How to Simplify a Product Range Without Losing Customers (2026)

Product range simplification is the deliberate work of removing products, variants or SKUs that add complexity without adding value, while keeping every customer need served by something you still sell. Doing it without losing customers takes a method, not a cull: score each item, map what each buyer was really after, then give that buyer an alternative before the original disappears.

That last part is where most range cuts go wrong. The operations side is well understood and badly executed in equal measure; the customer side is barely discussed at all, and it is the side that decides whether simplification ends with a tidier catalogue or a quieter business. Over the past few years I have watched teams with 400, 900, 2,000-plus SKUs try to cut 20% and end up defending an angry email inbox instead of celebrating a leaner P&L.

So here is the honest version of how to simplify a product range without losing customers. It takes seven steps, two of which (mapping by customer job, and protecting the buyers of what you remove) have almost nothing written about them anywhere. Budget three to six weeks for a range of a few hundred SKUs. The work is analytical rather than difficult, but it is unglamorous, and the discipline of writing down what you will not cut matters more than the discipline of what you will.

What You Need to Simplify a Product Range Without Losing Customers

The analysis needs four inputs, and the fourth is where most programmes start without it.

• SKU-level performance data. Twenty-four months minimum, if you can stretch to thirty-six. You need units, revenue, gross margin, returns, sell-through rate, inventory turns and stockouts per item, not per category. Twelve months is thin: it hides seasonality, and a seasonal SKU cut on a quiet-year read is how you end up out of stock in your best month.
• Cost-to-serve per SKU. Setup and changeover minutes on the line, picking and packing time, freight and duty, returns handling, customer service contacts, and the marketing or content spend that exists only because that product exists. Contribution margin after cost-to-serve is a very different number from revenue, and the gap is where your worst surprises live.
• Customer-side evidence. Search terms that land on each product page, “customers also bought” paths, sales-call objections, review text, support tickets, and the questions your team answers repeatedly. This is the input that tells you what a product is actually for, which is rarely what the spec sheet says.
• Substitution history. For every at-risk item, what do buyers buy instead? If you have click data, look at what follows a page view of a product you did not buy. In B2B, ask the account managers directly. This is the dataset almost nobody assembles and the one that does the real work later.

Two more things help. Write down a single commercial objective before you touch the numbers, because a range cut aimed at freeing working capital behaves very differently from one aimed at lifting conversion. And be honest about whether you are cutting breadth or depth. Breadth is how many options you offer. Depth is how many variants, colours or sizes sit behind each option. Cutting depth is usually cheaper and safer than cutting breadth, and most businesses try it in the wrong order.

Step-by-Step: How to Simplify a Product Range Without Losing Customers

1. Set the purpose and guardrails for the range

Simplification is not an end in itself, so write down what the range has to do for the business before you decide what to remove. Three or four lines: release working capital, cut changeovers, make the bestsellers easier to find, prepare for a channel change, exit a supplier. Each objective produces a different cut list from the same data.

Then name the groups you are not allowed to disappoint. For most businesses that is a short list: key accounts with contracts, the buyers who make up most of repeat revenue, the entry-price product that brings people in, and anything regulated, safety-critical or warranty-bound. Put the list in the document. If a proposed cut touches a protected item, it needs an explicit override, not a shrug.

Pick your measures now as well. Revenue per SKU, SKU productivity, gross margin return on inventory, conversion rate, and repeat purchase rate will tell you different things, and the last two are the ones that catch over-cutting. A range cut that improves every operational metric while repeat purchase falls was a mistake.

2. Audit performance, customer needs and each product’s role

Run the analysis in this order: classify, then score, then argue. Classify first, because the same sales number means completely different things depending on what role an item plays.

