More Choice Can Make the Assortment Weaker

Retailers often treat assortment growth as customer service. If five options are good, eight must be better. More colors, more brands, more pack sizes, more price points.

Each addition seems to create another possible reason for a customer to buy.

But assortment strength isn’t determined by how many choices you offer. It’s determined by how many meaningfully different reasons to buy those choices represent.

That’s the distinction that matters: productive choice versus duplicate choice.

Productive choice serves a different customer need, use case, price expectation, taste, fit, function, or purchase occasion.

Duplicate choice gives the customer another version of essentially the same solution while asking the retailer to support another SKU, another inventory position, another forecast, another replenishment decision, and another claim on working capital.

That difference is easy to miss because duplicate choice can look healthy at item level. The new SKU may sell. The problem is that some of those sales may not be incremental.

They may simply be redistributed from products you already carried.

A SKU can therefore appear productive while the assortment as a whole becomes less productive.

Imagine a retailer selling insulated stainless-steel water bottles.

The current assortment includes a 500 ml everyday bottle, a 750 ml bottle for longer outings, a premium lightweight model, and an entry-price bottle for value shoppers.

Those four options serve reasonably distinct needs.

Now the buyer adds three more 750 ml bottles from different brands.

All are stainless steel, all have similar lids, all sit within a narrow price band, and all target the same everyday hydration customer. Each new bottle gets some sales.

On a weekly report, none looks like an obvious failure.

Before the additions, the original 750 ml bottle sold 30 units a week. Afterward, it sells 16. The three new bottles sell 8, 5, and 4 units.

Total demand for that type of bottle rose only slightly, from 30 units to 33, but the retailer is now holding four inventory positions to produce those 33 sales.

That’s the mechanism: demand has been divided faster than demand has been created.

The retailer hasn’t necessarily created more customer choice in a useful sense. It has created more inventory competition inside the same need.

That’s why one of the most useful questions in assortment planning is not, “Will this SKU sell?” Many SKUs will sell.

The better question is, “What demand will this SKU add that we don’t already serve?”

That question changes the standard for adding products.

A new item doesn’t have to appeal to an entirely new customer. But it should add something meaningfully different enough to justify its place. Perhaps it opens a lower price point.

Perhaps it adds a premium performance feature. Perhaps it serves a different size requirement or use occasion. Perhaps it reaches a brand-loyal customer you currently lose.

Perhaps it fills a clear gap revealed by customer requests or shopping behaviour.

If the main justification is simply, “This one will sell too,” the retailer hasn’t yet established that the assortment needs it.

The useful mental model is to think in terms of need coverage, not SKU coverage.

A wide assortment can still have narrow need coverage if most of its products are slight variations on the same proposition.

A smaller assortment can have broader need coverage if each option earns its place by serving a distinct reason to buy.

Ten nearly interchangeable options may provide less useful choice than six clearly differentiated ones.

Customers don’t experience an assortment as a spreadsheet. They experience it as a decision. If the differences between products are meaningful, choice helps.

If the differences are weak, choice becomes noise. The retailer then pays for that noise through inventory while the customer may gain very little.

The hidden cost of duplicate choice

The obvious cost of an extra SKU is the inventory investment. The less obvious cost is fragmentation.

Every additional SKU takes a share of available demand. If it brings enough incremental demand, that fragmentation may be worthwhile.

If it doesn’t, the same category demand becomes spread across more items.

Unit sales per SKU fall. Demand can become harder to read because each item has a smaller sales signal.

Replenishment can become less efficient because the retailer holds shallower quantities across several similar items rather than useful depth in a proven item.

That raises the chance of being overstocked in some options and out of stock in others.

Markdown exposure can increase too. When demand is split across near-duplicates, weaker variants can accumulate inventory even while the category looks healthy overall.

Working capital becomes less productive because cash that could have supported stronger sellers, tested a genuinely new need, or funded another category is parked in items that mostly compete with one another.

This is where item-level thinking can mislead.

Suppose a newly added SKU sells $2,000 in a month. That sounds like evidence the decision worked.

But if $1,500 of those sales would otherwise have gone to existing products, the true gain was much smaller than the SKU’s sales suggest.

You may never know the exact amount of transferred demand, but you can look for signals.

Did category sales rise meaningfully after the item was introduced, or did sales simply spread across more SKUs? Did gross margin dollars improve? Did inventory turns weaken?

