Being a brand strategist is pretty easy. We only need to know how to do three things:

  1. Articulate what makes a brand different and desirable (brand positioning)
  2. Establish a clear role for the different parts of a brand portfolio (brand portfolio strategy)
  3. Communicate the relationships between different parts of a brand portfolio using appropriate visual and verbal signposts (brand architecture)

There are myriad tools for brand positioning and portfolio strategy. If you want to position a brand, you can use a two-by-two matrix, Maslow’s hierarchy of needs, archetype theory, Simon Sinek’s Golden Circle, or something more old school like a brand pyramid or brand platform. If you’re developing a portfolio strategy, then you might call upon the Five Ws, Jobs To Be Done, or the 80/20 Rule. But if you’re looking for a model to explain brand architecture, there is only one framework to turn to: The Brand Relationship Spectrum.

The Brand Relationship Spectrum was introduced to the world by David Aaker and Erich Joachimsthaler in 2000. In their own words, the spectrum is:

“a powerful brand architecture tool… intended to help brand architecture strategists employ insight and subtlety to subbrands, endorsed brands, and their alternatives.”

Over the intervening 20 years, much of the subtlety has been lost. The spectrum has evolved in people’s minds to a fairly simple choice between two opposites:

“Branded House or House of Brands?”

A Branded House uses a single master brand to sell a range of products or services, with only descriptive names to distinguish different parts of the offer, while a House of Brands contains independent, unconnected brands:

There are benefits to both approaches. Operating a Branded House makes it easier to concentrate your marketing budget in building a single brand, as well as making it easier to cross-sell between the different parts of your offer. On the other hand, a House of Brands gives you plenty of flexibility to acquire and sell brands, as well as allowing you to target a broader range of products and speak to a more diverse set of audiences. But each approach also has its drawbacks. A Branded House can be extremely monotone and very corporate, while a House of Brands can be chaotic and cluttered. The question of “House of Brands or Branded House?” suggests a choice between two opposites: rigid discipline on the one hand and chaotic freedom on the other.

Of course, the question is a silly one. It’s absurd to claim that there are only two types of brand architecture system in the world. And in any case, the choice presented above is a gross oversimplification of Aaaker and Joachimsthaler’s spectrum, which contains four basic strategies and nine substrategies:

According to Aaaker and Joachimsthaler’s spectrum, there are two intermediary types of brand architecture strategy between Branded House and House of Brands:

  • Subbrands: which introduce an additional layer of meaning to a master brand
  • Endorsed brands: in which the endorser brand provides assurance that the endorsed brand will live up to its promise

The Brand Relationship Spectrum is further divided into a set of nine substrategies that capture some subtle differences in terms of whether brands in a portfolio share a common identity, or name. Even after twenty years of living with this framework, I’m afraid I have very little positive to say about it. It is so utterly, utterly awful at capturing any of the insight and subtlety required to really understand how brand architecture works in the real world.

Here’s why:

There are almost no ‘pure’ examples of any of the nine types of substrategy (or even the four types of strategy). In other words, I had to cheat to create the above visual. Here’s what a more honest version would look like, if I included a couple more brands from each portfolio:

There are very, very, very few mature brand portfolios that exclusively employ any one of these strategies or substrategies. Accenture functions for the most part as a Branded House, but has acquired creative agencies that it has chosen to endorse rather than assimilate into Accenture divisions. Unilever largely operates as a House of Brands, but in 2013 the Group created a subbrand, Project Sunlight, in an attempt to motivate millions of people around the world to adopt more sustainable lifestyles. Almost every brand portfolio I have ever worked on has been a hybrid portfolio. And to be fair to David Aaaker, he admits as much in his 2004 book, ‘Brand Portfolio Strategy’:

“Nearly all organizations will use a mixture of all four branding routes on the brand relationship spectrum. A pure house of brands or branded house is rare… The challenge is not to create a single house, but a village where all the subbrands and brands fit in and are productive.”

The strategies and substrategies in the brand portfolio spectrum were never intended to describe entire brand architecture systems; they were supposed to be a way to think about individual brands within a system. But assigning arbitrary labels to individual brands is a pointless intellectual exercise. Developing great brand architecture requires us to think in terms of systems. It involves solving some of the biggest riddles in branding:

  • How do you grow a brand without losing what makes it special?
  • How do you get the biggest bang for your marketing buck?
  • How far can you stretch a brand before it becomes meaningless?
  • How can you cater to the wants and needs of diverse audiences without diluting what you stand for?
  • How do you direct people to the parts of your brand that they will love the most (and that will earn you the most money)?
  • And (for corporate brands) how do you balance cultural unity with the desire of different parts of your organisation to go about their work with a degree of autonomy?

A great brand architecture system will meet pretty much all of these challenges for you. That’s why it is one of the most powerful ways in which brands can create value. And thanks in part to the impracticality of the Brand Relationship Spectrum, even the most valuable brands in the world have architecture systems in obvious need of improvement. Here’s what Apple’s brand looks like:

Although typographically the brand system is neater than most, there’s a lot going on here – which is understandable for one of the most valuable brands in the world. To help, I’ve reorganised the portfolio based on the categories provided in Apple’s most recent annual report:

Some categories are messier than others. The services and wearables categories are particularly busy, which is possibly because these are also the fastest-expanding parts of the business, growing at 16% and 41%, respectively (based on the 2019 Form 10-K). I doubt it’s an accident that these two pillars contain the subbrands that are locked up against the Apple logo, as well as those that appear in block capitals, like WATCH and MUSIC. And a couple of random “+” signs thrown in for added emphasis. It’s the brand architecture equivalent of attention-seeking behaviour.

