Cutting range feels like cutting revenue. Every line on a shelf was added by somebody who could argue for it, and every line has at least a few customers who will notice when it goes. That is why assortment tends to grow in one direction only, and why the phrase “SKU rationalization” usually arrives in a business at the same moment as a cash flow problem.
This case study follows a regional retail chain of 41 stores that removed just over half of its active lines across eighteen months and finished the exercise with higher sales, better availability and a materially smaller working capital requirement. The company is not named here, and the figures are presented as a composite drawn from a range review of this type rather than as an audited public filing. The value is in the sequence and the decision rules, both of which travel well to other retailers.
What follows is the method: how the range got that big, what data ranked each line, how the long tail was protected rather than blindly culled, how the exit was sequenced so shelves never looked stripped, and what stopped the range creeping straight back.
In short
- Range grew by accident, not by plan. Roughly 4,100 active lines had accumulated over nine years with no formal exit process, so additions were permanent by default.
- Ranking used four inputs, not one. Unit velocity alone would have deleted profitable niche lines; the team scored velocity, margin per shelf week, basket linkage and substitutability together.
- The tail was cut selectively. Around 320 slow lines were kept because they anchored large baskets or defined the chain’s reputation in a category.
- Sequencing mattered more than the list. Exits ran in four waves by category, with facings widened on survivors before any line was pulled, so no shelf ever looked picked over.
- Working capital released funded availability. The cash freed from dead stock went into deeper cover on the top 400 lines, which is where the sales gain actually came from.
Why the assortment had grown to that size in the first place
The chain had no single decision that created 4,100 lines. It had nine years of small, individually reasonable additions and almost no deletions. A supplier launched a variant and offered introductory terms, a store manager reported a customer request, a buyer wanted a price-point gap filled, and each time the answer was yes.
What was missing was an exit rule. Adding a line required a short business case; removing one required a fight with whoever had championed it. In practice that asymmetry is the entire mechanism behind range bloat, and it shows up in retailers of every size. The same pattern is visible in how brands accumulate sub-brands and endorsement tiers when nobody owns the pruning decision, a dynamic covered in more depth in this guide to brand architecture in retail.
The three additions that did the most damage
Reviewing the additions log, the buying director found three categories of line that had multiplied fastest and contributed least. Pack-size variants were the largest group: the same product in three or four sizes where two would have covered the demand curve. Flavour and colour extensions were second, often introduced for a seasonal promotion and then never withdrawn.
Third were what the team called defensive listings, added purely so a competitor could not claim a point of difference. Several of these had been on shelf for years selling fewer than ten units per store per year. None of the three groups had a review date attached.
What the bloat was actually costing
The cost of a slow line is rarely visible in a category margin report, which is why range bloat survives so long. It appears instead as stock that does not turn, replenishment labour spread across too many picks, shelf space that cannot be given to the products customers actually want, and a distribution centre holding safety stock against 4,100 forecasts instead of 2,000.
The chain quantified four cost pools before touching the range: carrying cost on non-moving inventory, markdown and write-off on obsolete lines, store replenishment hours, and the opportunity cost of facings. That last one is the hardest to measure and usually the largest. A slow line does not just fail to earn; it occupies space a faster line could have earned from, which is a core idea in the broader approach set out in the modern brand playbook for retail and e-commerce.
The data used to rank every line before a decision
The instinct in most range reviews is to sort by unit sales and draw a line. The chain tried that first as a sanity check and immediately saw the problem: the bottom quartile by units included several lines that appeared in unusually large baskets, and a handful that were the reason certain customers visited at all.
So the team built a four-input score instead. Each active line was assessed on velocity, profitability per unit of shelf space, basket linkage and substitutability. No single input could condemn a line on its own.
Velocity, but normalised
Raw unit sales punish anything with a high price or a narrow customer base. The team normalised velocity by comparing each line against the median for its own subcategory rather than against the chain as a whole. A slow-moving line in a genuinely slow category was judged against its peers, not against a fast-moving staple in a different aisle.
They also excluded the first sixteen weeks of any line’s life from the calculation, because new lines have not yet found their level. That single adjustment saved several recent launches from being deleted before they had a fair trial.
Margin per shelf week
Cash margin per unit tells you very little on its own. The measure that decided most cases was margin per shelf week: the cash margin a line generated divided by the linear space it occupied, measured over a full trading week.
This metric turns two very different products into comparable numbers. A high-margin, slow-selling line that occupies one facing can beat a low-margin, fast-selling line that occupies six. It is closely related to the ratios covered in this explainer on inventory turnover and why retailers obsess over it, and the two are best read together because turnover measures speed while margin per shelf week measures whether that speed is worth the real estate.
