Brand Architecture: Which Name Leads?
A company adds a second product, and a naming question suddenly becomes a customer question. Should the new offer carry the company name, stand on its own, or show a smaller line saying who is behind it? The answer affects what a buyer expects before trying the product. It also affects what the company must explain every time it introduces another offer. The same issue appears when an acquired brand joins a portfolio or when outside developers build products that work with the company’s own.
Brand architecture is the set of roles and relationships among names in a portfolio. David A. Aaker and Erich Joachimsthaler describe it as the structure that specifies which brands have which roles and how they relate to one another. Their useful test is the driver role: when a customer says what they bought, which name chiefly explains the choice? A corporate name can lead the decision, share that job with a product name, merely reassure the buyer, or be almost invisible. Their brand relationship spectrum sets out these positions.
My starting choice is to give an established name as much work as it can do credibly. A separate name must earn the expense of becoming known. Yet the parent should not appear prominently just because the business owns the product. A name that promises the wrong benefit or blurs who made an offer can make a launch harder to understand. The decision is therefore about the buyer’s inference: what does each visible name help the buyer know, and what mistaken inference might it invite?
The architecture starts with the buyer’s answer
An internal portfolio chart tells employees what the company owns. A useful brand architecture tells a customer what the names mean together. If the company name appears first on every product, buyers are invited to transfer what they know about that name to each one. If distinct product names lead, each must carry more of its own meaning. An endorsement sits between those positions: the product makes its own promise, while a parent lends some reassurance. This is why the same company can own several products without presenting them as one family to the buyer.
Look at two documented portfolios. Atlassian’s product page groups Jira, Confluence, Loom, and Rovo under Atlassian. The names distinguish products, while the corporate setting makes their relationship visible. Procter & Gamble’s brand page lists Tide, Gain, Pampers, Gillette, and other distinct consumer brands. A shopper can understand Tide as a laundry brand without first learning the P&G portfolio. These pages show different presentations; they do not, by themselves, establish which structure performs better for either company.
The practical question is how much work the parent name should do at each point where someone chooses. A name in a corporate footer is doing a different job from a name at the front of the product label. A parent logo on packaging may identify ownership but leave the product brand as the purchase driver. A partner’s phrase “for Jira” may describe compatibility without making Atlassian the maker of that partner’s app. Position, size, wording, and context therefore matter as much as the names on a diagram.
Four relationships, from one lead name to many
At the closest end is a branded house. One master brand leads across offerings; additional words mainly tell buyers which product or service they are looking at. The advantage is that the company can apply recognition of the master name across launches. The constraint is reciprocal: a new offer must fit the expectations attached to that name, and the offer can change those expectations. Aaker and Joachimsthaler describe both the communication efficiencies and the risk of stretching a master brand too far. Their spectrum treats the branded house as one end of a continuum.
A subbrand gives a product a more active name while keeping it closely tied to a master brand. The two can share the buyer’s attention: the parent supplies a frame of reference, while the product name adds a distinct reason to choose. That is useful when a new offer belongs to the same broad promise but needs a recognizable proposition of its own. It also puts more strain on the parent than a light endorsement. If the product disappoints or points in a direction the parent cannot support, the close link makes the association harder to contain. Aaker and Joachimsthaler distinguish subbrands from endorsements partly by how much of the purchase decision the master brand drives. Their explanation of the driver role makes the distinction concrete.
With an endorsed brand, the product name leads and another brand signals who stands behind it. The endorsement can be prominent or modest. The intended reading is different from a subbrand: the buyer should first grasp the product’s own promise and then take reassurance from its backer. Endorsement is worth considering when a parent can add credibility but would distort the offer if it became the main name. It still requires a clear and consistent presentation. A tiny parent mark that buyers miss cannot be assumed to carry the full value of a highly visible endorsement; Aaker and Joachimsthaler explicitly caution against overestimating a weak token endorsement. Their discussion of endorsed brands and token endorsers explains that limit.
At the most separate end is a house of brands. Distinct names lead for distinct offers, while the owner stays in the background. The reward is freedom to give each offer a sharply defined promise and to avoid forcing incompatible associations under one name. The cost is that each brand needs support to become known and trusted. P&G’s consumer portfolio shows why this option is more than a textbook label: its current brand list places Tide, Gain, and Cheer in fabric care, alongside brands in other categories. That observation supports a description of its portfolio, not a claim that every owner of multiple products should imitate it.
These are relationship types, not four boxes into which every company must fit. Aaker and Joachimsthaler present a spectrum that includes gradations within endorsement and subbranding. NielsenIQ also describes branded house, house of brands, endorsed, and hybrid approaches. The right level of connection may differ between a core product, an acquired business, and a partner app in the same portfolio. What matters is whether each relationship says something deliberate and intelligible to the person choosing.
Choose the parent name only when it helps both sides
Begin with what the parent name contributes to the new offer. Does it suggest a relevant capability, a level of quality, or a familiar way of using the product? Would a buyer understand the new offer faster with that name attached? If the answer is yes, a master brand or subbrand may save the work of teaching an entirely new identity. If the parent’s associations are irrelevant, prominent use may only consume space in the name and make the promise less precise. Aaker and Joachimsthaler frame this as a choice between the value the master brand adds and the distinct association a separate brand can own. Their selection questions also ask whether the business can support a new name.