Did older items lose sales as the new item’s sales grew? Did stock depth become thinner across the set? Did markdowns increase?

Most importantly, did the new product attract a different need, or merely give the same customer another similar place to spend?

This leads to a practical rule:

A SKU earns its place by adding demand, protecting demand, or serving a distinct need. Selling something isn’t enough.

Protecting demand matters because not every good SKU must grow the category. Some products are strategically necessary even if their sales overlap with others.

A recognised brand may prevent customers from leaving for a competitor. A key opening price point may keep the assortment credible.

A specialist size may sell slowly but be essential to the category’s promise.

So the goal isn’t ruthless SKU reduction. It’s intentional differentiation.

The danger is not variety itself. The danger is unsupported similarity.

Test the role before adding the SKU

Before adding another option, force the decision through three questions.

1. What distinct customer need does this serve?

Be specific. “Customers want choice” is not a need. “Customers want a lower entry price,” “customers need a larger capacity,” “customers are asking for a sugar-free version,” or “this brand has a loyal following we currently don’t serve” are useful answers.

2. If we don’t add it, what sale are we likely to lose?

This exposes whether the SKU is solving a real gap or simply expanding the assortment. If the honest answer is, “Probably none, because the customer would buy one of our existing options,” the new item may mostly cannibalise current demand.

3. If it sells, which existing SKU is most likely to lose sales?

Retailers don’t ask this often enough. New products are usually evaluated in isolation, but assortment decisions are relational. Every addition changes the role of the items around it.

Identifying the likely donor SKU doesn’t automatically make the new item a bad decision. It does force the retailer to identify what improvement it expects in return.

Better margin? Better conversion? Stronger brand appeal? Higher average selling price? Access to a new customer segment? A stronger good-better-best structure?

If there is no clear benefit beyond shifting sales, the assortment may be getting wider without getting stronger.

This test can also change how line reviews are conducted. Instead of reviewing a wall of products and asking which ones look appealing, group the assortment by the job each product is supposed to do.

Entry price. Core volume. Premium trade-up. Specialist need. Seasonal use. Convenience. Performance. Fashion. The exact roles will vary by category.

Then look for two things: uncovered needs and over-covered needs.

Uncovered needs are genuine gaps. Over-covered needs are places where too many SKUs are competing for the same demand.

That map is often more revealing than a straight sales ranking. A slow seller may deserve to stay because it’s the only item serving a specific customer requirement.

A decent seller may deserve to go because three other SKUs are doing almost the same job.

That’s the deeper point. Assortment productivity isn’t simply about choosing winners and eliminating losers.

It’s about giving every SKU a reason to exist in relation to the rest of the assortment.

The strongest assortment isn’t the one with the fewest SKUs, and it isn’t the one with the most.

It’s the one where each SKU has a clear role and the difference between adjacent choices is meaningful enough to justify the inventory behind them.

So when considering the next product addition, don’t ask only whether customers might buy it.

Ask whether it creates a new reason to buy or merely creates another place for the same demand to land.

If you can’t clearly explain the new need, the sale you risk losing without the item, and which existing SKU it may cannibalise, you probably don’t yet have a strong enough reason to add it.

AI Solution

Prompt:

Act as a retail assortment strategist. Analyze the assortment information I provide and determine where customer choice is genuinely productive versus where similar SKUs may simply be dividing the same demand across too many inventory positions. Focus on need coverage rather than SKU count. For each meaningful group of similar products, identify the customer need or buying role each SKU appears to serve, flag products that seem to serve a genuinely distinct need, and identify likely duplicate-choice situations where several SKUs appear to solve essentially the same problem at similar price, feature, size, brand-positioning, or use-case levels. For every suspected duplicate, explain specifically why you think demand may be fragmented rather than expanded, and clearly label any conclusion that is an inference rather than something proven by the data. Then apply these three tests to the strongest findings: What distinct customer need does this SKU serve? If it were removed or never added, what sale might actually be lost rather than transferred to another existing SKU? If it sells, which existing SKU is most likely to lose sales? Use those answers to classify each important SKU or cluster as Keep, Investigate, Consolidate, or Gap Opportunity. Don’t recommend removing a product merely because it sells slowly if it appears to be the only option serving an important need, price point, size, use case, or customer segment. Likewise, don’t treat a SKU as productive merely because it generates sales if those sales may have shifted from existing products. Finish with the five most important assortment decisions I should investigate first, ranked by likely impact on inventory productivity, clarity of customer choice, and working-capital efficiency, and tell me what additional sales, margin, inventory, or substitution evidence would most help confirm each decision. Base the analysis only on the information I provide and don’t invent missing facts. Here’s the assortment to analyze: “Insulated water bottles: Brand A 500 ml at $24.99, Brand A 750 ml at $29.99, Brand B 750 ml at $31.99, Brand C 750 ml at $32.99, Brand D 750 ml at $30.99, premium lightweight 600 ml at $44.99, entry-price 500 ml at $16.99. The original Brand A 750 ml sold about 30 units a week before the three additional 750 ml bottles were added. It now sells 16, while the new bottles sell 8, 5, and 4 units a week.”

