The Brand Relationship Spectrum: Aaker and Joachimsthaler Mapped

The brand relationship spectrum is a framework for brand architecture: it describes how closely a parent brand should be connected to the products sold under it, arranged as a continuum and not as a binary choice.

Aaker and Joachimsthaler set it out in 2000, in Brand Leadership, to replace an argument that had been conducted as though there were two options — separate brands for everything, or one brand across everything — with a scale offering four positions and a number of gradations within them.

The practical question it answers is narrow and consequential: when a firm launches or acquires something, how much of the parent's name and equity should attach to it, and at what risk?

The four positions

House of brands

Separate, unconnected brands, each standing alone with no visible link to the parent. The parent is invisible to the consumer.

Why choose it. It allows brands to occupy positions that would be incompatible under one name — a premium and a discount brand in the same category, or products whose associations would damage each other. It contains risk: a failure or a scandal in one brand does not reach the others. And it permits channel-specific brands without the parent appearing to undercut itself.

What it costs. Each brand carries its own marketing investment with no shared equity, which is expensive and only viable at scale.

Two sub-positions exist. Not connected means no link at all. Shadow endorser means the connection exists and is not emphasised — a parent whose ownership is known to those who look but is absent from the packaging and advertising.

Endorsed brands

The product brand leads and the parent lends credibility from a secondary position.

Strong endorsement puts the parent prominently alongside the product brand. Linked name uses a shared element — a prefix or suffix carried across the range — rather than the full parent name. Token endorsement reduces the parent to a small mark on the packaging, present as reassurance rather than as a claim.

The purpose is to let the product brand develop its own associations while borrowing the parent's credibility on the dimensions where credibility is what the buyer lacks: quality, safety, durability, service.

Sub-brands

The parent leads and the sub-brand modifies. The sub-brand describes a variant, a segment or a level within the parent's territory.

Co-drivers share the driver role roughly equally, with both names doing real work in the purchase decision. Sub-brand as descriptor subordinates the sub-brand to a labelling function, where it distinguishes one offer from another without carrying independent associations.

Sub-brands stretch the parent into adjacent positions without requiring the buyer to learn a new name. The risk is that the stretch damages the parent, because associations run in both directions.

Branded house

One brand across everything, with descriptors rather than brands beneath it.

Different identity allows the shared name to carry somewhat different meanings in different categories. Same identity applies one consistent identity throughout.

The economics are the attraction: every investment supports every product, and the marketing cost of a new offer is a fraction of what a new brand would require. The exposure is the mirror image: a failure anywhere reaches everything.

Choosing a position

Aaker and Joachimsthaler give criteria instead of a formula, and four questions carry most of the decision.

Does the parent add credibility here, or borrow it? If buyers in the new category would find the parent's involvement reassuring, endorsement or a sub-brand transfers something real. If the parent is unknown or wrongly positioned in that category, the transfer is neutral at best.

Are the associations compatible? Two offers whose associations conflict — a luxury and a value line, a consumer and an industrial product — argue for greater separation. Incompatibility is a matter of what buyers infer, not of what the firm intends.

What is the risk of contamination? The more damaging a failure would be, and the more likely it is, the stronger the case for separation. Categories carrying safety, health or regulatory risk are systematically further towards the house-of-brands end.

Can the firm afford separate brands? This is the constraint that decides most real cases. Building an independent brand is expensive, and a firm without the volume to support one will end up at the branded-house end whatever the other criteria suggest.

Worked examples of each position

The positions are easier to hold apart against real portfolios, and the four below are the ones most commonly used in teaching because each is close to pure.

House of brands. A large consumer goods group holding dozens of cleaning, food and personal care brands, where the parent appears only in small type on the back of a pack and most buyers could not name it. Two of its brands frequently compete directly on the same shelf, which is only possible because no buyer connects them.

Endorsed. A hotel group whose individual properties or collections carry their own names, with the group name present as reassurance about standards and booking. The property brand supplies character and the endorsement supplies confidence that the room will be clean and the reservation will exist.

Sub-brand. A car manufacturer whose model names modify a parent that does the selling. Buyers choose the marque first and the model second, and a model that failed would damage the marque, which is the exposure the position carries.

Branded house. A financial services or technology firm applying one name across current accounts, insurance, mortgages and investments, with descriptors and not brands beneath it. A single reputational event reaches all of them, which is precisely what has happened to several such firms.

Mapping a portfolio this way is the framework's most productive use, because the incoherence shows up geometrically: a group holding an endorsed range, a branded house and two unconnected brands has three architectures and usually no stated reason for any of them.

What happens in an acquisition

Acquisitions are where architecture decisions are actually made, and the spectrum gives a sequence for them.

The immediate question is whether to retain the acquired brand at all. Retention preserves the equity the purchase price partly paid for and preserves the customer relationships attached to it; retirement captures the cost saving and the clarity of one fewer brand to support.

Where the brand is retained, the usual pattern is to move along the spectrum over time, not to jump. An acquired brand is first left unconnected, then endorsed, then reduced to a sub-brand, then absorbed as a descriptor and finally retired. Each step is reversible until the last and each transfers a little more of the equity to the parent.

The judgement is about pace. Moving too quickly loses customers attached to the acquired name before the parent has earned their trust; moving too slowly means paying to support two brands indefinitely while the acquisition's rationale was consolidation. The spectrum does not say how fast; it says what the intermediate positions are, which is what makes a staged plan possible at all.

Where firms get it wrong

Drifting rather than deciding. Architecture accumulates through acquisitions and launches, each sensible on its own, until a portfolio contains four positions on the spectrum with no rationale. The framework's most common use in practice is retrospective: mapping what exists in order to see the incoherence.

Over-branding. Creating a new brand for something that could have been a descriptor multiplies cost and divides attention. The test is whether the new name carries associations the parent could not, and if the answer is no, it is a descriptor.

Under-estimating reverse flow. A sub-brand or endorsement carries the parent's equity outward and carries the new offer's associations back. Firms model the first and ignore the second, which is how a parent brand ends up meaning something its owners never chose.

Deciding architecture in the marketing function alone. The choice commits legal, operations, procurement and customer service for years, and a position chosen without them produces an architecture the organisation cannot actually run — separate brands sharing one call centre that has to answer in two voices, or one brand across businesses with incompatible service standards.

Confusing architecture with naming. The spectrum is about the relationship between brands, not about what things are called. Renaming without changing which brand drives the purchase decision changes nothing the framework describes.

Limitations

The framework is descriptive and offers no way to weigh its criteria against each other when they point in different directions, which in practice they usually do. Two analysts applying it to the same portfolio can reach different positions and both defend them.

It also assumes the firm controls the architecture. Where a retailer's own brand sits beside the manufacturer's, or where a platform intermediates the relationship, the buyer's perception of who is behind a product is partly determined by others.

The framework also says nothing about internal cost. Each additional brand carries organisational overhead as well as marketing spend — separate agencies, separate approval routes, separate legal registrations in every market — and those costs frequently exceed the media difference that the analysis concentrates on.

And it predates the change in how brands are encountered. A framework built around packaging, advertising and shelf presence describes a world in which the consumer sees the architecture; where products are found through search and marketplace listings, the parent's endorsement may not be visible at the moment of choice at all.