THE BOARD DOESN’T CARE ABOUT YOUR BRAND ARCHITECTURE

It cares about concentration, risk, and return.

In late February, WPP announced WPP Creative — a new umbrella structure folding Ogilvy, VML, AKQA and Burson into a single division, targeting in the order of £500 million in cost savings over three years. Within forty-eight hours, every brand commentator with access to a microphone had taken a position. Some called it a Branded House. Some a House of Brands. Most a confused hybrid. Mark Ritson, with characteristic directness, called it strategically incoherent.

Everyone was arguing about where on a spectrum to plot the new structure.

Almost no one was arguing about the only question that mattered: does this new configuration of brands produce more compound value, at an acceptable level of risk, than the brand architecture it replaced?

Brand architecture used to be a typology question — what kind of architecture are we? Branded House or House of Brands? It now needs to be a portfolio question: what does this architecture earn, against what risk, compared with the alternatives? The instrument the CMOs take to their CEOs and boardrooms is not built to answer it.

That is a framework problem, not a WPP problem. It shows up every time a board is asked to make an architecture decision.

The framework that stopped evolving

The model most boards still use to reason about brand architecture — David Aaker and Erich Joachimsthaler's Brand Relationship Spectrum, produced in Brand Leadership in 2000 — is a serious piece of intellectual work. It gave businesses a shared language for choices that had until then been made by instinct or accumulated accident. It deserves the place it occupies in every postgraduate marketing curriculum or mini-MBA course.

It also deserves a more honest conversation about its limits.

To Aaker's credit, his own thinking moved beyond the spectrum. Brand Portfolio Strategy (2004) introduced the language of portfolio roles — silver bullets, cash cows, flankers, strategic brands — and acknowledged interaction effects more directly than the Branded House versus House of Brands spectrum had done.

The spectrum and its extensions were developed under a quiet but load-bearing assumption: that in the nineties, brands were still expensive to create, slow to build, and consequential to add. Adding a brand meant resources dedicated to multi-year identity programmes, agency rosters, retail roll-outs, considered media. The friction was real. The pre- and post-requirements of launching a brand forced executives into managing scarcity diligently.

My argument is that this assumption is now dead.

To create a brand today costs little more than a Canva subscription and forty minutes of attention. As a result, brands have proliferated. The old diligence is gone because boards no longer have meaningful approval over brand creation, since the cost has been engineered down to near-zero.

Portfolios are therefore expanding faster than the strategic thinking about them.

The consequence is that the spectrum now answers a question almost nobody is actually asking. Are we a Branded House or a House of Brands? is a typology question. The boardroom's question is a portfolio question: where is value leaking, where is risk concentrated, and what does the architecture produce as a number?

What the model cannot see

Three problems neither Aaker's traditional framework nor its extensions can, by their construction, fully address.

They evaluate brands one at a time. The spectrum tells you where each brand sits on a continuum from masterbrand to standalone. Portfolio roles tell you what each brand is for. Neither tells you what happens when you put two brands beside each other in the same portfolio. The interaction effects — the part of the system that determines whether the portfolio compounds value or suppresses it — are not the object of analysis.

They have no quantified concept of risk. Plotting a brand on a typology gives you a label. It does not tell you whether the brand makes the portfolio more or less vulnerable to a single shock — a regulatory event, a reputational crisis, a category collapse. In every other asset class a CFO touches, risk is the first variable. In brand architecture as conventionally practised, it is at best a qualitative consideration.

They do not produce a number. The output of an Aaker analysis is a structure. The output a board needs is a financial case.

There is empirical work that begins to address the gap. Rao (2004) and Srinivasan, Fournier and Hsu (2015) established that architecture choice has measurable consequences for stock returns, and that sub-branding outperforms branded house but at higher risk. That work matters. It is, by design, a comparison of strategies in aggregate — not an instrument for diagnosing a specific portfolio in front of a specific board with a specific decision to make.

The case the typology cannot explain

Compare two luxury portfolios.

By every conventional reading, Kering and LVMH are both Houses of Brands. Each owns a stable of distinct maisons with their own creative direction, identity, and customer. The typology says they are the same kind of architecture. They are not the same kind of business.

Kering's flagship Gucci accounts for around 40 per cent of group revenue and roughly 60 per cent of operating profit. The portfolio is overwhelmingly concentrated in luxury fashion and leather goods, at adjacent price points, exposed to the same Chinese aspirational-luxury demand cycle, the same wholesale dependency, the same creative-direction risk. When Gucci faltered, no other brand in the system had the equity transfer capacity, audience reach, or category breadth to absorb the impact. Kering's share price fell roughly 80 per cent from its 2021 peak, prompting the ReconKering turnaround announced by new CEO Luca de Meo last month.

LVMH is, by typology, the same. By portfolio dynamics, it is structurally different. Its brands span wines and spirits, fashion and leather goods, perfumes and cosmetics, watches and jewellery, selective retail, and hospitality. Champagne demand cycles do not correlate with handbag cycles do not correlate with travel-retail cycles. When one category falters, the others continue. The portfolio is genuinely diversified in the sense that matters — its brands are exposed to substantively uncorrelated risks. LVMH is the most valuable luxury company in the world, with a market capitalisation many times Kering's.

Two Houses of Brands, indistinguishable to the spectrum. Two portfolio profiles, materially different. One outcome predictable from the architecture; one a creative crisis the architecture amplified rather than absorbed.

This is what the typology cannot see — and what a portfolio instrument can.

A different instrument

At Strategic Brand Partners we have built an instrument designed to answer the portfolio question — the Brand Valency Model™. It measures how the brands in a portfolio interact across five dimensions, plots them against a sector-calibrated risk-return benchmark drawn from Modern Portfolio Theory, and produces a single risk-adjusted score that boards can defend in the language their CFO already speaks. The mechanics sit inside the model. What sits in the boardroom is a number.

The model sits comfortably with the most empirically validated brand research of the last quarter-century — Sharp on penetration and mental availability, Binet and Field on brand investment reducing revenue volatility, Rao on architecture's measurable effect on stock returns. It is in that tradition, not against it. What it adds is a portfolio-level diagnostic that converts those established findings into a single architecture decision a board can take.

The model is not a substitute for judgement. It produces a defensible, quantified case. It does not absolve a board of choosing. Where a lower-scoring scenario is the right answer for reasons of regulation, governance, succession, or geography, the model will say so — but the call remains a human one.

And not every portfolio problem is an architecture problem. Some of what looks like an architecture issue is in fact a creative direction issue, a leadership issue, a product issue, or a regulatory issue. The model includes a prior diagnostic — is architecture actually the lever here? — and when the answer is no, the most useful thing it produces is a recommendation not to restructure. The framework that recommends restructuring in every case is selling a service, not solving a problem.

The point

For twenty-five years, brand architecture has been governed by a framework built for a world in which adding a brand was a costly, considered act. That world is gone. Portfolios now expand faster than the thinking about them, and the question the boardroom asks has changed.

It is no longer what kind of architecture are we? It is: what does this architecture earn, against what risk, compared with the alternatives?

Whether our brands cumulatively add or destroy value is the question now. Where any single brand sits on a twenty-five-year-old spectrum is, with respect, a footnote.