How Retailers Build Assortments Customers Want

I keep an assortment fresh by adding in small batches and letting repeat purchases, not first clicks, decide what stays. Start a new product with limited stock and judge it after 30 to 60 days on one test: do customers come back for it? Slow sellers tie up cash, so set the rule in advance. For example, review any product with no sale in 90 days instead of deciding on instinct. To avoid overwhelming shoppers, keep each category to a handful of clear options and add something only when it fills a gap, not because it is new.

For a marketplace like Suspire, with clothing, kitchen and home products, this means a new brand partner is a trial before it is a commitment. When we locked in this "test-and-repeat" buying rule, our inventory turnover improved by 41%, slow-mover cash lock-up dropped by 35%, and the average order value on curated categories rose 22% within two quarters - proof that disciplined assortment decisions compound into healthier growth without flooding the shopper.

I would separate variety in the catalogue from money committed to stock. I once over-ordered 200 units of one jewelry piece. That experience changed how I approach adding products: I now make each piece to order after the customer has paid.

The lesson was that choosing an appealing design and proving demand for a large quantity are different decisions. I would keep a new assortment focused on distinct customer needs, then use paid orders to decide where to expand. A near-identical variation needs a reason to exist beyond making the catalogue look bigger.

Making to order does not remove production constraints, but it avoids making the same large inventory bet before demand is clear. For another retailer, my advice is to make the first stock commitment small enough that a mistaken forecast will not prevent the next buying decision.

I decide what to build next off my comment section. With 35k-plus subscribers, I keep a running list of what people ask me for by name, unprompted, more than once. The ideas I'm personally most excited about sit at the bottom of that list until somebody else brings them up.

I use the same test on the AI products I'm launching now. I don't watch first signups, because anybody will try a new thing once. I watch whether they come back and use it a second time without me reminding them. If the second use doesn't happen, the product is a demo, not a line.

Translated to a shelf, that's the test I'd put on every add and every cut. An item somebody buys once out of curiosity and never reorders is taking up attention and working capital without earning either. An item customers walk in and ask for by name has already paid for its slot before you commit the purchase order.

When I consulted on product development with large companies, the long assortments I saw were usually built from internal enthusiasm and whatever a supplier happened to have available. The cheap fix was to make somebody produce the customer request in writing before the buy was approved. Most of the time nobody could.

1. We view an assortment of products to sell through the same lens as a great show home. People want to be able to imagine living in a property, not have to start counting rooms! I try to keep the core offering of services that returning customers know and trust at CheapForexVPS very small and constant. The one "new" offering that we reserve for our assortment would rotate on a regular basis and would only be committed to the waitlist or even for pre-order only if there is sufficient interest first. I'd rather miss out on some growth potential than have money languishing in a sub-optimal offering that nobody asked for in the first place. Retirement of an offering would be communicated as early as possible to all existing customers and they would be moved to alternative offerings on the same terms as soon as is practical - how an offering is removed is more telling of a company than how an offering is introduced.

2. We attempted to design a budget entry point for our 'newer traders' some time ago. We thought it would be low margin, low use 'filler' for us. In reality, it became one of our most talked about products, not due to the specs, but due to the support we provided with setup of the trading platform around the clock. The support chats with customers were very revealing. People weren't worried about RAM or cores, but about setting up to trade at 2am without losing too much money! This gave us lots to think about. Before we add another feature to a product, we now ask, what worry does it remove from a customer's perspective and not just what additional feature does it add.

I can speak to how we decide which product lines to feature and how much attention each one gets.

The mistake is treating every product like it deserves equal space and explanation. When a buyer lands on a site with too many options presented the same way, they slow down or leave. I would rather lead with the products people actually ask about and make the less common options easy to find for the buyers who need them.

The best signal I have found for what customers truly want is what shows up in quote requests and sales conversations. When the same product type keeps coming up, like charms, jewelry pieces, or ornaments alongside our pins and patches, that tells us where to put more content, better photos, and clearer options. Watching what people ask for is a lot more reliable than guessing what might sell.

I treat assortment decisions as cash allocation above all else. Every SKU on a shelf or a product page is money I can't spend on something else, so I give each one a window to prove it. If a product hasn't earned its shelf space inside that window, it goes, regardless of how much I personally liked it when we bought it.

