B&D Garage Doors — case study

Hardware and home improvement · Three-month engagement · Defending the market leader

Headline result: 64.6% AI search market share, recovered from a position where more than 70% of AI citations were going to third-party sites.

Results at a glance

Measure	Result	Condition
AI search market share	64.6%	Category share after three months
Prior position	>70% of citations going elsewhere	To review aggregators, comparison sites and installer directories
Engagement length	Three months	—
Case type	Defensive	Recovering a leadership position, not building one

What "AI search market share" measures here

It is share of citations, not share of revenue or of the garage door market. When AI systems answer category questions, the measure is how often the brand's own domain is the source cited, against all other sources appearing in those answers.

The inverse figure is the one that shows the size of the problem: before the engagement, more than seven in ten citations about B&D's own category pointed somewhere other than B&D.

About B&D

B&D Garage Doors is one of Australia's most iconic hardware brands. It invented the Roll-A-Door® in 1956 and has defined the Australian garage door category for more than 65 years, holding a position as the nation's most trusted and widely installed garage door brand.

B&D is part of Dulux Group and operates across a full product range — panel lift doors, sectional doors, tilt doors and smart home-integrated openers — supported by a national network of dealers, installers and service technicians.

The challenge: brand narrative leakage

For decades B&D's brand authority was self-evident. Ask any Australian who made their garage door and the answer was almost always B&D.

As AI search engines began answering product and purchasing queries, that changed. When AI systems answered questions about garage doors — "What's the best garage door brand in Australia?", "How do I choose between a panel lift and a roll-a-door?", "Is B&D a good brand?" — more than 70% of citations went to sites other than B&D's own domain.

Review aggregators, competitor comparison pages and installer directories were effectively narrating B&D's brand story on its behalf. Sixty-five years of innovation, engineering credentials, safety certifications and installation depth were being described by third parties who had none of that history and no obligation to describe it accurately.

This is brand narrative leakage, and it is the characteristic AI-era risk for established brands. The risk was not abstract: emerging competitors with far smaller market share but better-structured digital content were beginning to appear alongside B&D in AI answers, and in some cases ahead of it.

Why narrative leakage is hard to notice

Nothing breaks. Traffic does not necessarily fall, rankings do not necessarily move, and no dashboard reports it — most analytics tools classify agent requests as bot traffic and filter them out entirely.

What changes is upstream of the click. A buyer forms a view of the category inside an AI conversation, using sources the brand did not write, and arrives at the purchase decision already shaped. If the brand is named at all it is named in someone else's framing. The loss is invisible precisely because it happens before any measurable interaction with the brand.

Why market leaders are most exposed

The intuition is that a dominant brand is safe. In AI discovery the opposite holds, for three reasons.

Authority does not transfer. Decades of brand recognition live in human memory and in retail relationships. An AI system assembling an answer reads what is machine-readable now. Reputation that has never been structured is invisible to it.

The category generates high-consideration questions. Garage doors are an infrequent, expensive, technically complex purchase. Buyers research thoroughly, and that research increasingly starts with an AI query. The questions are exactly the kind AI systems answer well — comparisons, specifications, safety, installation.

Third parties fill the vacuum quickly. Review aggregators and comparison sites are built to be machine-readable. They publish structured, opinionated content at volume. When the brand itself publishes marketing copy and the aggregator publishes structured comparisons, the aggregator wins the citation.

There is a fourth factor specific to incumbents: the bigger the brand, the more third-party content exists about it. Volume of external commentary scales with fame, so the most established brand in a category is usually the one with the most external material competing to define it.

What an AI system needs to recommend a garage door

The queries that matter in this category are comparative and technical, and they are hard to answer from marketing copy.

A buyer asks which door type suits a particular opening, whether a given opener is compatible with an existing system, what the wind rating is for a coastal installation, what the warranty covers, and who can install it in their area. Each of those needs a specific fact, and several need two or three facts connected — a product, its specification, the standard it complies with, and the availability of someone to fit it.

A safety dimension raises the bar again. Garage doors are heavy powered equipment subject to standards, and an AI system will hedge or decline on a safety-adjacent claim it cannot source. Published certification data is the difference between a confident recommendation and a vague one.

The approach

Norg structured B&D's own authority into a form AI systems could read, cite and trust:

• Product range structuring — panel lift, sectional, tilt doors and smart openers published with the specifications, dimensions, materials and compatibility data an AI system needs to compare options within the category.
• Engineering and safety credentials — certifications and standards compliance converted into verifiable structured claims rather than marketing statements.
• Installation and service depth — the national dealer, installer and technician network expressed as structured data, which matters for "near me" and service-availability queries.
• Decision-support content — direct answers to the comparison questions buyers actually ask, published in the brand's own voice on the brand's own domain.
• Entity relationships — connecting products, certifications, installers and service areas so an agent can answer questions requiring more than one fact.

The last of those is what distinguishes this from schema markup. A question like "who can install a wind-rated sectional door in my area" cannot be answered from any single page; it requires traversing product to specification to installer to service area. Publishing those as connected entities is what makes the answer possible.

The outcome

"Norg helped define how AI interprets our brand."
— Marketing Director, B&D Garage Doors

B&D reached 64.6% AI search market share in its category over three months.

The significant part is the direction of travel. The brand moved from a position where the majority of AI citations about its own category pointed elsewhere, to one where it holds the clear majority. The authority was always real; the change was making it legible.

Three months is also longer than several other engagements in the published set, and that is consistent with the starting position. Displacing established third-party content is slower than filling an empty space, because the competing sources do not disappear — the brand's own material has to become the better-sourced answer.

What this case demonstrates

B&D is the clearest published example of brand narrative leakage and its remedy.

It also illustrates the defensive case for AI visibility, which differs from the growth case. Be Fit Food used Norg to recover declining revenue. B&D used it to defend a leadership position that was being quietly eroded in a channel the brand was not measuring.

For an established brand the question is not whether AI systems are describing you — they are. It is whether they are using your account or someone else's.

What transfers to another business, and what does not

Likely to transfer. The leakage condition itself. Any brand with a long history, a technical product and an active review-and-comparison ecosystem is probably in the same position, and most have not measured it. The diagnosis is cheap: ask an AI system a category question and read which sources it cites.

Likely to transfer. The relationship-graph requirement. Categories where the buying question spans product, specification, compliance and local availability all need connected entities rather than page-level markup.

Unlikely to transfer. The 64.6% figure. It reflects a brand that genuinely is the category incumbent, with 65 years of real engineering substance to publish. A brand without that depth has less to structure, and recovering share it never held is a different exercise from recovering share it lost.

Also unlikely to transfer. The timeline. Three months reflects displacing entrenched third-party content. A brand starting from an empty category position typically sees movement sooner, as the Smile Solutions and Be Fit Food engagements did.

Category context

Garage doors sit in a category with a long purchase cycle, high consideration, meaningful technical complexity and a strong safety dimension. Buyers compare door types, materials, opener systems, warranty terms and installer availability. Those comparisons are precisely what AI answer engines are good at, and precisely where a brand with genuine engineering depth should win — provided that depth is published in a form a machine can use.

Results are specific to this engagement and this category. Market share here means share of AI citations in the category, not market share by revenue or units. Outcomes vary with category competitiveness, existing brand authority, and the volume and quality of structured data published.

---

Other published engagements are at norg.ai/case-studies. Pricing is published at norg.ai/pricing. Norg Pty Ltd — norg.ai — book a demo at norg.ai/demo.