Insights
Brand Architecture Is a Commercial Decision, Not a Design One
Four structures, three tests, and what the choice is actually worth
Ask a leadership team to draw the relationship between their brands and you will often get a long pause, followed by a diagram three people immediately disagree with.
This is not a design problem. It is a valuation problem wearing a design costume. Architecture decides what a business can extend into, what an acquisition is worth once you own it, and how much damage one bad year in one division does to everything else sharing the name above the door.
The four structures
Most organisations are in one of these without ever having chosen it.
| Structure | How it works | Recognisable examples |
|---|---|---|
| Monolithic | One master brand, one visual system. Extensions built with descriptors. The customer is trusting the name. | Mercedes-Benz, where every model is a descriptor beneath one name. Commonwealth Bank across its retail network. |
| Endorsed | The product has its own name and presence, but the parent vouches for it. Synergy without full dependence. | Courtyard by Marriott, AC Hotels by Marriott. Each has its own identity; the parent signature carries the guarantee. |
| Branded | Corporate identity kept separate from the brands. Brands have their own lifecycles and sometimes compete deliberately. | Procter & Gamble (Tide, Pampers, Gillette). Wesfarmers (Bunnings, Kmart, Officeworks). Most shoppers cannot name the owner of any of them. |
| Hybrid | A combination of all three, usually arrived at through acquisition. | Volkswagen Group: Volkswagen, Audi, Porsche and Skoda, each positioned separately, largely arrived at by buying them. |
The three tests
Choosing between them is not a matter of taste. Three questions decide it, and they should be answered in this order.
1. Synergy
Is the brand promise genuinely interlinked across this portfolio, or are you asserting a connection that customers never experience? A shared owner is not a shared promise. If the only thing two businesses have in common is a shareholder, the customer has no reason to transfer trust between them.
2. Equity
At which level does the value actually sit? Sometimes the parent. Often – particularly after an acquisition – it sits in the thing you bought, and absorbing it into the master brand destroys the asset you paid for. This is the test most frequently skipped, and the most expensive to get wrong.
3. Credibility
Will the customer believe this parent in this category? A name carrying real authority in industrial products may carry nothing in consumer goods. Asserting it anyway costs more than starting fresh, because you spend twice – once on the assertion and again on the correction.
Two ways it goes wrong
They are opposites, and most organisations are drifting toward one of them.
- The monolithic default. One name is cheaper to build and easier to manage, so it gets chosen without the tests being run. It works well while things go well – sub-brands ride the parent’s strength, visibility costs less, cross-selling is easier. The exposure only shows up on a bad day, when a single failure in one division reaches everything sharing the name.
- Extension until nothing is meant. The master brand stretched across so many unrelated things that it stops standing for anything in particular. Virgin is the most-cited case: airlines, trains, gyms, telecoms, spaceflight. Ask what the name promises and the honest answer is a tone of voice rather than a category. Not damaged by a crisis – diluted into wallpaper by trying to be everything to everyone.
The diversified holding company faces both at once
Where one ownership spans genuinely unrelated sectors, the honest architecture is almost always layered.
A group operating across property, retail, industrial operations, education and hospitality cannot run a single promise across all of it credibly. What works is a hierarchy:
- A master brand at ownership level, signalling scale and stability to partners, lenders and governments rather than to consumers.
- Sector-level brands where the sectors genuinely differ in customer, promise and purchase logic.
- Operating brands where a business has its own established market presence worth preserving.
- And products that carry no corporate endorsement at all, because the customer is buying the product and the parent adds nothing.
The discipline is deciding which level does the work for each audience, rather than applying every level everywhere.
The same framework in business-to-business technology
The tests do not change. The answer often does – which is what proves they are tests rather than rules.
A technical products business arrived at a three-tier structure: a corporate brand, a product brand for the hardware and software, and application brands distinguishing configurations for different industries.
The strategic decision underneath it was explicit and worth quoting in principle: the product is always the hero, and the corporate brand operates behind it as the catalyst. When partners and resellers promote it, the product leads and the company endorses.
That is the opposite conclusion to the consumer-goods case above, reached with the same three tests, because the equity sat in the product rather than the parent. Same framework, different context, different answer.
The practical test
One question exposes whether your architecture is real:
If we acquired a business tomorrow, would we know – from principle rather than debate – whether it keeps its name?
If the answer requires a six-week project every time, you do not have an architecture. You have a set of historical accidents that has been drawn as a diagram.