The mechanism that made this work for my team was a hard cap on total SKU count. New items only come in when something leaves. That single rule forced real arguments about what deserved a slot, and it kept the catalog from bloating into the kind of choice overload where a shopper browses, hesitates, and leaves without buying anything.

The buying decision that taught us the most was one where we bet on breadth. We added a wide range of variations in a category we thought customers wanted options in, and the orders clustered almost entirely around a couple of them. The rest just sat there absorbing cash.

We cut the range back and put the freed-up budget into testing new categories instead. Now freshness comes from rotation at the edges while the proven core stays stable, so shoppers see something new without having to re-learn the store every visit.

We bought a full size run of one style of boot in a very deep green, because the picture with the supplier made it look good. It idled on the shelf for months, as nobody wanted them in that particular shade, and the cash was tied up in a shelf of pairs, which I couldn't shift without a discount.

So, I decided to make one change. A new style can have a limited number first order in a few sizes and wait a couple of weeks before taking another order to see what happens before taking a big order. If it's popular, then it gets a big order. If not, I've only lost a little.

It's costing me something. Popular sizes selling out before a style has been proved is sometimes happening, and I have lost a few customers as a result. I can live with that more easily than with a green boot.

Achieving a balance between a new retail assortment and cash flow necessitates transcending gut instinct and integrating the ITR directly into the day-to-day operations dashboard. During my experience managing over 50 enterprise system implementations, a major reason retailers bind their cash in slow-moving stock has been the inability to confirm the SKU-level transparency regarding the speed of capital return to business. By integrating the ITR into the ERP, the emphasis shifts from what the purchase team thinks will sell to what the data shows is being sold preventing the error of over-ordering based on past trends that do not correlate to changing consumer behavior.

I witnessed this change in operations enabling teams to make different decisions regarding products to be culled. In one case of my large-scale retail implementation, the purchasing department realized through the ITR dashboard that the so-called bestsellers are actually killing the margins because storage costs combined with the slow turnover pace do not bring profits anymore. The team discovered that users do not need a broader choice but better availability of fast-moving goods. By utilizing this information for aggressive cutting of the slowest sellers, a lot of funds were released into investing in bigger supplies of the main products and trials of niche ones without increasing the total inventory expenses.

To keep the assortment up to date without overwhelming the customers, the retailers shall follow a one-in-one-out principle defined by turnover thresholds. This means that if a new product is introduced or bought, the worst-performing SKU in this category needs to be pulled out within the 90-day ITR window. This rule is to maintain the catalogue manageable for the customers and make sure that every dollar of the inventory works.

I keep assortment tight by focusing on core aesthetic pillars -- techwear, cyberpunk, Y2K and gothic -- and rotating limited, curated drops rather than broad, permanent ranges. That lets us introduce fresh pieces frequently without carrying large backstock, while analytics and on-site engagement tell us which motifs and fits convert.

A key buying choice was to prioritise functional, utility-led pieces and accessories over pure novelty. By narrowing to items that combine futuristic design with practical features we confirmed customers buy for both style and usefulness, which guides subsequent buys and reduces slow sellers.

One lesson from supplying Australian retailers is that more product choices don't necessarily create a better shopping experience. Too many similar options can occupy valuable display space without generating proportional sales.

At Mills Shelving, we've focused our range on modular shelving systems and compatible accessories rather than relying entirely on fixed configurations. Retailers can introduce new products by adjusting existing displays instead of investing in another complete fit-out.

I'd recommend testing new categories in a limited display area, monitoring sell-through and only expanding once demand is established. Unsold inventory already costs money; it shouldn't require additional fixtures and floor space as well.

The way we keep an assortment fresh without overwhelming customers or tying up cash is by treating every new product or feature like a test, not a permanent addition. The mistake many teams make is adding too much at once because variety feels like progress. In practice, too much choice slows decisions, makes merchandising messier, and leaves you investing in things customers only said they wanted but do not actually choose.

My rule is simple: add narrowly, measure quickly, and remove without emotion. We look for a small set of signals before expanding anything, such as clicks, conversion to trial or purchase, repeat usage, customer support questions, and whether the product creates demand for related items or just sits there. If something gets attention but not commitment, that usually means it is interesting, not valuable.

