Associated Keywords Offer Missed Potential in Search Advertising

A recent study by Rothschild, Needell, Veverka, and Yom-Tov (2025) has identified a significant gap in how search advertising is currently targeted, focusing on the concept of "associated keywords." These are terms that relate to a product category but not a specific brand or product—such as "license" or "battery" in the context of vehicles—and are frequently used by consumers well in advance of an actual purchase.

Key Findings

The researchers analyzed journeys across three major product categories: vehicles, laptops, and phones, using a large set of Bing search queries. They found that associated keywords are searched 50 to 80 days before purchase events and are linked with high probability of a subsequent purchase in the category, yet come at significantly lower advertising costs than brand or main category keywords.

Models incorporating even a brief window—just five days—of user search history, as opposed to only a single recent query, yielded an 8-15 percentage point improvement in predicting conversion. Marginal gains diminished quickly after a week of search history was considered.

Moreover, the impact of advertising varied by stage of the customer journey. Brand-related keywords were most effective very close to the purchase moment but offered less lift earlier in the journey. In comparison, associated keywords provided meaningful lift early on, during what the researchers call the “commitment gap”—the period when a user is determined to buy within a category but has not selected a brand. This is the stage where current advertising spend is minimal despite the highest opportunity for influence.

Associated Keywords: Defined and Illustrated

The paper defines associated keywords as terms (1) relevant in the context of a purchase journey, (2) carrying high probability of a category purchase, (3) low probability of a specific brand purchase, (4) lower in advertising cost, and (5) present well before the purchase event. For vehicles, keywords like "license" or "battery" consistently indicated strong future purchase intent months before any brand choice was made.

Advertisers are currently heavily focused on late-funnel users—those who are already likely to convert and are searching for specific brand or product keywords. As a result, opportunities to influence undecided customers upstream, using less expensive associated keywords, are routinely missed.

Implications for AI-Powered Search

With the advancement of large language models (LLMs) and AI-driven conversational interfaces, platforms are gaining the capacity to assess a broader scope of user context than a single keyword or search query. This development suggests that future targeting could shift from “searchers of X” to “users in stage Y for product Z,” further increasing the practical utility of associated keyword strategies in programmatic advertising and digital marketing campaigns.

Limitations and Simulation Insights

While the study utilizes Bing data and three specific product categories, limiting broad applicability, its simulated models robustly show the value of associated keywords in digital advertising. Graphical simulations support the findings that prediction accuracy improves significantly with a modest amount of user search history and that the best advertising effect for associated keywords occurs in early stages of the buyer journey.

Takeaways for Advertisers

Marketers managing search campaigns are advised to identify associated keywords within their sector, evaluate their cost-effectiveness early in the consumer funnel, and leverage a short window of search history (five to seven days) for improved targeting—balancing gains in accuracy against privacy considerations.

Reference

Source: r-bloggers.com