Nobody asks for more line changeovers. Or more planning noise. Or more supplier sprawl.
They ask for growth. Then the tail arrives.
At some point, what looked like commercial expansion starts behaving like operational drag.
That is the better phrase. “SKU complexity” sounds tidy. Operational drag is what the team actually feels.
The problem is not that the portfolio got larger
Some complexity is justified. Different markets need different formats. Different channels need different commercial configurations. Some variants are strategically important. Some local requirements are real.
So this is not an argument for a tiny portfolio or some simplistic anti-SKU ideology.
The real issue is whether the additional assortment is creating enough value to justify the operating burden it leaves behind.
That is the part companies usually under-measure.
Revenue is easy to see.
The drag usually is not.
Too many companies still treat this as a portfolio reporting issue instead of an operating problem.
The drag shows up where operators live
One reason complexity gets underplayed is that the language around it is too abstract.
On the ground, the problem is concrete.
It shows up as friction in the places operating teams actually live.
On the factory floor
More variants tend to mean more setups, shorter campaigns, and less freedom in the schedule.
That usually means lower efficiency, less room to absorb disruption, and more time spent accommodating the tail.
Firms like Bain and BCG have been making this point for years. The tax is operational before it is conceptual.
In planning
Long tails create more noise.
More low-volume items. More exceptions. More judgement calls. More debate over what deserves capacity and what does not.
At some point, planning stops steering the business and starts nursing complexity the portfolio created for itself.
In sourcing and operational support
A variant rarely arrives alone.
It usually brings extra inputs, components, approvals, and supplier interactions with it.
So what looked like a reasonable commercial choice upstream starts to show up downstream as more supplier sprawl, more qualification work, and more coordination overhead.
In inventory and working capital
As the tail expands, inventory tends to get thinner and more fragmented. You get more stock in awkward pockets, more edge cases, and more working capital trapped in items that do not earn the attention they consume.
So this should not be treated as a simple portfolio tidy-up. It is an operating-model and margin problem.
Why this keeps getting underestimated
There is a structural reason the issue stays hidden for too long.
No one function holds the whole bill.
Manufacturing feels the setups.
Planning feels the exceptions.
Procurement feels the sprawl.
Warehousing feels the fragmentation.
Finance sees margin pressure but not always the mechanics behind it.
Because the pain is distributed, the problem gets undercounted. Then the portfolio stops being managed and starts managing the business.
There is also a language problem.
“SKU intelligence” sounds like a feature category.
Operators usually describe the same reality much more plainly:
- too many low-value variants
- too much tail
- too many setups
- too much noise in planning
- too many materials and suppliers to manage cleanly
That language is less polished and much closer to the truth.
Stop counting SKUs. Start weighting complexity.
One of the quickest ways to get this wrong is to reduce it to a count.
A raw SKU number tells you very little. Two portfolios can have the same number of SKUs and totally different operating profiles.
The better question is not:
How many SKUs do we have?
It is:
Which variants create a disproportionate amount of drag relative to the value they contribute?
That is the management question.
Because some complexity is worth carrying. Some is not. Some variants are strategically necessary in one market and completely unjustified in another. Some products could share far more supply chain logic than they currently do.
Until you look at complexity in weighted terms, contribution, volume, setup burden, sourcing burden, duplication across markets, and service criticality, most simplification efforts stay too blunt.
They reduce visible count instead of removing meaningful drag.
Simplification is not pruning
The weakest version of this work feels like annual gardening.
Trim the tail. Delete some SKUs. Declare success.
That may remove noise, but it often misses the bigger opportunity. If you only cut count, the drag grows back.
Real complexity reduction is redesign, not subtraction.
It means asking better questions:
- Where are we carrying near-duplicate variants across markets?
- Which inputs or operating standards could be shared more broadly?
- Which low-volume variants create outsized changeover pain?
- Which commercial choices create operating burden nobody has really priced in?
- Where are we preserving variety out of habit rather than strategy?
That makes complexity reduction the stronger public framing.
It names what operating leaders actually care about: less drag, cleaner decisions, better throughput, and a portfolio the business can still run well.
Why this becomes urgent when the environment gets harder
When growth is easier, companies can absorb more mess.
When margins tighten, that changes fast.
Leaders are suddenly being asked to improve margins, simplify operations, free up capacity, and move faster without taking on service risk.
In that environment, the long tail stops looking like a background annoyance and starts looking like a strategic constraint.
The portfolio is no longer just a commercial artifact.
It is one of the biggest determinants of how hard the operation is to run.
Where systems still struggle
Most companies can produce a list of SKUs. Far fewer can join up the signals that actually matter:
- contribution and volume
- setup and schedule burden
- component variation
- supplier complexity
- market-by-market duplication
- service criticality
Simplification work still ends up in side analyses, spreadsheets, and workshops. The system of record can usually tell you what exists. It usually cannot tell you what is creating disproportionate operating pain.
That is the gap.
Where AI is useful, and where it is not
The weak AI story is: the model will tell you which SKUs to cut.
Nobody serious should buy that.
The stronger story is more practical. AI can help teams do the comparison work that is slow and painful at portfolio scale:
- normalize inconsistent product and material descriptions
- cluster similar variants across markets
- surface likely duplication or standardization candidates
- summarize where the drag appears to be concentrated
- draft a first-pass review pack that humans can challenge
That matters because complexity reduction is not a black-box optimization problem. It is a judgment problem with real commercial and operational tradeoffs. The job of the system is to make those tradeoffs easier to see, not pretend they do not exist.
Bad AI gives you a ranking. Good AI gives you a case a human can actually review.
The signal I would watch for
I would not wait until the portfolio “looks too big.”
That is usually too late and too vague.
I would watch for the moment the portfolio starts making the business slower to plan, slower to produce, harder to source, and harder to change.
That is the real threshold.
That is when assortment growth becomes operational drag.
The tail never announces itself as a strategic problem. It just makes the business slower until everyone can feel it.