What3words
What3words is a proprietary geocoding system founded in 2013 by Chris Sheldrick in London, United Kingdom, that divides the Earth's surface into a grid of approximately 57 trillion 3-by-3-meter squares, assigning each a unique, fixed combination of three words from a 40,000-word list to enable precise location identification without relying on traditional addresses or latitude-longitude coordinates.[1][2][3]
The system originated from Sheldrick's experiences in the music industry, where imprecise location data for events led to logistical challenges, prompting the development of a human-friendly alternative optimized for verbal communication and voice assistants.[4][5]
What3words has gained adoption among emergency services in over 100 organizations globally for faster incident location, logistics firms for deliveries in unaddressed areas, and navigation apps, with its mobile app downloaded more than 60 million times as of 2025.[6][7]
Despite these implementations, the system has drawn criticism for algorithmic flaws increasing the risk of address confusability—potentially leading to errors in safety-critical scenarios—and its closed-source model, which limits independent verification and has prompted legal actions against open-source alternatives.[8][9][10][11]
History
Founding and Initial Development
What3words was founded in 2013 in London, United Kingdom, by Chris Sheldrick, who serves as CEO, along with co-founders Mohan Ganesalingam, a mathematician, and Jack Waley-Cohen.[4][2] The company's inception addressed longstanding deficiencies in traditional addressing systems, particularly for precise location sharing in scenarios lacking standardized street addresses.[12] Sheldrick's idea stemmed from his prior role in the music industry, where he organized live events and repeatedly faced logistical failures, such as delivery trucks and equipment arriving at incorrect sites due to ambiguous or incomplete venue directions.[2][12][13] To resolve this, the team partitioned the globe into roughly 57 trillion squares measuring 3 meters by 3 meters each, assigning every square a unique sequence of three words selected from a 40,000-word lexicon to enable simple, memorable identification without reliance on coordinates or lengthy codes.[4][2] Ganesalingam developed the core algorithm for word encoding and assignment, ensuring algorithmic efficiency and global uniqueness while minimizing memorization demands.[2] The system launched publicly with a free mobile app and website on July 2, 2013, supported by an initial seed funding round of [$500,000](/page/500).[7][14] Early development emphasized computational mapping of Earth's surface for accuracy, including handling polar regions and ensuring the grid's proprietary nature.[4] By November 2013, an online API was introduced, facilitating developer integrations and marking the transition from prototype to scalable tool.[12] These steps laid the groundwork for broader adoption, though initial challenges included refining word lists for linguistic neutrality and computational demands of generating trillions of addresses.[4]Key Milestones and Global Expansion
What3words was founded in August 2013 by Chris Sheldrick, Jack Waley-Cohen, and Mohan Ganesalingam in London, United Kingdom, with the initial system dividing the Earth's surface into approximately 57 trillion 3-meter by 3-meter squares, each assigned a unique combination of three words.[4][3] The company secured its Series A funding round of $3.5 million in November 2015, led by Intel Capital, enabling early product refinement and initial partnerships.[14] A pivotal early milestone came in 2016 when Mongolia's national postal service, Mongol Post, adopted what3words as its official addressing system in June, marking the first national-scale implementation and addressing the country's challenges with traditional addresses in nomadic and rural areas.[15][16] This was followed by the first major emergency services integration in 2015 with Everbridge, expanding to broader adoption in public safety sectors globally.[4] By 2018, the system supported 14 languages, facilitating international usability.[4] Global expansion accelerated through automotive and logistics partnerships, including a 2020 integration with Mercedes-Benz vehicles for in-car navigation and a collaboration with HERE Technologies starting that year, which expanded in January 2024 to enhance location services across industries like delivery and mapping.[4][17] Additional Series B funding of $45 million in 2021 supported scaling to over 60 languages and operations in 193 countries.[4][1] Recent developments include significant U.S. market traction by early 2025, with integrations in e-commerce, ride-hailing, and emergency response, alongside partnerships with brands like Jaguar Land Rover, Subaru, and DHL.[18][1]Recent Developments and Innovations
