Marketers have long treated branded search as an effective proxy for product demand. Increasingly, that assumption seems less reliable.
Evaluative purpose of search. The product has an effect where that intention goes. Relationships have not changed enough that branded searches are a reliable indicator of product demand.
Our data reinforces that relationship. Across our client portfolio, branded customer acquisition costs are, on average, 76.6% lower than unbranded acquisition costs. However our data also reveals a very consistent trend. In the past month, search demand fell by 11.1%, despite a flat auction area.
If branded search remains a reliable indicator of product demand, the explanation is straightforward: demand itself must be weak. Evidence suggests otherwise.
10X your SEO with Semrush for Enterprise.
The world’s most powerful SEO platform, built for the Enterprise.
Request a demo
When the representative stops showing the truth
To understand why, it helps to eliminate some of the more obvious explanations.
- The first possibility is that consumers enter the market with a weak choice of product. Based on general data from Google Keyword Planner and Semrush, the evidence does not support that conclusion.
- The second possibility is to refuse to interact with the search results. Research from Rand Fishkin suggests that users click less throughout the search. Notably, however, navigational searches, as defined by Semrush, have never had a meaningful drop in reported click-through rates.
- A third possibility is macroeconomic pressure. If deteriorating economic conditions suppress demand, we can expect a corresponding decrease in search volume. The available data does not show that.
After considering these explanations, a different conclusion makes more sense. Consumers show little demand. They express the need in a different way.
AI-powered search provides a more emphatic explanation. Less than 12 months ago, AI Overview appeared in 57.2% of the marketing searches for a keyword query for one of our clients. By June 2026, that number had increased to 95.9%.
As AI systems answer questions, compare alternatives, and aggregate information before users conduct further searches, keyword searches take a smaller part of the decision-making process. The physical metric changes even if the underlying preferences don’t.
At its core, this is a problem of balance rather than a problem of demand.
Why this changes marketing decisions
This distinction is important because organizations allocate money based on metrics they trust. That dynamic also helps explain why performance marketing commands a disproportionate share of marketing investment.
If a metric appears to provide a straight line between costs and measurable results, capital naturally follows. If that metric is a less reliable representation of the underlying objective, the risk allocation is skewed.
If branded search underrepresents product demand, marketers run the risk of concluding that their products are weakening when, in fact, consumer behavior is shifting upward. The result is predictable: less investment in activities that create long-term choices.
The search measures revealed the intention. What has changed is that the expressed purpose is no longer the same as the basic need. AI has added an extra layer between consumer preference and visual search behavior. Search by name is still useful, but it should no longer be taken as definitive.
The implications go beyond search. Whenever a long-standing proxy underrepresents the object it was intended to measure, organizations run the risk of promoting the proxy instead of the objective itself. Marketing is unlikely to be different.
Rethinking how product demand is measured
If branded search becomes a less reliable measure of product demand, that doesn’t mean abandoning the metric. It is a rethinking of the framework through which it is interpreted.
Branded search should always be one indicator of product demand, but it may not be your direct proxy. Greater emphasis should shift to measures that directly capture preferences, including unassisted awareness, search content sharing, and product conversion rate.
If preferences come before a measurable objective, investment in a different area, original research, thought leadership, and continuous product development should not be considered close to performance. It’s the jobs that decide.
Changing capital allocation also changes when competitive advantage is created. As AI becomes the intermediary between consumers and information, authority extends beyond the channels directly controlled by the product.
Competitive advantage may depend less on generating more content than on being part of a reliable body of evidence that AI systems constantly reference. Increasingly, visibility may come more faithfully than distribution.
The search will likely continue to measure intent. Named search, however, seems to capture a smaller reflection of product demand than before. Organizations that recognize this shift early may simply measure product performance more accurately. They can fund more effectively because they understand the difference between the metric and the underlying need they want to represent.
