The brand architecture trap: when a house of brands becomes a mess

Australian companies are good at buying brands and bad at finishing the sentence. The deal closes. The slide arrives. Boxes, arrows, a masterbrand in a sober typeface, a row of acquired names that still have to earn their keep in the half-year. Everyone agrees the architecture is "a journey". Then the next acquisition lands before the last logo has been taken off a truck.

Architecture is not a poster of what you own. It is a rule for what the customer is allowed to confuse. In a market this small, confusion is expensive. You do not have American room to run twelve consumer brands at television weight. You have a handful of retailers, a handful of banks, a super system that already looks like an alphabet, state governments that re-cluster every term, and universities that collect campuses the way conglomerates collect SKUs.

The trap is treating the slide as strategy. The slide is a filing system. The strategy is the set of decisions the filing system is there to enforce.

How the mess usually gets built

The common path is acquisition without a retirement plan. A listed retailer buys a specialist. A bank buys a broker or a digital brand it does not want to contaminate with branch history. A university absorbs a college. A state government folds three departments into a cluster and keeps all three logos "for continuity". Each decision can be justified in the week it is made. The portfolio is what happens when nobody is paid to undo them.

Wesfarmers is the local proof that a house of brands can be a discipline rather than a collage. Bunnings, Kmart, Officeworks and the chemicals and lithium names do different jobs for different buyers. The group brand is for investors and employees. Customers are not asked to love a holding company while they buy a bag of soil. That clarity is the point of a house of brands. It is also the reason the model fails when it is copied by organisations that want holding-company warmth and specialist credibility at once.

Woolworths Group and Coles Group live with a different pressure. The supermarket name is the thing customers already have a relationship with. Adjacent brands — big-box general merchandise, liquor, a loyalty scheme, a media network, a fulfilment promise — either borrow that trust or dilute it. Every time a new name is stood up for a service the shopper thinks of as "the supermarket, but on a phone", you are spending architecture on a channel. Channels are not brands. They are ways of keeping a promise.

Banks have spent twenty years learning this the expensive way. A digital sub-brand can be a useful quarantine while you build a different service model. It becomes a mess when the quarantine never ends, the rates are the same, the complaints still hit the parent, and the customer is asked to hold two relationships for one loan. The ACCC and Australian Consumer Law do not care that your architecture slide shows a dotted line. If the advertising, the product and the entity the customer can sue are different stories, you have a conduct problem with a design budget.

Three models, used as tools

A branded house puts one name on the work. The advantage is efficiency. Every dollar, every staff badge, every tender response compounds. The risk is contamination. A failure in one line of business teaches the customer something about the others. Qantas discovered, publicly and repeatedly, that a parent name attached to loyalty, domestic flying, a low-cost sibling and a freight operation cannot choose when it would like to be one company. Customers already decided.

A house of brands keeps names apart so that trust, price and risk do not travel automatically. The advantage is permission to be unlike yourself. The cost is media, legal entities, packaging, trade marks and the management overhead of people who will lobby for their logo. In Australia that cost is higher than the global textbooks admit, because the audiences overlap. The same household shops at the supermarket, walks into the hardware warehouse, holds a super account and deals with a state agency. They will notice if you launch a fourth name for a job they already filed under the third.

An endorsed model tries to have both. The specialist name does the selling. The parent stands nearby as a quality mark, a government mark, a mutual mark or a university mark. Endorsement is the most abused of the three. It is what teams reach for when they cannot bear to kill a name and cannot bear to hide the group. Used well, it tells a true story: this college is of that university; this regional airline is of that carrier; this insurance product is issued by that authorised entity. Used poorly, it is a signature block on an advertisement that still cannot explain who is responsible when something breaks.

Choose the model by the job, the risk and the buyer — not by the aesthetic of the slide.

Retail and the weekly shop

Grocery in this country is a duopoly with a discounter that already trained shoppers to treat own brand as a feature. Architecture here is mostly about whether the parent name can stretch into general merchandise, liquor, media and fulfilment without teaching people that the food hall is now a holding company. Private label is the quiet architecture decision. A house brand that pretends to be a specialist, a value brand that pretends to be a farm, a premium brand that pretends to be imported: each of those is a promise. If the promises collide on the same bay, the customer learns that names are decoration.

Bunnings is a branded house that behaves like a category. The warehouse is the offer. Adjacent services — trade desks, hire, a cafe that exists to keep you on the concrete — sit under the same name because they are in service of the same job. That is the test. If the new thing is in service of the same job, stretch the name. If it is a different job, you are about to invent a brand you will have to feed forever. A name invented so the group can claim it is designing a category is still a name. Feed it or do not launch it.

Super, advice and the alphabet of funds

Superannuation already looks like a house of brands that nobody designed. Industry funds, retail funds, successor funds, legacy products that still send a statement: the member is asked to navigate a system that policy, not branding, created. After a merger, the temptation is to keep both names so that nobody feels acquired. That kindness is how you pay for two apps, two call-centre greetings and a decade of "which fund am I in?"