In September 2025, what3words launched what3words Pro, a premium AI-powered upgrade to its mobile app and web mapping service, featuring AI Chat for handling location tasks via natural language processing. This innovation allows users to input voice or text commands to recognize, interpret, and organize what3words addresses, supporting applications like route generation, ride booking, and batch extraction of addresses from scanned files or documents.[19][20][21] The AI capabilities build on compatibility with large language models, enabling seamless integration into voice assistants for conversational navigation, such as querying precise locations like "the entrance to the stadium parking" and receiving the corresponding three-word address without manual lookup. This addresses limitations in traditional GPS systems by reducing errors in ambiguous verbal descriptions and enhancing hands-free usability in automotive and logistics contexts.[22] In July 2025, navigation provider Sygic integrated what3words into its GPS app, announced in advance of CES 2025, to enable users to share and navigate to three-word addresses for improved accuracy in urban and off-road scenarios.[23] By March 2025, what3words expanded its U.S. footprint through deepened partnerships in delivery and emergency response, facilitating precise location sharing where standard addresses fail, such as remote sites or complex facilities.[18]Technical Design
Core Grid System and Resolution
The core grid system of what3words partitions the Earth's surface into a fixed array of approximately 57 trillion discrete squares, each designed to measure precisely 3 meters by 3 meters.[1] This resolution was selected to enable location references granular enough for practical applications, such as pinpointing a specific building entrance, vehicle parking space, or small outdoor area, surpassing the precision of traditional street addresses or postal codes in many regions.[1] The grid encompasses the globe from latitudes roughly -85° to +85°, algorithmically adjusting square boundaries to maintain near-uniform 3m dimensions within centimeters of accuracy across most populated areas, thereby minimizing distortions from Earth's sphericity without relying on dynamic projections.[24] Unlike latitude-longitude grids, which produce varying physical sizes due to meridional convergence, the what3words system employs a proprietary algorithmic mapping to generate orthogonal, equal-area squares overlaid on a global coordinate framework.[25] This approach ensures that each square remains a consistent 9 square meters in area, facilitating straightforward encoding and decoding for human-readable addresses while covering approximately 510 million square kilometers of land and ocean surfaces.[26] The total of 57 trillion squares arises from the mathematical necessity to tile the planet at this scale, calculated as the surface area divided by the 9 m² per square, with edge cases at polar regions handled by extending the grid pattern to avoid gaps or overlaps.[1] Resolution limitations stem from the fixed 3m scale, which does not adapt to sub-meter needs like indoor navigation or micro-precision surveying, nor does it inherently account for vertical positioning such as building floors.[26] Empirical tests and geospatial analyses confirm the grid's consistency, with deviations under 2% in square sizing even at higher latitudes, validating its utility for global standardization over variable traditional grids.[24] This design prioritizes memorability and error resistance in location sharing, as the uniform resolution supports the subsequent word-assignment process without requiring user knowledge of geographic projections.[27]Word Selection and Encoding Process
What3words employs a fixed dictionary of approximately 40,000 unique words per supported language to encode locations into three-word combinations.[8] The words are selected from a corpus filtered for frequency (appearing at least three times), length (at least four characters), and commonality, prioritizing shorter, easier-to-spell and pronounce terms while excluding rare variants like British spellings in favor of American ones (e.g., "favor" over "favour").[8][26] The selection algorithm sorts words by length, distinctiveness, frequency, and phonetic simplicity to facilitate memorability and reduce transcription errors, with the dictionary designed to cover over 57 trillion 3m × 3m grid squares globally.[26][28] For encoding, geographic coordinates are mapped to a unique integer index representing the grid square's position within larger cells (approximately 4,638 m × 2,885 m, yielding 37,324,800 cells worldwide).[29] This index n incorporates offsets q tailored to high-density areas and is computed as n = q + 1,546 × x + y, where x and y are intra-cell coordinates.[29] The index is then shuffled via a linear congruential transformation, such as m = (9,401,181,443 × n) mod K³ (where K varies by density band, e.g., 2,500 for urban "band 0"), to pseudorandomly distribute word triples and separate similar combinations by thousands of kilometers, minimizing error risks (e.g., 1 in 2.5 million chance of ambiguity in the UK).[29][28] Finally, m is factored into three indices (i = floor(m / K²), j = floor((m / K) mod K), k = m mod K), each indexing the sorted word list to yield the ordered triple (e.g., indices 1,940; 910; 2,140 map to "offers.trail.medium").[29][8] In multilingual implementations, separate dictionaries are created per language, with an algorithm assigning simpler, more frequent words to regions where that language predominates to enhance usability.[28] The entire mapping is proprietary and fixed, ensuring unchanging addresses despite urban changes, though reverse-engineering reveals the shuffling prevents predictable patterns.[29] Bands adjust dictionary subsets (e.g., first 2,500 words for cities) to optimize for context-specific simplicity.[29]Ambiguity Handling and Error Tolerance
