Inventory optimization: consulting engagement or decision platform?

A supply chain VP and a CFO walk into a steering committee. Inventory keeps slipping — markdowns deepening in fashion, overstock rising in DIY, ruptures climbing in beauty. The question on the agenda is "do we bring in a consulting firm, or do we buy a decision platform?" Both sides have their answer ready, and both are partly right.

The honest version is that the two options solve genuinely different problems, and the choice that survives the next 24 months depends on which problem you actually have. This article unpacks the distinction in finance terms, gives you the five questions to ask before committing, and stays explicit about the hybrid pattern most mid-to-large retailers end up running.

The two genuinely different problems

The first move in any honest consulting-vs-platform conversation is to refuse the false equivalence. A tier-one strategy firm and a decision platform are not two flavors of the same thing. They are two interventions sized for two different bottlenecks. Pick the wrong one for your bottleneck and you spend a year discovering the mismatch — usually after a seven-figure check has cleared.

Consulting wins when the bottleneck is political

There are situations where the hardest part of fixing inventory is not the math. It is getting buyers, merchandisers, supply chain, finance, and store operations to agree on a new operating model — and to live with the org-chart changes that follow. The bottleneck is not "what is the right replenishment rule?" It is "who owns it once we change it, and how do we tell the four people who used to?"

That is consulting work. A tier-one firm earns its fee in those rooms. The diagnostic that names the dysfunction, the change-management sequencing, the political cover of an outside signature on a recommendation internal leaders couldn't deliver. One-off restructures, post-merger integrations, organisational redesigns, regulatory transformations: these are jobs the right consulting engagement does well, and a platform doesn't do at all.

The test is whether the work has a defined end state. If the question is "how do we reorganize the inventory function across three newly merged brands?" — that has an end state. Consulting fits. If the question is "how do we take 50,000 markdown decisions a season consistently?" — that has no end state, only a steady-state flow. Consulting is the wrong shape.

Platform wins when the bottleneck is cognitive bandwidth

Most inventory pain in a mid-to-large retailer is not political. It is a volume problem dressed up as a quality problem. The team makes the right call when it has time to think. It runs out of time around SKU 200 of 40,000, and from there decisions get made by template, by gut, or not at all. The cost of that compromise compounds week over week into the markdowns, the carrying cost, and the lost sales every CFO eventually has to explain.

You cannot consult your way out of a bandwidth problem. No 200-page deck on better replenishment calls changes the fact that no human team can make 40,000 of them a week to standard. What changes the math is a decision platform. It takes the repeated decisions at scale, embeds the rules the team defends, and propagates each call into the systems that execute it.

This is the structural reason platforms exist in this category. They are not a better consultant; they are the only intervention that scales linearly with decision volume. The work the platform does next year is the same work it did last year, on more SKUs and more stores, at near-zero marginal cost. That is the opposite of a consulting engagement, and it is the right shape when the bottleneck is volume.

The cost economics: opex vs capex, and what each buys you

CFOs care less about the philosophical distinction than about how each option lands in the budget. The two have structurally different financial signatures, and pretending they're comparable on a single ROI line is exactly how people pick wrong.

Consulting is pure opex. A six-to-twelve month inventory optimization engagement at a mid-size chain typically lands in the low-to-mid six-figure range, sometimes higher with data work or process design. Cost scales linearly with scope: more categories, more countries, more weeks. It ends when the engagement ends — there is no annuity. The IP usually leaves with the consultants, sitting in slides on a laptop that gets reorganized away inside two years.

A platform is a mix of capex and opex. A decision platform for the same chain typically lands in the low six to mid seven figures all-in over three years, depending on scope, integrations, and network size. The cost scales non-linearly with decision volume: one license decision powers many SKU/store decisions, year over year. The IP is persistent — the business rules, the integrations, the learning loop accumulate. The trade is real lock-in risk, which is the one honest concern that belongs in any platform business case.

Notice what's not in either bullet: a comparable cost-of-quality. A consulting engagement that produces a deck no one operationalises has cost you its fee and bought you nothing. A platform that ships in nine months without a closed decision-execution loop has cost you its first-year fee and bought you a recommender. The real number to model is not the sticker price; it's the total cost of industrialization measured in margin lift actually banked over three years.

For the frame that anchors these conversations in P&L terms, see the CFO's guide to retail AI ROI — which P&L line moves, by how much, and when.

The hybrid pattern most retailers actually want

In practice, the binary framing of "consult or buy" is almost always wrong. The pattern that ships value in most mid-to-large retailers is a sequence, not a choice — and it uses each intervention for the problem it actually solves.

The sequence runs in three beats. First, a short consulting engagement to scope and select. Two to three months, not twelve. The output is a sober read on the bandwidth bottleneck, a shortlist of platforms whose architecture fits the stack, and an honest baseline against which value will be measured. Consulting at its best: diagnostic, decisive, with a defined end state.

Second, platform-led for the ongoing operational decisions. Replenishment, allocation, markdown timing, store transfers, returns disposition — the decisions that recur every week and that no team can hand-arbitrate to standard. The platform owns the steady state; the internal team owns the rules, the overrides, and the strategic shifts. Decisions move from a meeting agenda to a managed flow, with the loop closed back to outcomes.

