Brand Extension: when it works

How to evaluate whether a brand extension will be successful: criteria and metrics

Kevin Keller’s Dartmouth model reduces the evaluation to three questions, and the order here is important — most companies check them in the wrong sequence.

1. Desirable (desired) — does the consumer want this product specifically from this brand. Not "is such a product needed at all," but "are they expecting it from you."
2. Differentiated (differentiated) — how the extension differs from what competitors are already selling in this category. According to research data, this is a critical factor: an extension that simply repeats an existing offer under a different name is doomed to fail, regardless of brand strength.
3. Deliverable (deliverable) — is the company actually able to produce and maintain this quality on its capacities, competencies, and channels.

In practice, this framework is supplemented by quantitative testing: concept testing with an assessment of purchase intent, transfer of associations testing (which attributes of the parent brand a consumer automatically 'borrows' for the new product), and separately — a test for differentiation relative to category leaders. If the expansion passes the first two tests but fails the third — you are facing a classic case of 'looks similar but not the same,' which is discussed further.

Fit versus simple similarity: why these are different things

The error that destroys more extensions than brand weakness itself — confusing product similarity with brand alignment. Xerox produced both copiers and computers — technologically these are close products. However, the Xerox brand was built on the association of 'copying, document flow,' not 'computing equipment,' and consumers did not transfer trust to the new category despite the technical kinship.

Fit is a multidimensional concept, and evaluation based only on one or two dimensions systematically yields biased results. Researchers identify six dimensions:

It is indicative that usage-occasion fit — correspondence by usage situations — gives the weakest effect among all six. Companies often mistakenly rely on it ("this is for the same life situation"), while in reality image-based and function-based measurements work. And another counterintuitive fact: improving fit gives a greater increase in the probability of success (61,4%), than improving the equity of the parent brand itself (60,6%). A strong brand does not save a weak fit — this is the main practical conclusion for those evaluating the risk of expansion.

Nike in fitness trackers worked, Samsung in cars — did not: analysis of the mechanics

Nike+ FuelBand looked like a risky move by a shoe manufacturer into electronics. But by image-based and function-based fit the match was almost perfect: the Nike brand is associated with measuring personal progress in sports, not with footwear as such — the fitness tracker logically continued exactly this association, rather than the technology of producing sneakers.

Samsung Motors and Xerox's attempt to enter the computer market failed for a mirror reason: technological proximity (feature- and resource-based fit) was high, but image fit and market fit were low. Car buyers do not automatically transfer trust from consumer electronics, and computer buyers in 80-s did not consider expertise in copying equipment relevant to computing tasks.

From this, we get a direct answer to the question 'what is the difference between similarity and fit': similarity is the coincidence in one dimension (usually technical), while fit is the coincidence in several dimensions at once, especially image and function. Technological kinship without image — is almost a guaranteed failure.

How expansion hits the parent brand

An unsuccessful expansion does not always harm the parent brand — and this should be understood before launch, not after. Reputation dilution occurs only when all three conditions coincide simultaneously:

• the experience with the expansion is noticeable and unambiguous — the consumer clearly remembers that the product was bad, not 'something didn't like';
• this experience contradicts expectations from the parent brand — it breaks what the brand was valued for;
• the consumer perceives the failure as diagnostic — that is, as an indicator of overall company quality, not as a coincidence in one category.

A simple weak, unexpressive extension usually goes unnoticed and does not cause damage — the consumer ignores it, rather than re-evaluating their overall perception of the brand. It is the striking and unambiguous failure that is dangerous, which appears systemic.

A separate practical compromise concerns downward price extension. If a brand launches a cheaper line, consumers perceive the new product itself as higher quality (name transfer effect), but at the same time, the overall perception of the parent brand's quality decreases. The short-term benefit turns into a cost for the main line — this should be taken into account when launching budget versions of premium products.

When to use a sub-brand instead of direct extension

A subbrand is a way to take a portion of the parent's equity, but to shield it from direct impact in case of failure. Fairfield Inn by Marriott works exactly this way: the consumer perceives Fairfield Inn as a separate product with its own identity, and even if the experience with it turns out to be negative, the assessment of Marriott as the parent brand is not reviewed — there is a connection, but it is not strong enough to automatically transfer the negative sentiment.

A subbrand is worth choosing in three situations:

• fit is weak on one or more dimensions, but the category's potential justifies the risk;
• a new category implies a lower level of quality or price compared to the main line (protection from the downward extension effect);
• the target audience for the new product differs significantly in expectations from the parent brand's audience, and the direct name may create cognitive dissonance.

Direct extension of the name, on the contrary, is justified when fit is high on several dimensions at once — especially image and function — and the company is confident in its ability to meet the stated quality level. In this case, a subbrand would simply weaken the transfer of trust that would have worked at full strength anyway.

In summary: how to make decisions in practice

Before launching an extension, it's worth taking a short check: assess fit across all six dimensions, not just the two or three that first come to mind; check differentiation relative to current category leaders; determine whether there's a risk of downward-effect on the perception of the parent brand's quality; and decide whether a sub-brand is needed as a shield. The statistic in 30% survival rate does not mean that extensions are a bad strategy. It means that most companies launch them without this check, relying only on the power of the name. Fit is almost always more important than equity — and this variable is exactly the one that should be considered first.