Devil Take the Hindmost (or How Yahoo Lost the Internet Search Market)

Athena Intelligence
4 min readFeb 6, 2024

Lessons for Reinsurers from Google and Catastrophic Wildfire Risk Assessment

In the ever-evolving landscape risk management, lessons from other rapidly changing industries can be found. Specifically, parallels between the seismic shift from Yahoo’s directory-based approach to Google’s superior search algorithm, which caused a rapid shift between one widely used approach and a new, radically different approach to Internet search. In the context of assessing escalating threat of wildfires, catastrophic reinsurers assessing wildfire risks are about to grapple a similar change.

In the 1997 and 1998, the Internet was still new to business and the public. Yahoo! dominated search with directory-based approach. On September 4, 1998 a new search algorithm was introduced by Google. By the end of 1999, about 2.5 million Internet users had tried Google’s search engine, representing less than 1.0% of worldwide users. For context, AOL had 26.7 million subscribers in 1999.

Photo from Unsplash & Matt Howard

Nevertheless, a superior search engine gathered users. By the end of 2000, Google had an estimated 28 million users, and 24 months later (2002), it had 179 million users, or 28.4% of Internet users. Using the same “raw data” of existing websites, how did Google offer such a superior experience than Internet users rapidly changed their patterns.

Today, the property insurance industry is grappling with the intricacies of prefire wildfire risk. The existing models for US properties are based on the WHP map, the Wildfire Hazard Potential maps which is published every 5 or 6 years by the US Forestry Service. This represents the best thinking of academics who specialize in wildfire risk.

The WHP is static and overgeneralized, which prompts our comparison to the human curated indexing of Yahoo! in the 1990’s. Google Search, from its beginning, dynamically assessed links, content, and user behavior, making it dynamic and very specific to the user’s intent and search goal. By prioritizing relevance, Google’s revolutionary algorithm, based on profiles, rapidly came to dominate an indexing approach.

Devil Take the Hindmost

There is an interconnected narrative about how quickly people switched to better insights, and how quickly businesses will switch to a tool to higher profitability. For property insurers and property cat reinsures this describes the speed at which better insights into wildfire risk will change underwriting. The early adopters will have better insight into pricing, risk avoidance and capital efficiency. For the next few years, in wildfire risk, the asymmetrical information will result in higher profits for Athena’s customers.

Devil take the hindmost, is an idiom that predates the formation of Lloyds of London, and expresses a sentiment those earliest names would have understood. A sense of competition to write the best risks, leaving the more challenging underwriting to those who are slower, less informed to suffer the consequences of their lack of insight.

New or better information, whether it arrives by carrier pigeon or geospatial, machine language generated algorithms is the key to better risk management.

Google’s profiling algorithm provided users with the most pertinent and up-to-date information. This focus on precision transformed user experience, setting a benchmark for a new approach: algorithmic strategies that prioritize dynamic adaptability over static indexing.

Reinsurers today, looking at wildfire risk, face a seismic shift that runs parallel to this. Athena creates conditional profiles (30 sq meter pixels) with unique alphanumeric identifiers, which describe all of the characteristics for miles around, which relate to that location’s ability to host a fire. The profiling algorithm, then looks at historic fires and those conditional profiles and historic fires to create probabilities.

Each pixel thus has a conditional profile (which changes with vegetation growth, bioregion, weather, burn scars or wildfire mitigation efforts) and a multiyear history and a forward risk projection. This approach allows dynamic risk insights, even as the frequency and intensity of wildfires are escalating and challenging traditional risk assessment models, based on static data, is failing.

Much like Yahoo’s indexed approach, traditional catastrophic risk assessment models suffer from limitations in adapting to contemporary risks. Wildfires, influenced by climate change and complex environmental factors, demand a level of relevance and precision that traditional models struggle to provide.

Reinsurers can draw inspiration from Google’s success with profiling algorithms and look for dynamic, adaptable models in prefire risk assessment. The ability to dynamically assess environmental data, incorporating real-time information on climate, vegetation, and human factors, is crucial in the face of evolving wildfire risks.

The transition from Yahoo’s indexed approach to Google’s profiles which resulted in a superior search algorithm serves as a metaphorical roadmap for reinsurers navigating the challenges of catastrophic wildfire risk assessment.

By embracing dynamic adaptability, prioritizing relevant current data, and applying algorithmic thinking to risk assessment, reinsurers can improve capital efficiency and emerge as proactive stewards in the face of an evolving risk landscape. In the nexus of technology lessons and risk management imperatives, the echo of Google’s triumph resonates as reinsurers endeavor to craft a resilient future in the wildfire era.

About Athena Intelligence

Athena Intelligence, headquartered in Sacramento, is a data vendor with a proprietary, geospatial, conditional profiling tool that integrates vast amounts of disaggregated wildfire and environmental data. This tool generates spatial intelligence, providing a digital fingerprint of wildfire risk. The company’s innovative solutions empower industries such as power, insurance, and financial services to make informed decisions in risk assessment, portfolio optimization, and loss cost probabilities. Learn more at

Follow us here, or on Linked In (Group Athena Voice of the Acre) or reach out to us at‍



Athena Intelligence

Athena Intelligence weaves vast amounts of disaggregated environmental data. Drop us a line (, or visit