If the product, the fee stack and the trustee are becoming one, the name should follow. Endorsement can cover a transition. It should not be a permanent apology. APRA will care about the entity. Members will care about the login. Architecture that satisfies neither is a communications plan.

Government clusters and the citizen who just wants a licence

State governments reorganise. The brand work often arrives as a visual identity for a cluster that will be unpicked at the next election. Citizens do not live in clusters. They want a licence, a clinic appointment, a concession, a train that runs. Service NSW and its cousins in other states are, at their best, a branded house for transactions: one front door, many back-end departments. That only works if the front door is allowed to own the experience. If every department keeps a campaign brand, a program brand and a microsite, you have not built a house. You have built a street of pop-ups.

Procurement makes this worse. Tenders ask for "brand alignment" and then specify a program name that must appear on every artefact so a minister can announce it. The honest architecture move is to treat program names as dated, disposable and nested under the service the citizen already knows. Date them. Retire them. Stop designing crests for things that will not outlast a budget cycle.

Universities and the campus that wanted its own crest

Australian universities collect. A Group of Eight name is a branded house with a research reputation attached. Dual-sector institutions run TAFE and higher education under tensions that a single monogram cannot settle. Regional campuses, residential colleges and research institutes each arrive with a lobby that wants a mark. The question is not whether the college has history. The question is whether a seventeen-year-old, a research partner and a state education department are choosing the same thing.

If they are choosing the university, endorse and keep the campus language for wayfinding. If they are choosing a specialist institute that the university cannot operationally stand behind, you have a governance problem, and a logo will not fix it. Architecture cannot launder a reporting line.

Decide with costs, not adjectives

The useful questions are ugly.

Does the customer think this is a different job? A home loan and a credit card may be one relationship. A hardware warehouse and a department store are not. If the jobs are the same, extra names are vanity.

Does risk travel? Aviation, banking, health, food and childcare contaminate. A cheap sibling brand will still teach the parent something if the safety or conduct story breaks. The ACCC, the Therapeutic Goods Administration and the food standards regime do not read your endorsement hierarchy. They read the conduct and the claim.

Who pays to keep the name alive? In this market, a consumer brand that cannot hold a national media presence, a pack face and a service promise will drift into trade-mark maintenance. That is not a portfolio. That is a graveyard with renewal fees.

What does the shop floor say when asked who they work for? If the answer requires a slide, the architecture has already failed. Store managers, nurses, lecturers and call-centre leads are the distribution system for whatever you drew.

What will a lawyer do with the claim? If the advertised brand is not the contracting entity, Australian Consumer Law will eventually ask you to explain yourself. Architecture should make that explanation boring.

These questions belong in the same conversation as positioning. A name that cannot constrain a decision is a sticker. A portfolio of unconstrained stickers is the trap. The practice of writing the permissions table is more useful than another architecture workshop.

A house of brands is a decision to spend money keeping names apart. If you cannot say who pays, you do not have a house. You have a drawer.

Killing a brand is a brand decision

Retirement is the neglected craft. Australian groups will spend a year on a launch and then leave the old name on a facade because a regional manager is nervous. The nervousness is information. Sometimes it means the old name still does a job. Often it means nobody has written the sentence that tells staff what to say on the Monday the signage changes.

Write that sentence before you design the lock-up. Tell staff which customers are being moved, what happens to their points, their warranties, their course codes, their concession cards. Put the old name in a dated archive, not in a "family of brands" that implies it might come back. Families that never lose members are not families. They are accumulations.

Trade marks should be searched, filed and killed with the same seriousness. IP Australia will let you hold a graveyard. Customers will not shop one. If you are keeping a mark only so a competitor cannot have it, say so in the legal file and stop asking design to give it a living identity.

For a listed group, retirement and launch are disclosure events when the change is material to the equity story. That is a different sequence, and it is why rebranding an ASX company is not a design sprint. Architecture decisions that surprise the market will be treated as strategy surprises, because they are.

A working architecture is a list of permissions

Draw the slide if you must. Then attach a one-page permissions table that a non-designer can use.

Which names may appear on a price ticket. Which names may run paid media. Which names may hire. Which names may speak to government. Which names may issue a product that carries a regulated warning. Which names are allowed to have a colour of their own, and which must sit in the parent's system like a division.

The table will start arguments. Good. Arguments at the table are cheaper than arguments in a store. They are also how you discover that you do not have three brands. You have one brand and two projects that wanted a logo.

If the table cannot be written, you are not ready to brief identity. You are still in strategy, which is the only honest place to be. The visual work can wait. The next acquisition will not. Write the retirement rule before the bankers send the next teaser.

A house of brands is a decision to spend money keeping names apart. A branded house is a decision to spend reputation when any part fails. An endorsement is a decision to tell the truth about parentage. None of those is a personality. All of them can be done well in this country. The mess begins when a group wants the savings of one name, the protection of many, and the political ease of never disappointing a division that brought a crest to the merger.

Pick the cost you are willing to pay. Print it. Then stop adding boxes. If you are living inside a portfolio that cannot be explained without a slide, the guest guidelines want the ugly table, not the pretty one.