Your commercial director brings good news: "Premium Customer grew 40% this year. It's our most profitable customer at 28% aggregated margin." You approve more commercial resources for Premium Customer. Six months later, EBITDA drops. Why? Premium Customer IS profitable in aggregate — but the 40% incremental you sold was Product C through Distributor channel: −5% margin per intersection. You sold more to a "profitable" customer but through the most loss-making cell of your matrix.
This happens because aggregated profitability deceives in companies with multidimensional mix. Typical FP&A reports say "profitability by customer" or "profitability by product" — metrics mathematically correct but strategically incomplete. The real lens is the INTERSECTION matrix: customer × product × channel. Where each cell reveals whether that specific combination generates or destroys value.
In this module you learn to build and read the multidimensional profitability matrix. When to apply it, what decisions to make on each cell (renegotiate price, redirect channel, withdraw product from customer), and why the discipline of "cell-by-cell surgery" is very different from "discontinue the customer" or "kill the product."
Let's go.
Customer A generates $500K/year at 22% aggregated margin. Bought by intersection: $400K in Product X at 28% margin, $100K in Product Y at −2% margin. Your commercial team proposes raising Customer A volume 30%. Good plan?
In plain language
Before the mechanics, the four basic questions.
Why do this at all?
Because aggregated profitability by customer, product, or channel is mathematically correct but strategically misleading. Mix, pricing, expansion, and discontinuation decisions are made on the wrong dimension when only aggregate is seen. The intersection matrix reveals the 5-10 cells that determine EBITDA destiny — where 1-2 are big winners, 1-2 are big hidden losers, and cell-by-cell surgery improves EBITDA 3-8pp in 12-18 months without changing the apparent portfolio.
Who builds it and who uses it?
Senior FP&A builds it, ideally with IT/data support to extract revenue and cost per intersection. Commercial provides context for why each cell exists. Commercial director + CFO + CEO USE it for pricing, mix, expansion, and discontinuation decisions. Without commercial team in the conversation, the matrix becomes analysis without action — because actions (renegotiate price, redirect channel) are executed by commercial.
When does it show up?
Quarterly in executive mix review. Annually with portfolio planning. Before every big expansion decision (entering new customer segment, launching new product, changing channel mix). In M&A due diligence, the buyer always asks for the matrix by cohort — to understand mix quality, not just aggregate. And critically: when aggregated profitability stays flat while mix is changing — that's the signal that cells are moving invisibly to the aggregate.
What if we never build it?
Three predictable patterns: (1) The company pushes volume in "profitable" customers without noticing incremental goes to loss-making cells. EBITDA falls while revenue grows. (2) Discontinues "marginal" products that are actually profitable in certain cells (e.g.: Product C is good only for Premium Customer direct) — destroys real options. (3) Subsidizes loss-making customers without knowing. The "customer-centric" company that swears to defend the difficult customer discovers late that customer never paid the real cost of serving it.
Andina S.A. — the matrix that changed strategy
Andina reports profitability by customer and by product every month. The executive committee approves strategies on those aggregates. EBITDA was flat two years despite 18% revenue growth. Something didn't add up.
CFO commissions multidimensional matrix: 4 customer segments × 4 product lines × 4 channels = 64 cells. Analysis reveals 8 cells are highly profitable (35%+), 12 cells healthy (20-35%), 30 cells in watch zone (5-20%), and 14 cells have NEGATIVE margin that aggregate hid.
One critical cell: Premium Customer × Line C × Distributor channel = −5% margin. Total revenue of that cell: $2.3M/year. Means Andina loses $115K/year selling a product to a premium customer through a wrong channel. But aggregated report showed "Premium Customer = 28% margin" because the customer's other purchases compensated.
Cell-by-cell surgery: in 6 months, Andina (a) renegotiated Line C price for Premium Customer raising 15% (customer accepted because they recognized value); (b) redirected that cell to Direct Sales channel (better margin); (c) discontinued the Mass × Services cell (−8% margin, $20K revenue — wasn't worth maintaining). Result at 12 months: +2.8pp of total EBITDA. Without changing products, without losing customers, without changing apparent mix. Just surgery on 6 cells.
The board conversation changes. No more "sell more to Premium Customer" as blanket strategy — but "sell MORE OF Product A and B to Premium Customer, sell LESS OF Product C to Premium Customer, and never through Distributor." Strategy becomes cell-specific.
The visual below lets you see a simplified Andina matrix and toggle between Customer × Product and Customer × Channel views.
Matrix, live
Andina 4×4 matrix with two views. Customer × Product and Customer × Channel. Each cell shows margin % and revenue $K. Green = healthy, amber = watch, red = negative.
The critical experiment: alternate between the two views. Look at the "Premium" row — in product view, Line C has −3% margin and $30K revenue (loser at that intersection). In channel view, Premium × Partner has $0 revenue (customer doesn't buy through that channel). The matrix reveals WHERE the losses are, not just HOW MUCH.