Basket linkage and substitutability
Basket linkage asked a simple question: when this line sells, what else is in the transaction, and how large is that transaction compared with the store average? Lines that consistently appeared in baskets well above average value were flagged for protection regardless of their own velocity.
Substitutability asked the mirror question: if this line disappeared, would the customer buy something else in our store, buy nothing, or go elsewhere? The team estimated this from three sources: prior stockout behaviour on the same line, the number of comparable products in the same subcategory and price band, and store manager judgement collected in a structured survey rather than an open conversation.
How the four scores combined into a decision
The four inputs produced a placement on a simple grid rather than a single ranked list. The grid mattered because it forced a specific action for each line instead of a binary keep or kill.
| Category | Profile | Share of lines | Action taken |
|---|---|---|---|
| Protect | Strong on velocity or margin per shelf week, high basket linkage | 18% | Widen facings, raise stock cover, never substitute |
| Fix | Adequate demand, weak margin per shelf week or poor availability | 14% | Renegotiate cost, reduce facings, or reposition on shelf |
| Replace | Weak on all four, but the subcategory is healthy | 21% | Exit and reallocate space to a better line in the same need state |
| Exit | Weak on all four, high substitutability, low basket linkage | 39% | Sell through and delist, no replacement |
| Review | Conflicting signals, usually a protected tail line | 8% | Held for twelve months with a hard review date |
The two exit categories together accounted for 60% of lines, which is where the headline “half the range” figure came from once the Replace group was partially backfilled. Note that the Review bucket was deliberately small. A large Review bucket is a sign that a range team is avoiding decisions rather than making them.
Protecting the tail that quietly drives basket size
The most expensive mistake available in a range review is deleting the long tail wholesale. The tail is where a retailer’s reputation for having what you need actually lives, and it is invisible in a velocity report by definition.
Around 320 lines in the bottom quartile by units were kept. They fell into three groups, and each group had a different reason for survival.
Destination lines
Some lines are the reason a customer chooses one store over another. In this chain they included a small set of specialist ingredients and a handful of hardware items that competitors in the region did not stock. Their own sales were trivial. The baskets they appeared in were roughly 2.4 times the store average.
Deleting a destination line does not lose you that line’s revenue. It loses you the trip. That distinction is the single most useful thing a range team can internalise before it starts cutting.
Credibility lines
A second group existed to make a category look complete. A category that stocks only the three fastest lines reads as thin, and shoppers respond by treating the whole aisle as unserious. This is the same mechanism that makes range depth part of the value proposition for discounters and for own-brand programmes, as discussed in this piece on how grocery private label is winning shelf space.
The team kept credibility lines at minimum facings and minimum cover. They were not promoted, not given extra space, and were reviewed annually. The point was presence, not performance.
Lines protecting a customer segment
The third group served specific customer groups whose overall spend was high even though the individual lines were slow. Loyalty data made this group visible: a few hundred households whose annual spend sat in the top decile bought at least one of these lines regularly.
Losing a high-value household to save shelf space on a slow line is a bad trade at almost any volume. The same logic that governs retention economics elsewhere applies here, and it echoes the post-purchase findings in this case study on cutting acquisition cost by fixing post-purchase, where the cheapest growth available came from customers the business already had.
Sequencing the exit so shelves never looked empty
A range cut executed badly looks like a business in trouble. Gaps appear, facings stay unchanged, and customers draw the obvious conclusion. The sequencing was therefore treated as a separate workstream from the analysis, with its own owner.
Four rules governed the rollout. Survivors got their extra facings before any exit line was pulled. Exits ran category by category rather than chain-wide. No more than two adjacent subcategories were touched in the same four-week window. And every exit line was sold through rather than withdrawn, except where a supplier agreed to take stock back.
Three ways to exit a line, and what each costs
| Method | Speed | Cash impact | Customer visibility | Best used when |
|---|---|---|---|---|
| Sell through at full price | Slow (8–16 weeks) | Best margin retention | Very low | Line has steady if modest demand and no expiry pressure |
| Managed markdown | Medium (4–8 weeks) | Margin loss, cash recovered fast | Moderate | Seasonal or dated stock, or space needed for a replacement |
| Supplier return or swap | Fast (2–4 weeks) | Best cash outcome where available | None | Supplier relationship is ongoing and terms allow returns |
The chain used supplier returns wherever the relationship allowed it, which turned out to be about a third of exit lines by value. Negotiating that was easier than expected: a supplier facing a delisting will often prefer taking stock back over a markdown that damages its price positioning across the region.
