Ecommerce Merchandisers Share How to Trim Assortments Without Losing Sales

We've learned that curating the perfect menu is all about balance. When we streamline our assortment at Equipoise Coffee, we use a single guiding rule: Every variant must serve a distinct, non-negotiable customer ritual. If a product variant does not satisfy a unique taste profile or brewing method that can't be easily replicated by another, it gets cut.

When we evaluate our single-origins like Mexican La Laja Honey, Ethiopian Yirgacheffe, or our signature Cavaliers Blend, we look at why people buy them. We don't need five different bag sizes or redundant grind options that clutter our site. We focus on the core offerings that represent our dedication to roasting science. By removing the middle-ground variants that don't clearly map to a specific brewing method, we make the decision process simpler for our home brewers.

This rule protects our basket size because we pair it with clear communication. We explain these tradeoffs directly to our customers through our educational content. Instead of offering endless minor variations, we guide them on how to achieve their desired cup with our core offerings. If we cut a specific pre-ground option, we educate them on why buying whole bean and grinding fresh elevates their morning ritual.

To protect search relevance, we never delete the search footprint of a retired variant. We redirect those queries to our comprehensive brewing guides or our main collection pages. This keeps the traffic flowing and builds trust. Our customers appreciate the clarity, and we avoid the decision fatigue that kills conversion rates. It is a simple, clean approach that keeps our small-batch operations running smoothly.

When we streamlined our menswear assortment, the rule that proved durable was to cut by demand concentration, not by gut feel about which variants should sell. We pulled the data and found the familiar pattern: a small share of size and color variants drove the large majority of sales, and a long tail of variants each sold almost nothing while still costing us in inventory, photography, and choice overload at checkout. The rule of thumb we settled on: protect the variants that together make up roughly the top 80 percent of units sold, scrutinize everything below that, and only keep a slow variant if it serves a real strategic purpose, like completing a size range so we do not look like we exclude certain bodies. That last caveat is what kept the cull from hurting us. We did not delete slow variants blindly, because removing the smallest and largest sizes would have cost basket size and search relevance from people who needed the full range and shopped the whole brand because of it. On search specifically, we kept the variants whose pages actually ranked and got found, and removed the duplicates that split relevance. The durable rule: cut the true dead tail by the numbers, but never cut a variant that protects size inclusivity or a page's search footprint, even when its own sales look thin. Fewer, better-chosen variants raised conversion and cut cost without shrinking the baskets that mattered.

My catalog kept growing by addition. A supplier would send a new size option, a new color, a slight formulation tweak, and it would go live. But when I looked at how customers bought across my channels, the variants selling in thin, unpredictable trickles were pulling down the listings next to them. I watched those weak siblings dilute review velocity on parent pages, split click-through across too many options, and make the whole product family look indecisive to both algorithms and shoppers.
So the rule I landed on was blunt. If a variant can't sustain its own purchase frequency at a pace that keeps it in stock without manual intervention, I cut it. Sporadic sellers create restocking headaches, age out of warehouse slots, and drag conversion rate down on the entire product family.
Cutting them felt risky because each one technically had some revenue attached. But once those listings tightened up, the remaining variants sold at a higher pace, collected reviews at a higher pace, and ranked better organically. I could see the lift in the parent listing within weeks of removing the dead weight.
The durable part is that the rule works the same whether I'm selling on my own site, through a marketplace, or into a retail account. I apply the same purchase-frequency threshold everywhere I list.

We started by mapping variants against shopper intent instead of internal attributes. We then measured how much they overlapped in search results. If two or three variants attracted the same queries and one consistently won the sale, the others created friction. Before removing anything, we checked if weaker variants helped increase basket value through bundles, attachments, or repeat behavior.

The rule that proved durable was simple: Every variant had to earn its place by winning discovery or order value. If it did not succeed across the season, we treated it as clutter. This approach helped us build a cleaner assortment and improve search performance overall.

At EV Cable Hub we had drifted into stocking too many near-identical cables, small length and connector variations that looked like choice but were just clutter. When I trimmed the range, the rule I held to was to keep any variant that pulled its own search traffic or solved a real fitment need, and cut the ones that only existed because a single customer had once asked. A variant has to earn its slot by being found and bought, not by being theoretically available.

The way I protected search relevance was to look at which exact lengths and connector types people searched and landed on before removing anything. A 5 metre and a 10 metre Type 2 each bring in their own buyers, so they stay. A handful of in-between lengths nobody searched for were quietly retired and the demand simply consolidated onto the standard options either side, so basket size held rather than fell. The fear that pruning loses sales usually assumes every line is wanted, when a lot of them are just dead weight you are paying to carry.

The durable rule that survived a full season was this: if a variant had not sold in 90 days and brought no search traffic of its own, it went. That single test cleared the dead lines without me agonising over each one, and the season after, the slimmer range was easier to merchandise, faster to keep in stock, and converted better because buyers were not paralysed by twelve almost-the-same choices. Fewer, clearer options nearly always beat a long list that exists only to look complete.

• How Retail Teams Make Omnichannel Inventory Tradeoffs That Protect Sales
• Ecommerce Returns: How Leaders Balance Instant Refunds With Risk
• Make Better Retail Pricing and Discount Calls Without Hurting Brand