Product Portfolio Design

Why this node matters

Product portfolio design is the node where commercial breadth either compounds margin or accumulates complexity cost. The selection determines which SKUs the portco invests production capacity against, which products carry the gross margin contribution the EBITDA bridge depends on, and which complexity the operational organisation must absorb. Offer creation, pricing, channel design, and sales motion all inherit their economic logic from this layer.

Most Mittelstand portcos manage the portfolio on accumulated history rather than empirical criteria. Products were added when customers asked, retained because someone still buys them, and reviewed only when manufacturing complexity or inventory cost becomes operationally painful. These are the variables that surface through complaint rather than analysis. They predict portfolio profitability weakly because they ignore activity-based costing, cannibalisation, and strategic role. The economically meaningful portfolio criteria are contribution per SKU, strategic role, customer-job coverage, and operational capacity fit, not what the firm happens to currently produce.

The unexamined assumption underneath in PE operating practice is that the existing portfolio represents accumulated customer demand and therefore reflects market reality. The portfolio is an accumulation of historical decisions, most of which were made without portfolio-level discipline.

The empirical literature is unambiguous. Cooper, Edgett, and Kleinschmidt (1999), in their Journal of Product Innovation Management study of 205 U.S. firms, established four canonical objectives for new product portfolio management (maximise value, achieve balance, align with strategy, fit capacity) and found firms using pure financial methods systematically underperform firms using multi-criteria methods. The objectives generalise to existing portfolio rationalisation, where the same multi-criteria logic applies. Strategy& (2019), analysing CPG portfolios, found 35 percent of SKUs drive zero incremental profitability and the bottom 15 percent actively destroy gross margin. The 20/225 distribution holds at the product level as well as the customer level.

The node operates across four selection layers. Each carries a different decision weight.

Economic criteria carry primary decision weight. Contribution margin per SKU under activity-based costing, inventory turnover, cannibalisation rate, and customer concentration per SKU. Kaplan and Cooper (1998) established the 20/225 pattern at the product level: top 20 percent of SKUs generate 150 to 300 percent of profits, bottom 10 to 20 percent destroy 50 to 200 percent. Strategy& (2019) confirmed the pattern empirically in contemporary CPG: 35 percent of SKUs at zero profitability, 15 percent destructive. Standard gross margin hides this because it allocates overhead by revenue rather than activity. The mechanism is direct. Economic criteria reveal which products are subsidised by which other products in the portfolio.

Strategic criteria carry predictive weight. Strategic role (Henderson's 1970 BCG classification: Stars, Cash Cows, Question Marks, Dogs), lifecycle stage distribution, innovation pipeline coverage, and strategic fit with capability. Henderson's framework remains operationally useful despite acknowledged limitations in fast-changing markets (BCG Henderson Institute, 2014); the underlying logic that a portfolio needs balance of cash generators, growth investments, and selective bets holds. Cooper, Edgett, and Kleinschmidt (2001) identified pipeline gaps as the single most common portfolio failure mode in their study of 205 firms. The mechanism: portfolios concentrated in any one strategic role produce systematic risk at the lifecycle transition.

Customer criteria carry filtering weight. Customer-job coverage, cross-sell and bundle architecture, and product line breadth tested against contribution. Kekre and Srinivasan (1990), in Management Science, found broader product lines produce significant market share benefits across 1,400 business units, but Quelch and Kenny (1994), in their seminal Harvard Business Review analysis "Extend Profits, Not Product Lines," refined this into the curvilinear relationship: beyond a category-specific threshold, additional SKUs produce diminishing returns and rising complexity costs that exceed the revenue benefit. Christensen, Hall, Dillon, and Duncan (2016) frame the customer-coverage question as job-to-be-done scope. The mechanism: portfolios that cover the customer's full job retain at 2 to 3x the rate of single-product portfolios, but only up to the breadth threshold where complexity begins to dominate.

Operational criteria test executability. Manufacturing complexity, inventory and working capital intensity, channel complexity, and sales force coverage capacity. Fisher and Ittner (1999), in Management Science, established that manufacturing efficiency declines roughly proportionally to SKU count above the optimal level. The empirical sales force constraint: reps cannot effectively sell more than 7 to 12 products in active rotation, which means portfolios with 50+ SKUs systematically produce coverage gaps where 70 percent of the portfolio receives 20 percent of sales effort. The mechanism is the same as the segmentation literature: complexity costs above a threshold dominate the revenue benefits the complexity was designed to capture.

The output of the node is a core portfolio of SKUs that pass all four layers, a redesign candidate list of SKUs that pass economic and strategic layers but require customer-coverage or operational redesign, and a discontinuation or divestiture list of SKUs that fail the foundational tests. The selection is a precondition for every downstream node. Pricing inherits its margin structure from this layer. Sales motion inherits its product mix. Channel design inherits its complexity envelope.

The thesis: a product portfolio that has not been rationalised against contribution, strategic role, customer coverage, and operational capacity is not a portfolio. It is an accumulation.