Role	What it looks like	What low volume usually means
Traffic driver	Low margin, high new-customer share, often the cheapest item in the range	Working as intended; the sales report hides the value
Bestseller	Top quartile on units and usually on contribution	Rarely relevant; protect these absolutely
Attach or accessory	Sells mainly with something else, tiny standalone revenue	Do not judge it on its own sales line
Niche or specialist	Small audience, high loyalty, strong review scores	May be your most profitable customer segment per pound spent
Key-account item	Low or no public demand, high account dependency	Concentration risk dressed up as a slow seller
Duplicate or cannibal	Overlaps another SKU on the same job and price point	The clearest and safest cut in the whole range

Now score each SKU on four things: contribution margin after cost-to-serve, sell-through and inventory turns, role (from the table above), and substitution risk, meaning how easily a buyer can move to something else. A simple two-axis plot of contribution and substitution risk sorts the range into four quadrants and takes about an hour once the data is clean.

The practical instruction: score every SKU on all four criteria, not only the ones that are obviously weak. Scoring only the weak ones is how ranges end up with a quiet layer of unexamined products that nobody will defend in a review.

3. Map the range by customer job and substitution

This is the step that protects customers, and it is the one almost every guide skips. Organise your range by the job each product does for the buyer rather than by specification. Three colours, one product. One volume, one product. Someone whose job is “get a small amount quickly” does not care which of your six bottle sizes technically serves them best.

Once products are grouped by job, overlap becomes obvious in a way that spec-based catalogues hide. Two SKUs serving the same job at the same price with different packaging are duplicates, not choice. That single reframe usually finds more safe cuts than any margin analysis, because it catches the items that cannibalise rather than the items that merely underperform.

Build the substitution map alongside it: one row per product, with the job, the buyer who buys it, the survivors that meet the same job, and the gap if none does. If a product has no survivor, it is not a cut decision, it is a range gap you are about to open. That is a different conversation and sometimes the right answer is to keep the item.

4. Choose what to keep, merge, bundle or retire

Only now does the cut list get written. Five actions, not two, and the middle three are where most of the retained revenue comes from.

Score pattern	Action	Customer protection required
High contribution, low substitution risk	Keep, and give it better navigation and more content	None. This is the range’s spine
Low contribution, high strategic role (traffic driver, attach, key account)	Keep, and measure it on the role metric instead of revenue	None, but write down the role so the next review does not relitigate it
Low contribution, low substitution risk, clear overlap	Merge into the survivor as a variant, size or colour	Redirect the old URL and keep the variant discoverable
Moderate contribution, no clear overlap, still sells steadily	Convert to a bundle, accessory set or made-to-order offer	Explain the change before it happens; bundles often outperform the individual items
Low contribution, low turnover, no role, high substitution risk	Retire, with a dated sunset	Full treatment: named alternative, notice, migration path
Low volume but sole supplier of a customer need	Keep or move to made-to-order	Longer lead time is fine; losing the capability is not

Use the Pareto view as a starting point, never as the answer. The 80/20 principle says a small share of products produce most of the results, and in product ranges that pattern usually holds. The failure mode is treating the tail as worthless, because the tail frequently contains the niche items, the entry prices and the accessories that keep the rest of the range coherent. Forum threads on ABC analysis keep circling the same technical question, by the way: weight by units or by value? It changes which items land in which quartile, and there is no clean answer without knowing your own range.

And there are cases where you should not simplify at all. Regulated ranges with prescribed variants, contractual B2B obligations, safety-critical and warranty-bound lines, deliberate loss leaders, and ranges built around a small high-value segment with real switching costs all argue for restraint. Practitioners who admit to over-cutting, and I include myself, tend to have learned it in one of those five places.

5. Design clearer navigation and recommendations

A smaller range only earns its keep if customers can find their way through it. Group by job, name products for the use rather than the feature, cut filter combinations that return one result, and replace dead comparison tables with a good-better-best tier with a clear reason to step up.

This is also where a simplified range quietly lifts conversion, because the hesitation that choice overload creates has somewhere to go. The best-known evidence remains the jam study reported in the Harvard Business Review, where shoppers shown six varieties bought far more often than shoppers shown twenty-four, and the more recent finding that narrowing selection raised conversion by as much as 27% while cutting decision regret by 35% in health and intimacy categories. Both are category-specific, so treat the percentages as a direction rather than a forecast for your range.