Output:

The assortment has one very clear pressure point: the 750 ml segment. Before the three additional 750 ml bottles were added, Brand A sold about 30 units per week.

After the expansion, the four 750 ml bottles collectively sell 33 units:

The assortment went from one inventory position producing 30 units to four inventory positions producing 33.

That doesn’t prove the new SKUs are unproductive, but it is a strong signal that demand has been divided much faster than it has been created.

The key number isn’t that the new bottles sell 17 units collectively. It’s that total 750 ml demand increased by only 3 units.

Productive Choice vs Duplicate Choice

The strongest productive choices are the ones whose roles can already be explained without stretching: opening price, mainstream smaller size, larger capacity, and premium lightweight.

The weakest part of the assortment is where the retailer has four answers to essentially the same question:

“I want a 750 ml insulated bottle for around $30.”

Unless Brands B, C and D bring meaningful differences that haven’t been provided, the customer need appears substantially over-covered.

What the sales movement is telling you

Brand A’s 750 ml bottle fell from 30 units to 16 after the other three were introduced, a decline of 14 units.

The three newcomers sell 17 units combined.

So:

• Original assortment: 30 units
• Expanded assortment: 33 units
• Additional sales: 3 units
• Additional inventory positions: 3 SKUs

That is a particularly important distinction.

A retailer looking only at the newcomers could say:

“These three new products generate 17 sales a week.”

A retailer looking at the assortment could say:

“We added three SKUs and generated only three additional category units.”

The second interpretation is much more useful.

We cannot conclude that exactly 14 of the newcomers’ 17 sales were transferred from Brand A. Other factors could have changed during the period.

But the sales pattern is consistent with substantial cannibalisation, and that should now become the working hypothesis to test.

Applying the Three Tests

Brand B 750 ml at $31.99

What distinct customer need does it serve?

None can be identified from the information supplied. It is the same capacity as Brand A and only $2 more expensive.

If it disappeared, what sale might actually be lost?

Unknown. Some Brand B customers might specifically require that brand, but there is no evidence supplied showing brand loyalty or another unique reason for choosing it.

A significant proportion could potentially transfer to Brand A, C or D.

Which existing SKU is it most likely taking sales from?

Brand A 750 ml is the strongest candidate because it was the established product whose sales declined after the expansion.

Classification: Investigate

Do not eliminate it yet. First determine whether Brand B brings customers the retailer couldn’t otherwise capture.

Brand C 750 ml at $32.99

Distinct need?

Not demonstrated. It is the most expensive of the standard 750 ml bottles, but its $32.99 price doesn’t establish a true premium role when there is a clearly differentiated premium lightweight bottle at $44.99.

Lost sale if removed?

Unknown. Its five weekly sales may represent genuine Brand C demand, or they may largely migrate to another 750 ml product.

Likely donor SKU?

Brand A 750 ml, and potentially Brands B and D.

Classification: Investigate

Brand C needs a reason to exist beyond generating five weekly sales.

Brand D 750 ml at $30.99

This may be the clearest duplication candidate.

It sits just:

• $1 above Brand A
• $1 below Brand B
• $2 below Brand C

Without another differentiating attribute, those price differences are unlikely by themselves to establish four clearly different customer needs.

Distinct need?

None apparent from the information supplied.

Lost sale if removed?

Potentially very little if customers perceive the alternatives as substitutes. That remains an inference requiring testing.

Likely donor SKU?

