have seen the best retailers view shelf space as a scarce resource that requires justifications based on trends, as opposed to just relying on past decisions. Utilizing the integration of the Point of Sale system and the Warehouse Management System to create a continuous feedback system is key in this decision-making process. For instance, when a particular product category has a good movement in a suburban area but is slow in a high-profile city location, it is ideal for the ERP system to call for the shift of this shelf space to quicker-moving items.

One of the assortment decisions I made was changing a large distribution client from a single procurement process to a clustering-based plan. Instead of relying on geographical criteria alone, we categorized stores according to their specific movements of products. This helped us realize that some heavy products were taking up necessary shelf space in areas where fast transactions took place. We subsequently decided to cut back on these slow movers in order to leave space for the proper convenience products that would meet the needs of the consumers in this locality. This resulted in the reduction of slow-moving inventory as well as freeing up the capital previously tied up in unsold inventory.

The success of these implementations relies on the understanding of the operations and IT directors of the way replenishment triggers should be calculated.

When customer preferences vary by location, I think the smartest shelf-space decisions come from looking at what customers actually buy at each store rather than applying the same assortment everywhere. I would compare sales velocity, repeat purchases, seasonal demand, and how long individual products sit before deciding which items deserve more space. In moving operations, I have seen a similar pattern with packing supplies: customers shipping vehicles and household goods to Hawaii often need different materials than customers making smaller moves, so keeping too much of a low-demand item only ties up space. We adjusted the mix by keeping more of the packing materials customers consistently requested and reducing quantities of supplies that moved slowly. That helped us use storage space more efficiently while avoiding unnecessary excess inventory. My advice to retail teams is to treat each store as its own demand environment and earn shelf space based on local sales data rather than company-wide assumptions.

Shelf space should be earned through sales performance, not allocated equally across every store. A product that sells quickly in a suburban supermarket might barely move in a convenience store near an office district.

At Mills Shelving, we've prioritised modular gondola systems that allow retailers to change shelf heights, add hooks and adjust product facings without replacing entire fixtures. This gives retailers the flexibility to expand strong-selling categories while reducing space allocated to slower products.

My advice is to measure sales per shelf metre alongside replenishment frequency. The goal isn't to display more products. It's to make every metre commercially productive.

We changed our late-season planning from adding more inventory to keeping options open. Instead of committing heavily to every early winner, we reserved capacity for new signals as regional events and social conversations began shaping demand. This approach meant starting with smaller quantities and setting clear reorder triggers based on daily sales. It allowed us to respond to changing demand without committing too early.

The result was less unsold inventory without losing flexibility. Early demand can be misleading in a short seasonal cycle because excitement does not always last. By separating discovery from commitment, we gave the merchandising team time to learn before making decisions. We recommend setting triggers in advance based on sales pace, minimum margins, and remaining calendar days to make decisions faster and clearer.

Shelf space is literally a decision about who's walking into that store, and most retail teams still make it off what sold there from last season.

At Nike the account relationships were absolutely everything, what they did was let us build a different assortment for each door based on the customer persona it served along with their psychology and how they actually made a decision. We split the market by the role each store played in culture which allowed marketplace segmentation at its best. There were two Tier Zero accounts in Australia for the tastemakers, then eight to ten Energy and Quickstrike doors for the early adopters, then our best accounts nationwide for the watchers, the people who look to someone else to tell them what's cool. At the end of that ladder are the skeptics, who only buy once enough people are wearing it that it feels safe.

The assortment decision that made the biggest difference was actually taking product out. We regularly pulled product from the doors serving skeptics, even though it would have sold there very easily, and we saw double-digit growth year on year as a result of protecting product and not selling it everywhere we could. Have to be clear here also, I too have got this wrong. The Roshe Run for example stayed in market way too long, I didn't turn it off and it was over-distributed, and once too many people had it the silhouette died almost overnight and we were left with an inventory problem.

What this tells you is that your sell-through data shows what somebody bought in that store, but it can't show you who they were being when they bought it. So before you give a product more space, you have to ask who that store's customer is when they make the decision. Are they the ones leading, or the ones waiting until it feels safe enough?

I'd be bold enough to say slow stock usually starts as good product sitting in front of the wrong customer. The extra space just makes it obvious.

Nick Gray, Founder & CEO, IGU Global | Rethink Retail Top Retail Expert 2026

We run marketplace storefronts rather than physical shelves, but the problem is identical: finite attention, and demand that does not look the same in every location.

Our version of shelf space is the parent listing and the inventory we are willing to hold in each marketplace. The signal we trust most is the Sales and Traffic by Child Item report in Seller Central, read as two columns together. Sessions tell you what shoppers are looking at. Unit Session Percentage tells you whether they buy once they get there. A child variation pulling heavy sessions with a weak unit session percentage does not deserve more inventory, it deserves a listing fix. A child with a strong unit session percentage and thin sessions is where the stock and the ad budget should go.

The assortment decision that mattered most on one catalog was cutting variations by marketplace instead of globally. A size and colour combination that moved steadily in the US sat dead in the UK, and because it hung off the same parent it kept eating FBA storage and pulling the catalog toward the aged inventory surcharge. We split the parent by marketplace, removed the dead children from the UK catalog, and shifted that storage allocation to the two variations already converting there.

The part teams skip is the second half: the slow mover has to come out of the inventory plan, not just off the listing. That is where the carrying cost actually sits.

Local assortment is harder for independent stores than for chains, because they rarely have the buying scale or data to test a range. An approach that tends to work is building a core range from what the store's own customers actually buy, then testing a small set of adjacent items each season rather than reshaping the whole shelf. Smaller minimum orders let a store trial a new brand without committing its whole budget. Reviewing sell-through monthly and cutting slow lines early frees space for what the neighborhood is asking for. The principle is simple: let local demand set the range, and use low-risk ordering to test it.