Reallocating space before removing stock
The operational detail that made the rollout invisible to customers was the order of the two moves. Planograms were redrawn first, extra facings assigned to the Protect group, and the additional stock ordered and delivered. Only then were the exit lines allowed to run down.
This costs a few weeks of slightly congested backroom space and it is worth every hour. The alternative, pulling lines and then fixing the shelf, produces a fortnight of visible gaps in every store, which is exactly when a competitor’s flyer lands.
What the store teams were told
Store managers were briefed with the reasoning, not just the list. Each store received its own before-and-after planogram, the specific lines affected, and a short script for the three or four customer questions the head office expected. Managers could escalate a deletion they believed was wrong, and 27 lines were reinstated on that basis.
That escalation route did more than correct 27 errors. It stopped the exercise being experienced as something done to the stores, which is usually what determines whether the new discipline survives contact with the second year.
Working capital released and where it went
The financial case for a range review is usually made on working capital, and this one was no exception. Removing slow lines converts stock into cash once, and then permanently reduces the stock required to run the same sales.
The chain released a mid seven-figure sum in inventory over the eighteen-month programme. Roughly 60% of that came from the one-off sell-through of exit lines, and the remaining 40% from the structural reduction in safety stock across the distribution centre once there were fewer forecasts to hedge.
Why the money did not go to the balance sheet
The decision that shaped the results was what happened to the released cash. The finance team’s default position was to pay down the revolving facility. The commercial argument, which won, was that a meaningful share should be reinvested in stock cover on the Protect group.
The reasoning was straightforward. Availability on the top 400 lines had been running in the low nineties, and every point of availability lost on a fast line costs far more in sales than a slow line ever earned. Buying deeper cover on those lines was the highest-return use of the cash available to the business at that moment.
The split that was actually agreed
The final allocation put roughly 45% into deeper cover and improved replenishment frequency on the Protect and Fix groups, 35% into debt reduction, and 20% into a category refresh programme that funded the Replace backfills. That third bucket mattered because Replace lines need to be genuinely better than what they succeeded, and a backfill funded from nothing tends to be whatever the incumbent supplier offers.
Retail inventory levels are tracked in aggregate by public statistical agencies, and a business running a range review can usefully benchmark its own stock-to-sales ratio against the sector series published by the US Census Bureau. Sector averages will not tell you what your range should look like, but a stock-to-sales ratio far above the sector norm is a strong signal that the range, not the demand, is the problem.
Availability, sales and margin twelve months later
The measurement period ran for twelve months from the completion of the final exit wave, compared against the same twelve months two years earlier to avoid contaminating the baseline with the transition itself. Like-for-like store count was held constant by excluding two stores opened during the period.
| Metric | Before | Twelve months after | Change |
|---|---|---|---|
| Active lines | ~4,100 | ~1,980 | Down 52% |
| Like-for-like sales | Baseline | Up | +4.1% |
| On-shelf availability, top 400 lines | 91.6% | 97.2% | +5.6 points |
| Inventory units held | Baseline | Down | -31% |
| Gross margin rate | Baseline | Up | +70 basis points |
| Markdown as share of sales | Baseline | Down | -90 basis points |
| Store replenishment hours per week | Baseline | Down | -12% |
Where the sales growth came from
Attribution work suggested the sales gain was not evenly spread. The great majority of it came from availability on the Protect group, where lines that had previously been out of stock for part of most weeks were now consistently on shelf. A smaller contribution came from the Replace backfills, several of which outperformed the lines they succeeded.
There was a real offset. Roughly 1.8 points of sales were lost on deleted lines that customers did not substitute, which is the honest cost of the exercise. The net figure of 4.1% is after that loss, not before it.
The margin story is mostly markdown
The 70 basis point gross margin improvement looks like better buying, and buying terms did improve modestly on the consolidated volume. Most of it, though, was the disappearance of chronic markdown on lines that were never going to sell at full price.
A range that is 52% smaller has far fewer opportunities to end up with obsolete stock. That is a structural margin benefit rather than a one-off, and it continued into the second year at a similar level.
What did not improve
Two things were flat. Basket size barely moved, which surprised the team given the basket linkage work, and customer counts were unchanged. The reasonable interpretation is that the tail protection successfully avoided losing trips rather than gaining any, which was the objective.
Supplier relationships took a measurable hit in the first six months. Two suppliers withdrew promotional support, and one delisted the chain from a national activity. Both effects had faded by month twelve, but a business running this exercise should budget for a period of commercial friction rather than assume it away.
Rules that stopped the range creeping back
Range bloat is a process failure, so a one-off cut without a process change simply resets the clock. The chain adopted five rules at the end of the programme, and they are the part of this case study most worth copying.