Brand A 750 ml appears most likely, followed by the other standard 750 ml products.

Classification: Investigate, with strongest consolidation potential

Its four weekly sales shouldn’t automatically protect it. The important question is how many of those four purchases would disappear from the retailer versus simply move one foot along the shelf.

Brand A 750 ml at $29.99

This product deserves different treatment.

It already demonstrated approximately 30 units of weekly demand before the other products appeared. It therefore provides the best available benchmark for the underlying need.

Distinct need?

Yes. Larger-capacity insulated bottle at a mainstream price.

Lost sale if removed?

Potentially meaningful because the original sales demonstrate established demand, although some would undoubtedly transfer to the other 750 ml products if they remained.

Likely donor relationship?

The evidence suggests Brand A has probably donated demand to B, C and D rather than primarily gaining demand from them.

Classification: Keep / Benchmark

Unless its margin or other economics are materially inferior, I would use Brand A as the control against which the other 750 ml SKUs have to justify themselves.

Premium lightweight 600 ml at $44.99

This is what productive choice looks like structurally.

It isn’t simply another capacity at another nearly identical price. “Premium lightweight” gives it a different proposition, and $44.99 places it well above the mainstream assortment.

The customer can understand why this product exists.

Classification: Keep

Its unit velocity could even be lower than one of the standard 750 ml bottles and it might still deserve its place because it performs a different job.

That’s why sales ranking alone is insufficient for assortment decisions.

The Five Decisions I’d Investigate First

1. Determine whether four 750 ml SKUs are actually serving four different customers.

This is the highest-priority decision.

At present, the assortment evidence says “probably not.”

Look for SKU-level margin, customer switching behaviour, brand-specific demand, stockout substitution and sales before versus after each introduction.

The most revealing question would be: When one 750 ml brand isn’t available, what do customers do?

If they simply buy another 750 ml bottle, you have strong evidence of substitutability.

2. Test whether one or two of Brands B, C and D can be removed without materially reducing category sales.

Don’t ask which has the lowest sales and automatically delete it.

Instead, ask which contributes the least incremental demand.

The weakest seller isn’t necessarily the least valuable SKU. The least valuable SKU is the one whose customers would most readily transfer elsewhere without harming category economics.

Evidence needed: controlled removal or temporary stockout results, sales transfer patterns, gross margin dollars and customer requests.

3. Compare the economics of 30 units through one SKU versus 33 units through four.

Three extra weekly units may or may not justify three additional inventory positions.

Compare:

• inventory investment
• inventory turns
• gross margin dollars
• stockouts
• weeks of supply
• markdowns
• minimum order quantities
• replenishment complexity

A small sales increase can become economically unattractive if significantly more cash has to sit in inventory to produce it.

4. Protect the roles that are genuinely differentiated.

The entry-price 500 ml and premium lightweight 600 ml shouldn’t be judged by exactly the same logic as the 750 ml cluster.

They appear to broaden need coverage rather than simply SKU coverage.

Confirm that assumption by examining their customer behavior, margin and substitution patterns, but don’t consolidate them merely to reduce SKU count.

5. Look for an actual assortment gap before adding anything else.

There isn’t enough information here to identify a specific missing product, so I wouldn’t manufacture one.

Instead, map the existing assortment by customer job:

• Entry price
• Mainstream smaller capacity
• Mainstream larger capacity
• Premium/performance

Then investigate whether customers are regularly requesting something those roles don’t cover.

Only then does a Gap Opportunity exist.

That prevents the common mistake of spotting whitespace in a product grid and assuming it represents unmet customer demand.

The decision I’d make from the current evidence

I wouldn’t immediately delete Brands B, C or D because we don’t yet know their margins, brand-specific demand or substitution behavior.

But I also wouldn’t accept their existing sales as proof that all three deserve to remain.

The 750 ml cluster should go under immediate assortment review.

You added three inventory positions and category volume appears to have increased from roughly 30 to 33 units. That makes incremental demand, not individual SKU sales, the issue that needs to be proved.

The rule I’d use from here is simple:

Don’t ask, “Does this SKU sell?” Ask, “What sales would we actually lose if this SKU disappeared?”

If the answer for Brand B, C or D is “very few because customers would buy another 750 ml bottle,” you’ve found duplicate choice. And that’s where reducing choice could actually make this assortment stronger.