6. Test the smaller range with real customers

Do not announce a cut list before something has tested it. Cheapest test: put a prototype shelf or range page in front of twelve to twenty customers and watch how long they take to choose and what they ask. If they cannot pick between two survivors, your labelling is doing its job badly, not your range.

Structured interviews work too, with a short list of removed products and one question: what would you have done instead? Answers that name nothing reveal a real need you were about to delete. Then pilot in one region, one channel or one market segment, and hold a clean baseline first so you are not comparing a good month to a bad one.

7. Roll out in phases and monitor the retention signals

Practitioners are near-universal on one point: phased, never a single mass cut. Most of the chatter in buying and supply-chain communities is about the sequencing, and the sequencing is what protects you.

Phase	Timing	Action	What to watch
Baseline	Weeks 1-2	Freeze the range, record metrics, confirm the substitution map with sales and support	Are sales, ops and finance scoring on the same criteria
First cut	Weeks 3-4	Retire only the unambiguous duplicates and cannibalising items	Conversion rate, support contacts on the affected pages
Restructure	Weeks 5-8	Merge variants, launch bundles and made-to-order offers, fix navigation	Average order value, attach rate, time on site
Full rollout	From week 9	Retire the remaining low-risk items with named alternatives for each	Repeat purchase rate, account retention, complaint volume
90-day review	Month 3	Compare against baseline on every measure you set in step 1	Revenue per SKU, inventory turns, net customer count
Quarterly rhythm	Ongoing	Repeat the classification; add back anything that now fails a role	Drift, since ranges always creep back

The metrics that matter after the change are simple: revenue per SKU, SKU productivity, conversion rate, repeat purchase rate, stockouts in the survivors, and complaint volume. Two of those catch mistakes the others miss. If revenue is flat, conversion is up and inventory turns improved, the cut is working. If operational metrics improved while repeat purchase fell, you cut something a real customer relied on, and it will cost more next year than the working capital you freed saved this year.

Give every retired product a migration path: a permanent redirect, a genuine alternative with a named reason, a dated sunset, and for contracts, a transition clause. Customers forgive a discontinued product. They do not forgive finding out from a third party.

Common Mistakes That Cost You Customers

• Cutting by sales alone. Revenue hides cost-to-serve and hides the role entirely. A high-revenue item with heavy changeover and returns can contribute less than a quiet item beside it.
• Removing choice before you understand the need. If you have not mapped what each buyer came for, you are guessing at substitutes. This is the mistake behind nearly every over-cut story.
• Changing the range and the messaging in the same quarter. When both move, nobody can tell which change caused the result and you cannot roll anything back.
• No substitute, no explanation. Deleting the item without naming what covers the need converts a neutral decision into a betrayal.
• One mass cut instead of phases. A single cut concentrates all the risk and leaves no room to learn.
• Cutting breadth before depth. Removing whole categories strands more customers than collapsing sizes or colours ever will.
• Letting the cut list rot. If nobody owns the decision to add a SKU back, customer-specific requests grow the range again within a year.
• Treating simplicity as a permanent state. A simplified range is also an easier range for competitors to copy. As one SaaS founder put it, simple products face fiercer competition because the barrier to entry is lower. Focus and simplicity are not the same thing as defensibility.

If you only remember the sequence: set the objective and the protected list, classify by role, score on margin, turnover, role and substitution risk, map every product to a customer job, pick one of five actions per item, protect the buyers of anything you retire, then phase and measure.

Frequently Asked Questions

Conclusion

Start with the substitution map, not the cut list. Before you decide anything, go through your at-risk products and write down the customer job each one does and which survivor meets that job. Everything else in this method follows from that page, and it will also tell you which products you were never really allowed to cut.

Score the range on contribution margin after cost-to-serve, sell-through, role and substitution risk. Pick one of five actions per item, phase the changes so nothing irreversible happens in month one, and give every retired product a named alternative and a dated sunset. Then watch repeat purchase as closely as you watch working capital, because that is the number that tells you whether you simplified the range or simply annoyed the customers who valued it.