Brand hierarchy: branded house or house of brands?

The structure is decided before the name is

A brand architecture, or brand hierarchy, answers one question that usually arrives disguised as a naming question: what should this new thing be called in relation to everything else we have? The answer decides what kind of name the job needs. Under one model the thing needs a descriptor. Under another it needs a name that can stand entirely alone.

That is why the structure comes first. A name can be excellent and still wrong for its place in the hierarchy, and no amount of creative quality repairs that. The same sequence applies here as in a brand strategy: the decision about difference and audience is made first, and the names are judged against it.

One name on everything: the branded house

A branded house puts a single master brand on everything and tells the products apart with descriptors. Google is the familiar case: the master name stays constant, and Maps or Search say what the product does. Every product benefits from every other product's visibility, and the marketing budget compounds in one place.

The cost sits in the same place. Reputational damage travels across the whole range, and a product that does not fit the master brand's meaning has no room to differ. When Google reorganised under Alphabet in 2015, Larry Page gave the reason as being able to run things "independently that aren't very related" to Google. When the products stop sharing a meaning, a branded house stops being the right container.

For naming, the discipline is restraint. Under this model the job is usually descriptor work, and every invented sub-name spends attention the master brand could have kept.

Independent names under a silent parent: the house of brands

A house of brands does the opposite. The parent stays in the background and each brand stands on its own. Procter & Gamble's brand list in its 2018 filing names Tide, Pampers, Gillette and Ariel side by side, and a buyer of one has no reason to know the others share an owner.

The advantage is separation. Each brand can address a distinct audience, and trouble in one does not travel. The price is that every brand pays for itself: its own marketing, its own trademark work, its own story. It also makes the heaviest demand on naming, because each name has to do the whole job alone, with no parent to lean on. It suits portfolios whose audiences really differ, and it is a poor fit for a company with one audience and four products.

In between: endorsed brands and sub-brands

Two models sit between those poles, and they are easy to confuse. An endorsed brand has its own name and identity and visibly carries the parent as a guarantor. Marriott's filing lists Delta Hotels by Marriott and Protea Hotels by Marriott beside brands such as Courtyard and Residence Inn. The endorsement lends trust to a brand that can still speak to its own guests.

A sub-brand is tied more closely to the master brand and adjusts what it stands for: the iPhone beneath Apple is the usual example. The practical difference is the direction of dependence. An endorsed brand could in principle walk away from its endorser. A sub-brand is partly defined by its parent and cannot.

Earning a name of your own

The harder question is how many names a company should have at all, and we would answer it with a bias toward fewer. A new thing earns its own name when it has an audience of its own, a story of its own, or a lifespan of its own. A feature that shares the buyer and the story of the main product needs a descriptor, and calling it something clever creates a second thing to explain.

Lexus is a useful reference for the opposite case. Accounts of its origin say the Toyota board first meant to launch the new luxury car as a Toyota, while the head of Toyota's US division pressed hardest for an all-new luxury model. Launching in 1989 under its own name let the car speak to a different audience than the parent did. That is a decision about the hierarchy, taken before any name was chosen.

Most hierarchies are accidents

Look at a company ten years old and the hierarchy is rarely a design. It is the sum of naming decisions taken one at a time, by different people, each reasonable alone. The result is a feature with one name in the interface, another in the documentation and a third in the sales deck. A written naming architecture stops that: who gets a name, what kind of name, and how it sits beside the parent.

Aaker and Joachimsthaler's brand relationship spectrum, published in 2000, describes the choice as a range with several positions in between. WARC notes that in practice it has shrunk to a choice between two opposites, branded house or house of brands. The wider reading is more useful. Most real portfolios sit somewhere between, and the useful question is how much of the parent each product should carry.

Four questions that show which structure fits you

Which model suits a company is answered by its audiences and its budget, not by taste, and that is also why the structure matters for a company of any size: every new product forces the question, and a structure written down once answers it each time. Work through these in order, and stop at the first answer that is clear.

1. Do the products share a buyer and a story? If the same person buys them for the same reason, one master brand with descriptors is the natural fit. If the buyers or the stories differ, a single name has to stretch over both and usually fits neither.
2. Would trouble in one hurt the others? If a failure, a recall or a bad review in one product would travel to the rest, separate names are a form of insurance.
3. Can you pay for every name? Each independent brand needs its own marketing and its own trademark work. If that budget does not exist, a descriptor under the master brand is the honest answer.
4. Does the parent help or hurt in this market? Where the parent is trusted and fits, endorse the new brand with it. Where its associations work against the product, keep the parent out of sight.

The answers tend to agree with each other. When they do not, the first two carry more weight than the last two, because audiences are harder to change than budgets. To develop the structure, list every name you use today, group them by buyer and story, pick a model for each group, and write the rule for what happens at the next product. That document is the naming architecture, and it is short.

Decide it with the second product

For a new company the answer is simple. You begin as a branded house of one, and you stay one for as long as audience and story stay the same. The decision becomes real when the second product appears, and that is the cheapest moment to write the rule down: this will be a descriptor, this will stand alone, this will carry the parent.

When it is a name that has to stand alone, it needs everything a first name needs, including screening for language, culture and category. Each additional independent brand multiplies the registration work, which is one more reason to make the choice deliberately. Our AI agents treat the structure as part of the briefing, so a name is cut against the position it has to hold. See how it works, or the naming module for what a run produces.