So, how would I apply the Brand Relationship Spectrum to try and improve the Apple brand architecture?

I wouldn’t.

Trying to work out which brands belong to which strategies and substrategies would distract from a big question lurking in the visual above:

How does Apple intend to grow?

For context, although services only account for around 18% of sales, they deliver double the profit of the products business (iPhone, iPad, Mac, wearables, home & accessories) and consequently contribute around 30% of gross margin. Instead of thinking about brand architecture in terms of a Relationship Spectrum, I find it far more useful to think in terms of a hierarchy. And that hierarchy is different for every brand. So, the first step in developing brand architecture is to define the brand’s hierarchy: how many layers should it contain? The Apple brand architecture centres on a master brand, but this is supported by product brands, service brands, product variants, proprietary apps and operating systems. There may be missing layers in the current system: for example, Apple currently sells a single monthly subscription package for Apple TV+, as well as separate monthly charges for iCloud storage. So, it’s possible that in future the brand may want to expand the types of subscription packages it sells to cover multiple services – or even multiple devices. Consequently, I’d be tempted to think about adding a ‘subscription’ layer to the hierarchy. Partner brands also seem to be largely absent in the Apple world – but if Apple is keen to move further into Financial Services, then it may need to collaborate more closely with a range of FS partners. Moving into lifestyle services may also require partnering with lifestyle brands, so I’d also consider adding a layer to deal with co-branded partnerships. Really great brand architecture is built on an understanding of all the likely ways a brand may grow in future.

The visual above shows how ill-defined some of the layers in the hierarchy currently are. Ideally, you’d want to see a similar approach to naming and design of brands within each layer. But there are glaring inconsistencies at every layer. There’s a really basic tidying up job to do here. Beyond that, there’s a second important step: to determine the order of importance of each layer in the hierarchy. At this point we bump into another big question:

What kind of a brand does Apple aspire to be?

  1. A product brand that cross-sells services to its users?
  2. A service brand that uses products as a delivery mechanism?
  3. Or a rounded ‘experience’ brand that seamlessly brings together products and services?

Each of these approaches will have consequences for the brand architecture because they suggest a different ordering of products and services in the hierarchy:

The schematic above shows how each of the options might affect Apple’s brand architecture. I’ve visualised a couple of representative brands to bring each layer of the hierarchy to life, but please excuse the very rough PowerPoint design skills on display.

In scenario A, products are given prominence in the architecture, with services relegated to a minor role. This very much echoes the Apple brand of the past, in which our primary relationship with the brand is through the devices we fetishise. Product variants are also prominently branded because the aim is to encourage people to buy the latest version – which is less likely if these are hidden towards the bottom of the hierarchy. Given their most recent financials, Apple would be bonkers to adopt this approach – even if the current “shouty” Apple Watch logo does suggest a move in this direction.

In scenario B, services are given greater prominence in the architecture. The idea here would be to cross-sell as many services as possible, with the product families playing a supporting role: my iWatch and iPhone and iPad are merely the means through which I access all of my media and Cloud content. This also feels a little dodgy – given that a significant number of people still seem to want to own the very latest Apple products.

Scenario C is much closer to a plan for world domination. Here, a new type of sub-brand would be introduced that provides subscriptions to Apple products and services on a rolling basis. I’d imagine that at the top end this would involve something along the lines of unlimited iCloud storage, access to all music and video content and instant upgrades of all Apple devices. In this scenario, the distinction between products and services becomes less important, as these are both components of the subscription packages.

I’m not trying to predict the future here, but hopefully the three scenarios demonstrate that brand architecture is a powerful tool for planning and facilitating how brands grow. None of these scenarios would have been achieved by applying the Brand Relationship Spectrum, or by asking Tim Cook if he is building a Branded House or a House of Brands. Developing a great brand architecture isn’t easy: it’s one of the thorniest and most frustrating parts of the job – but it’s also one of the most rewarding. Taking a hot mess of inconsistent brands and evolving them into a coherent system makes organisations more efficient and more effective – and there are very few brand architecture systems out there that can’t be improved. Trying to assign individual brands in the system to four or nine boxes along a pre-defined spectrum is a distraction. A far better starting point is to consider the four following lines of thought:


  • What does the organisation hope to achieve?
  • Which brands in the system are creating the lion’s share of the value, and where will the value come from in future?


  • Who are the different audiences that the brands in the system needs to interact with?
  • How similar or different are their needs?
  • Is the ultimate aim to encourage audiences to deal with multiple parts of the system (e.g. maximise cross-sell), or is it better to create dedicated parts of the system for each audience group?


  • How does the organisation want to manage risk: if one brand in the system fails, how should this affect the other parts of the system?
  • Which brand do you want employees to feel they work for?


  • How many layers are appropriate for the hierarchy: what types of brands, partnerships, ingredients, variants, ranges, campaigns, programmes and initiatives does the organisation want to engage in?
  • What’s the order of importance of each of these in supporting the organisation’s ambition?
  • And what is the most appropriate way to signpost the brands that sit in each layer of the hierarchy – visually and verbally.

I appreciate this is infinitely more difficult than the simplistic question of “Branded House or House of Brands?” but when it comes to brand architecture there are no simple pre-packaged solutions.

Uniqueness matters.

If you’ve enjoyed this article and would like to download a PDF version, just click on the button below:

Leave a Reply