Interactive visual
Multi-dimensional profitability · customer × product × channel
Aggregated profitability hides per-intersection losses. A 'profitable' customer can be buying our least-profitable product through our most expensive channel = net loss. The heatmap reveals exactly where to look.
What you are seeing
Three critical lessons: (1) Aggregated profitability by customer or by product hides loss-making intersections. Premium Customer can be 28% margin in aggregate but buy Product C through Distributor channel with −5% margin. Without the intersection lens, the deal of 'sell more to Premium Customer' looks good and turns out destructive. (2) The action is NOT to discontinue the customer or the product — it's to renegotiate the INTERSECTION. Raise Product C price when sold to Premium customer, or redirect the channel, or renegotiate terms. Surgery is by cell, not by aggregate. (3) Without multidimensional matrix, FP&A reports 'profitability by customer' that is mathematically correct but strategically misleading. The real discipline is the intersection heatmap — once per quarter — to identify the 5-10 cells defining EBITDA destiny.
The critical reading of the visual: every row or column aggregate hides individual cells that can be very different. Premium Customer aggregated is 28% — but contains a −3% cell. Line C aggregated may be 15% — but contains cells from +25% to −5%. Without the matrix, decisions are made on average; with the matrix, on specific cells.
Second reading: the action is NOT "discontinue the customer" or "kill the product." Premium Customer in aggregate is valuable. Line C in aggregate may be strategic. The action is CELL-BY-CELL SURGERY: renegotiate Line C price when bought by Premium Customer, or redirect that cell to another channel, or offer Line C in the bundle with Line A where combined profitability does justify. The company becomes surgical, not butcher.
And critical: the matrix is built once per quarter, not continuously. Effort is 1-2 days of FP&A per quarter. Output informs 5-10 surgery decisions. Each cell improved from −5% to +15% over $500K revenue = +$100K annual contribution. Doing 5 surgeries per year = $500K-$1M of EBITDA recovered, without touching apparent portfolio. It's the highest-leverage discipline in mature FP&A.
The mechanics: how to build the multidimensional matrix
- Start with two dimensions, not three. Customer × Product is most useful to start. Once the team internalizes the reading, add Channel as third dimension. Launching all three simultaneously overloads the team and reading becomes confusing.
- Define cells at SEGMENT level, not individual customer. 4 segments × 4 products × 4 channels = 64 analyzable cells. 200 customers × 50 SKUs × 5 channels = 50,000 cells nobody can manage. Granularity is the enemy here.
- Per-cell margin must use ABC costing (Module 2.10), not traditional. Without ABC, the matrix inherits cross-subsidies from traditional costing and shows "profitable" cells that actually lose money. Multidimensional matrix without ABC is mathematically suspect.
- Document revenue AND margin per cell — both matter. A cell with +35% margin but $5K revenue doesn't deserve strategic attention. A cell with −2% margin but $2M revenue is immediate priority. Matrix must allow filtering by "materially important cells" — typically >$200K revenue.
- Identify the 5-10 cells that define EBITDA destiny. Pareto applies here too: few cells determine most of value (positive or negative). Focus on those few. The other 50+ cells are second order.
- For each problematic cell, evaluate 3 options: renegotiate price, redirect channel, withdraw product from customer. NOT the fourth option: "discontinue the customer" or "kill the product" — those are aggregate actions that throw the baby out with the bathwater.
- Refresh quarterly. Compare quarter to quarter the SAME cells. Per-cell trajectory reveals trends before they appear in aggregate. A cell moving from +15% to +5% over 3 quarters is early signal — act before it reaches 0% or negative.
- Aggregated profitability. Aggregated margin by customer, product, or channel. Mathematically correct, strategically partial.
- Intersection profitability. Margin per specific cell (customer × product × channel). Where cross-subsidies and real opportunities are seen.
- Cost-to-serve (module 2.13). Real cost of serving a specific customer or channel. Includes pre-sale, sales, post-sale, returns, support. Essential for intersection margin to be defendable.
- Cross-subsidy. When a profitable cell subsidizes a loss-making cell within the same customer or product. Matrix reveals it; aggregate hides it.
- Cell-by-cell surgery. The discipline of acting per specific cell (renegotiate price, redirect channel) instead of per aggregate (discontinue customer).
- Material cell. Cell with sufficient revenue to justify strategic attention. Typically >0.5% of total revenue — the rest is noise.
- Practical rule: cell-by-cell surgery in mid-market improves EBITDA 3-8pp in 12-18 months without changing apparent portfolio. It's the most underutilized FP&A lever.
Adversarial check
Adversarial check
1.Your commercial director says "Premium Customer grew 40% this year, our best customer." Proposes more resources for Premium Customer. What do you ask BEFORE approving?
2.Your analysis reveals Mass Customer × Services = −8% margin. What is the right answer?
3.Why does the multidimensional matrix need ABC costing instead of traditional?
Exit checklist
Suggested re-review: quarterly with executive mix review. Annually with portfolio planning. And always before approving big expansion, launch, or discontinuation decisions — those are the decisions where the matrix changes the outcome.
Optional
Go deeper
Sources and books to dig into the original material