Rule one: every new line requires a nominated line to exit, or an explicit written exception approved at director level. This is the single highest-impact change and it is also the one businesses abandon first.
Rule two: every listing carries a review date, no more than eighteen months out, recorded in the item master rather than in a spreadsheet. A line with no review date cannot be created.
Rule three: the four-input score runs quarterly and automatically, not annually as a project. Removing the project framing was deliberate; a review that requires a mobilisation never happens on time.
Rule four: the Protect list is published internally and its availability is reported weekly alongside sales. What gets reported keeps its stock cover.
Rule five: exceptions are counted and published. The chain found that simply publishing the number of exceptions granted per buyer per quarter reduced them by more than any approval threshold did. These operating disciplines sit alongside the wider commercial choices described in the modern brand playbook, and they work because they change the default rather than relying on periodic willpower.
The one rule they got wrong
The chain initially set a hard cap on total line count, and it caused problems within two quarters. A fixed cap makes range a queue, and buyers began timing their submissions rather than judging them on merit. Categories with genuinely expanding demand were held back by categories that should have been shrinking further.
The cap was replaced in year two with category-level margin per shelf week targets. That let strong categories grow and forced weak ones to justify their space, which is what the cap was trying to achieve in the first place.
A note on how far this travels
This was a bricks-and-mortar chain, where shelf space is a hard physical constraint and the cost of a slow line is unambiguous. Online retailers face a softer version of the same problem: infinite shelf space, but real costs in photography, content, forecasting, returns handling and the dilution of search relevance within their own catalogue. The four-input score adapts, with margin per shelf week replaced by margin per unit of pick and hold cost, and with basket linkage doing considerably more of the work.
The Pareto pattern underlying all of this, where a minority of lines drives the majority of sales, is a general commercial regularity rather than a retail-specific one, and the background is summarised in the Pareto principle. What retail adds is the tail: the small set of slow lines whose removal costs you the trip, not just the line.
FAQ on SKU rationalization
What is SKU rationalization?
SKU rationalization is the structured review of every product line a business sells, scoring each one on its commercial contribution and then keeping, fixing, replacing or removing it. The goal is not a smaller range for its own sake. It is to reallocate finite shelf space, working capital and replenishment effort toward the lines that earn most from them.
How much of a range can safely be cut?
There is no universal percentage, and any consultant quoting one before looking at your data is guessing. Businesses that have never run a formal review commonly find 30% to 50% of lines contributing very little, while a range that is reviewed quarterly may have almost nothing to remove. The defensible answer comes from scoring the range, not from a target set in advance.
Will cutting lines reduce sales?
Some sales are always lost, because a share of customers will not substitute. In this case study roughly 1.8 points of like-for-like sales were lost on deleted lines and were more than offset by availability gains on the lines that were kept. Whether the net is positive depends almost entirely on whether the released capacity is reinvested in the surviving range or simply banked.
Which metric matters most when ranking lines?
Margin per shelf week decided the largest number of cases here, because it compares products of different sizes and price points on the same basis. It should never be used alone, though. Basket linkage is what stops a scoring model from deleting the slow lines that bring high-value customers through the door.
How do you identify a destination line before deleting it?
Look at basket composition rather than line sales. Pull every transaction containing the line over twelve months and compare average basket value against the store average, then check how many of those baskets come from repeat households. A line whose baskets run well above average and skew to loyal customers is a destination line, whatever its own velocity says.
How long does a range review of this size take?
The analysis in this case took about ten weeks including data cleanup, which is usually the slowest part. The rollout ran eighteen months because exits were sequenced in waves by category rather than executed at once. Compressing the rollout is possible but it raises the risk of visible gaps and supplier friction happening everywhere simultaneously.
What happens to supplier relationships during a cut?
Expect friction, particularly from suppliers losing several lines at once. This chain saw two suppliers withdraw promotional support and one pull the business from a national activity, with the effects fading within a year. Giving suppliers advance notice and offering stock returns rather than markdowns reduces the damage considerably, because a markdown harms their price positioning across the whole region.
Does the same approach work for online retailers?
The framework transfers, but the cost of a slow line is different. Online, the constraint is not shelf space but content production, forecasting accuracy, returns handling and internal search relevance, so margin per shelf week is replaced by margin net of pick, hold and return costs. Basket linkage carries more weight online because substitution behaviour is easier to observe directly.
How do you stop the range growing back?
Change the default rather than relying on periodic reviews. The most effective single rule is that a new listing requires a nominated exit or a written director-level exception, combined with a mandatory review date recorded against every line in the item master. Publishing the count of exceptions granted per buyer proved more effective at this chain than any approval threshold.