Third, occasional consulting for major restructures. A merger, a vertical expansion, a new geography, a regulatory shift — anything that breaks the operating model — pulls a firm back in for the one-off political work. The platform keeps running underneath, because daily decisions don't care that the org chart changed.

The economics of this sequence are usually better than either pure option. The consulting bill is smaller and time-boxed, because the platform absorbs the recurring work. The platform pays back faster, because the consulting engagement scoped it honestly and built executive air cover. And the IP stays — in the platform, in the rules, in the team that runs it — instead of leaving with whoever's currently on the engagement.

Five questions the CFO should actually ask before committing

Every consulting-vs-platform conversation eventually narrows to a binary. Before it does, run it through these five questions. Each isolates a variable that decides the answer.

1. Is the bottleneck a repeated decision or a one-time restructure? If the same call has to be made 10,000 times a season, that's platform shape. If a single defining call has to be made once and lived with, that's consulting shape. Most networks have both — but the dominant cost lives in only one.

2. Will the IP and the model leave when the engagement ends? This is the question consultants don't volunteer. Where does the optimization logic live after month twelve — in a system the team operates daily, or in slides nobody reopens? If the answer is slides, the cost compounds against you.

3. Can the org operate the new way without the consultants in the room? Run the thought experiment of the engagement ending tomorrow. Does the better way of working continue, or does it quietly snap back? Real transformations survive the engagement; theatrical ones don't. The honest version of this question is whether you're buying a capability or renting a feeling.

4. What is the marginal cost of one more decision next year? Consulting is linear: another category, another quarter, another fee. A platform is sub-linear: the cost of decision 50,001 is essentially zero once decision 1 is in production. Which curve your business needs depends on how the decision volume is moving — and in most retailers it is moving in only one direction.

5. Who owns the decision logic in three years? This is the question that should anchor the whole conversation. If the answer is "the platform we operate, the rules our team maintains, the loop our system closes" — that's a defensible asset. If the answer is "we'd have to bring the firm back to figure it out" — that's a recurring tax, dressed as flexibility.

The classic anti-pattern: the deck on the shelf

The failure mode worth naming is the one most CFOs have already seen at least once. A multi-million-euro engagement scopes inventory optimization across the network, runs nine to twelve months, and lands a 200-page deck — categorised, prioritised, footnoted, executive-summarised. It is genuinely good work.

Six months later the deck is on a shelf. The recommendations are partly implemented in spreadsheets that don't talk to each other. Inventory KPIs look roughly as they did the day the engagement started.

The failure is not in the analysis. It is in the structural mismatch: a deck is a one-time artifact, while the decisions it recommends are a continuous flow. Without an operational system to take the daily call and propagate it into the ERP, WMS, and pricing engine, the recommendations decay the moment the network moves. And a retail network moves weekly. This is the same pattern that turns retail data useless without a decision layer: insight that cannot be operationalised at the cadence of the business converts to nothing.

The right read of this failure is not "consulting was wrong." It is "consulting was the wrong shape for the underlying problem." The deck would have worked for the political-restructure problem it implicitly solved. It failed at the cognitive-bandwidth problem it was sold against — because no deck can solve a bandwidth problem.

Solya in context

Solya is a decision platform, with lightweight services around scoping and implementation. We are not a consultancy, and we do not pretend consulting engagements have no place. We say the opposite. There are situations — post-merger integrations, regulatory transformations, deep org redesigns — where the right move is a tier-one engagement and a platform is the wrong tool. We say so on those calls.

What we are built for is the steady state. The repeated SKU/store decisions, the markdown timing, the allocation arbitration, the store transfers — work that recurs every week and that no team can hand-arbitrate to standard. We are built so you do not need a consulting engagement to run inventory decisions — only to handle the occasional structural change that breaks the operating model. For the architecture that makes this possible, see the orchestration layer.

The question to settle before the next steering committee

The consulting-vs-platform debate is sharper than most committees admit, and softer than either side wants to claim. Sharper, because the two options solve genuinely different problems and picking wrong costs a year and a budget. Softer, because the right answer for most mid-to-large retailers is a sequence, not a binary — short consulting to scope, platform for the ongoing flow, occasional consulting for restructures.

Settle one question: what fraction of your inventory pain comes from decisions made once, and what fraction from decisions made ten thousand times? That ratio is the variable that decides the answer. Every other discussion is downstream of it.

---

Pressure-test your consulting-vs-platform case

At Solya, we offer supply chain and finance leaders a 30-minute diagnostic to map your inventory bottleneck against the consulting-vs-platform decision. Vendor-neutral, grounded in your network, your category mix, and your decision volumes.

You'll walk away with:

• A read on whether your dominant bottleneck is political (consulting-shape) or bandwidth (platform-shape)
• A three-year cost-and-IP comparison framed against your own decision volumes
• A view of where a hybrid sequence — short consulting plus platform — would fit your roadmap