Then reverse the direction of the question. What will the new product teach people about the parent? A close naming link is a two-way transfer. A product that strengthens the parent’s existing promise can justify a visible connection. An offer aimed at a very different need may weaken that promise, even if the established name would give the launch an initial lift. Aaker and Joachimsthaler warn that extending a name for a new business unit’s immediate benefit can damage the associations the parent needs elsewhere. Their account makes the effect on the master brand part of the choice.
Related products do not settle the answer on their own. Jungju Yu’s model in Marketing Science distinguishes relatedness on the supply side, such as shared production capability, from relatedness on the demand side, such as correlated customer preferences. In that model, a shared brand can let buyers use a firm’s track record across product markets; under some conditions, separate names can instead support investment in each market. This is a theoretical result with stated assumptions, not a measured rule that close product categories should always be split. Its value for a manager is the sharper question it poses: what exactly can buyers transfer from their experience of one product to an expectation about the next?
My preference, where the parent has a relevant promise and the new offer reinforces it, is to keep the connection visible. Start with the parent as the lead name if a descriptive product label is enough; move to a subbrand when the product needs its own durable recognition. If the offer requires a distinct promise but the parent still lends useful reassurance, let the product lead and endorse it clearly. Use a separate brand when the parent would mislead, constrain the offer, or confuse the channel in which it appears—and only if there is a realistic plan to build that name. The price of the separate choice is repeated explanation of who the product is, what it does, and why buyers should trust it.
That sequence is a decision aid, not a rule that every portfolio must start with a branded house. It depends on facts a company must know about its own buyers: which name they recognize, what they think it promises, and whether recognition carries into the proposed category. A diagram cannot supply those answers. If those facts are unknown, I would avoid a portfolio-wide renaming and first test the competing presentations with people who would actually choose the product. The unresolved issue is not whether the names look orderly together; it is which presentation makes the offer understandable without promising the wrong thing.
Give product names and partner names different jobs
Atlassian’s portfolio illustrates a common source of confusion. Its product page presents Jira, Confluence, Loom, and Rovo within the corporate portfolio. A customer can encounter a distinctive product name and still see the Atlassian connection. That does not mean an outside developer can use the same naming pattern as if its app were an Atlassian product. The ownership relationship changes the message that a name must carry.
Atlassian’s Marketplace partner guidelines make that boundary unusually explicit. Partners are told to present themselves as third parties and use their own marks. The page allows a format such as “Project Planner for Jira” but rejects starting a partner app name with “Jira,” because that could imply Atlassian made the app. The preposition “for” is small, yet it tells the buyer that the partner’s product works with Jira rather than being Jira itself. This is brand architecture at the level of a real listing, not just a corporate chart.
The lesson reaches beyond a software marketplace. A company with distributors, licensed products, or integration partners should decide whether another organization’s name indicates ownership, endorsement, compatibility, or merely a commercial relationship. Those meanings should appear in the wording and visual order people actually see. Do not use the parent logo as a vague stamp of approval when the company has not taken the role a buyer would infer from it. Do not hide a genuine parent relationship if that relationship is the main reason a new offer is credible. The goal is an accurate reading at the moment of choice.
Make the choice usable at every point of contact
Once the relationship is chosen, write it down as a short set of decisions for each name. State which name leads, which supports, what the product name promises, and where the parent appears. Then look at the places that put those decisions under pressure: a website menu, product page, search result, package, sales proposal, or partner listing. A name that seems balanced on a portfolio slide can read quite differently when it is truncated in a search result or appears next to a competitor’s product on a shelf. This review is an application of the chosen relationship, not a separate branding theory.
For a new offer, I would settle the customer-facing sentence before approving a visual system: “This is [offer], made by [company]”; “This is [company’s] [offer]”; or “This works with [product] and is made by [partner].” Each sentence assigns a different job to the names. If the team cannot agree which one is true, more design variants will only disguise the disagreement. The final name, copy, and placement should all tell the same story.
For an existing portfolio, changing every name at once is a costly way to seek tidiness. First identify the names that already drive customer choice and the ones that exist mainly as internal labels. Then inspect the exceptions: products with a distinct audience, acquisitions with recognition of their own, and partnerships where ownership could be misread. NielsenIQ describes architecture as a way to codify component roles and relationships and notes that implementation may involve a launch or rebranding. Its overview also treats later adjustment as part of managing the portfolio. That supports making the structure explicit, while leaving the actual choice dependent on the portfolio in front of you.
The strongest architecture is therefore easy to state in ordinary language. A buyer should be able to tell which name identifies the product, which organization stands behind it, and whether another name signals a family resemblance or a separate maker. If the parent helps the buyer make a sound inference, let it lead. If the product needs its own promise, give it a visible role. If the relationship is only endorsement or compatibility, say exactly that. The structure has done its job when the names clarify the offer without asking the customer to learn the company’s org chart.