One buying decision that taught us a lot was resisting the urge to broaden our offering too early. In SaaS and digital product building, the equivalent of a retail assortment problem is feature sprawl. We had moments where it was tempting to add more options because a few users requested them. Instead, we kept the core offer tighter and tested demand in smaller ways first. What we learned was that customers rarely wanted the biggest catalog. They wanted the clearest path to the outcome they came for.

That changed how I think about assortment. A fresh assortment is not the one with the most newness. It is the one where each item has a job: bring people in, convert them, complement a top seller, or increase retention. If a product does none of those, it is taking up space, attention, and working capital.

The best buying teams stay disciplined by reviewing assortment regularly, setting exit criteria before launch, and protecting room for winners instead of defending slow sellers for too long.

We separate discovery inventory from proven inventory. Discovery buys are limited and assigned a short evaluation period. That prevents enthusiasm from becoming a warehouse problem. It also gives the team permission to test demand signals without pretending that every experiment deserves a commitment.

The key is deciding in advance what success means. We set targets for sell-through, margin, returns, and customer engagement. An item can sell steadily and fail if it creates confusion or attracts the wrong audience. Conversely, a smaller item can earn depth when it brings decisive behavior. This approach keeps our assortment responsive because investment follows validated momentum, not the optimism that often surrounds a new idea.

Buying a fresh assortment is really hard for a few reasons. First, it's incredibly expensive to just cover "the basics". Partnering with great vendors who can support a count-and-fill stock business is essential to reduce markdowns and buy as you need to keep cash available to take chances. I also believe in buying about 20% of the seasonal OTB in "marketing" items, pieces or brands that have a great following and visibility, to see if clients transact on the hype. Also, it's helpful to take a proven winning silhouette, and buy both a classic fabric and a fresh trend pattern and color. I'll usually buy the classic 2 or even 3 to 1 over the trend piece. The final thing that's helpful is to talk to the vendor about fabric availability and time to remake an item. If you know a fresh piece might be great, bring it in early, test the market, and set yourself a reminder to re-order the piece so that it can deliver before your season really gets rolling. For instance, if it's 2 months to re-produce and the vendor is likely to have extra fabric, I'll try to get it in for July 1, and if it hits we can place a re-order by 8/15 and get it in November (post Italian August vacation), and be ready for American Holiday shopping.

I'd start with the rule that every product has to earn its place under one of our four pillars at APMZEE, Action, Performance, Movement and Sleep. That decides what gets added before anyone argues about taste: what does this do for someone's week, under which pillar, and does it overlap with something we already sell? If it overlaps, it replaces something or it doesn't launch. It keeps the range readable for a shopper in their forties who doesn't want to compare thirty labels. Cash is the other half of the decision, and supplements make it sharper because stock has a shelf life, so a slow seller costs more than its unit price, since it ties up money and then expires. I'd rather launch later with one format people understand, like creatine gummies, than put out three versions of one idea. The pillars are still the first filter anything new goes through.

I'm Charles Liu, founder and Marketing Director of Cubic Promote, supplier of branded uniforms, giveaways and corporate gifts. For us, the key is expanding the message without abandoning what customers already know us for. We have traditionally ranked well and been known for promotional giveaways, but we have been putting more effort into showing that we also supply corporate gifts and workwear. The challenge is making sure customers do not interpret that as us moving away from promotional products altogether.

So whenever we update our messaging, we continue mentioning giveaways alongside corporate gifts, staff uniforms and workwear. We try to clearly show the full range rather than replacing one category with another. That has helped us preserve familiarity for clients who already order promotional merchandise from us, while also making them more aware that they can come to us for things like gift hampers or staff uniforms. The lesson for me is that a messaging refresh should broaden what customers associate with your brand, not erase the part of the brand they already trust.

Charles Liu Founder and Marketing Director Cubic Promote

We judge an assortment by contribution per unit of working capital, not by how full the range looks. On a marketplace the carrying cost is explicit, so the maths is easier than it is in a store. Every SKU has a sell through rate and an age in the FBA Inventory Age report. Aged inventory surcharges start once units cross the long storage threshold. Your Inventory Performance Index score sets your restock limits. So a slow seller does not merely tie up cash, it takes storage capacity away from the SKU that would have turned. That is the cost most buyers never price in. The rule we use is blunt. A SKU earns its place by clearing a minimum weekly sell through at a price that still carries its ad cost and its referral fee. If it misses, it gets one intervention on content and price, and then it gets cut. On overwhelming shoppers, the lever people miss is the variation family. Ten standalone listings for ten colours splits the reviews, splits the sales history, and hands the shopper ten decisions. One parent with ten children consolidates review count onto a single page and gives the shopper one decision followed by an easy one. Same assortment, far less cognitive load, better organic placement. What taught us most was watching which children in a family actually converted. Buyers keep choosing the middle of a range, not the cheapest and not the top. Depth belongs in the middle. The ends exist to make the middle look like the sensible choice. Jimi Patel, Director, eStore Factory LLC

The hardest buying decision is saying no to a good product, because every item I add ties up cash and shelf space and adds one more question my team has to answer by email. What I check is whether a customer who arrives not knowing her hair can find the right product quickly. Our UK Curl Report found that 63% of UK women have wavy, curly or coily hair, and that 24% of women with textured hair still aren't sure of their pattern, so the shopper's problem is usually too much choice and not too little.

I look at the range by curl pattern and need. If two conditioners suit the same 3B hair, one has to earn its place with a clearly different formulation or price, or it goes. I'd rather reorder shallowly on anything new until it has sold through once, and let the questions customers send us show where the gaps are, because the inbox is the most honest research I have.

I'd keep a range fresh by adding alternative versions of plants it already carries, at a different origin or price, and explaining how they differ so more choice doesn't become confusion. In June 2026 we added a Bulgarian Helichrysum italicum alongside our Corsican one, and its launch post presented it as a way for more people to experience the plant. In 2020 we offered neroli both pure and as a 10% dilute to make it more affordable, and by April 2023 we offered five chamomile oils, including a CO2 blue chamomile added early that year, with a blog post comparing them. I'd also let some variety come on request: we offer to make a blend or dilution for a customer who asks, and any of our blends can be made into a body oil. For learning what customers want, I'd listen to their questions: in 2023 customers often asked me which of our juniper berry oils suited them, so I wrote a short guide to help them choose.

I keep our range fresh by recombining styles we already stock before I commit cash to anything new. Our bundles group existing pairs: The Changing Tides, for example, puts Odessa, Shanti and Lilah together in one set, and each of those three can also be bought on its own. A bundle gives shoppers something new to look at without a new design to develop or buy in, and choosing stays simple because there's nothing unfamiliar to weigh up. The buying decision behind it is to hold off on a new style until we've seen which combinations people actually pick, rather than guessing at a new design and finding out from the slow sellers. Freshness doesn't always mean more products; sometimes it means a better way into the ones you already have.

We keep the range deliberately small: the five categories riders need on every ride, which are helmets, jackets, gloves, boots and pants, instead of thousands of products from hundreds of brands. Most of the catalogue is listed through Shopify Collective, which lets us sell other Shopify makers' products without buying stock up front, so trying a new style doesn't tie up cash in slow sellers. Inside those five categories every product is hand-selected, and if it doesn't meet our standards we don't carry it. The reasoning is simple: when a shopper scrolls through dozens of jackets and still doesn't know which one to buy, a shorter, curated list is part of the service.

At Hey Horst GmbH, we operate five automated retail machines at two locations in Bavaria, Germany. Our biggest lesson has been that customers often buy very differently from what we initially expected.

We use real-time vending telemetry and sales data to identify bestsellers and slow-moving products. Since every machine has limited space, each product needs to justify its position in the assortment.

One particularly interesting discovery was that turkey products performed better than pork products in certain categories, while some traditional grocery staples, such as butter and yogurt, attracted less demand than expected. At the same time, energy drinks, popular soft drinks, and regional ready-to-eat meals proved to be important parts of our assortment.

These insights changed our buying decisions. Instead of ordering large quantities based on assumptions, we test new products in smaller batches, monitor their sales performance, and expand only when customers demonstrate genuine demand.

We also adapt the assortment to each location rather than offering exactly the same products everywhere.

My advice to other retailers is simple: don't confuse a large assortment with a good assortment. Customers appreciate relevant choices, not unnecessary complexity. Let actual sales data guide your decisions, keep testing, and don't hesitate to remove products that aren't performing.