The What3words system employs a curated dictionary of approximately 40,000 words selected to minimize phonetic and orthographic ambiguities, such as avoiding excessive homophones, though residual confusions persist (e.g., "sense" and "cents").[8] The encoding algorithm uses linear congruential transformations to assign word triples to grid squares, prioritizing the separation of similar combinations—such as placing ///table.chair.lamp in Australia and ///tables.chair.lamp in the United States—to reduce the risk of local confusion from minor variations.[30] [8] Official claims assert a low ambiguity rate, with similar-sounding codes having a 1 in 2.5 million probability of overlap in regions like the United Kingdom.[30] Error tolerance relies on exact matching rather than fuzzy algorithms, ensuring that input errors—such as a single misspelling or word substitution—map to locations typically hundreds of kilometers away, prompting users to detect and correct via visual discrepancy on the app's map interface.[31] The AutoSuggest feature mitigates partial errors by offering up to three alternative valid addresses, displaying their distances from the user's approximate location or contextual landmarks (e.g., nearest city) to facilitate selection.[31] [30] In emergency contexts, protocols recommend verbal confirmation of words alongside GPS data to counter transcription risks from mispronunciation or accents.[30] Critiques highlight vulnerabilities, including the absence of robust error-correcting codes, leading to potential failures in high-stakes scenarios; for instance, analyses model transmission errors (typos, homophones) and find that 20-25% of addresses have more than three confusable alternatives, with some erroneous mappings occurring within 10 kilometers due to algorithmic clustering.[8] Empirical reports from rescuers document cases where input errors directed teams to incorrect sites, attributing issues to the system's intolerance for variations without built-in Levenshtein-distance-like corrections.[32] [8] These limitations contrast with official endorsements from 100% of surveyed UK emergency services citing reliability when properly verified, underscoring dependence on user diligence over inherent fault tolerance.[30]Business Model
Revenue Streams and Partnerships
What3words generates revenue primarily through licensing its API to businesses, enabling integration of the three-word addressing system into products and services for sectors such as automotive, logistics, e-commerce, and emergency response.[33][34] This B2B model includes tiered pricing for API usage, with plans starting at £7.99 per month for 1,000 API calls in the Basic tier and scaling to £235 per month for 75,000 calls in the Premium tier, while enterprise solutions are offered via annual licenses or per-device fees for applications like vehicles or drones.[35][36] Additionally, the company offers premium Pro access for individual users and customized business plans, though the core consumer app remains free to drive adoption and data collection.[37] The firm reported £1 million in revenue for 2023, marking a 28% increase from the prior year, largely from these licensing agreements amid growing integrations.[7] Non-profits and charities qualify for free API access up to 75,000 monthly conversions between words and coordinates, supporting humanitarian uses without direct revenue contribution.[38] Partnerships form a key pillar, with thousands of businesses worldwide integrating What3words to enhance location accuracy in navigation, delivery, and recovery operations.[39] Notable collaborations include an expanded global agreement with HERE Technologies for hosted API services in mapping applications, and integrations with logistics firms, courier services, and driving apps like Revved for precise addressing in e-commerce and transport.[40][41] Emergency sector ties, such as with Norfolk Fire & Rescue Service for hydrant location, further embed the system in public safety without always involving fees, prioritizing societal impact alongside commercial growth.[41] These alliances, spanning automotive, e-commerce, and recovery, leverage What3words' proprietary grid to differentiate partner offerings, though revenue dependency on API uptake exposes the model to competition from open alternatives.[39][34]Funding, Financial Performance, and Sustainability
What3words, founded in 2013, secured initial seed funding of US$500,000 in November 2013 to support early development. Subsequent rounds included a Series B investment of $8.5 million in June 2016 led by investors such as Intel Capital, Aramex International, and Force Over Mass.[42] The company has raised approximately £60 million in total equity funding across multiple rounds, including equity crowdfunding campaigns on platforms like Crowdcube, with participation from strategic investors such as Mercedes-Benz and Sony's venture arm.[7][43] A 2020 funding round valued the company at £250 million.[7] Financial performance has shown revenue growth amid persistent losses. In 2022, annual revenue was approximately £0.8 million, increasing to over £1 million in 2023 (a 28% year-on-year rise).[44][45] Revenue doubled to £2.2 million in 2024, driven by expanded partnerships and licensing.[46] Net losses narrowed progressively: £31.5 million in 2022, £16.5 million in 2023 (halved from prior year), and £10.6 million in 2024, supported by a 25% reduction in administrative expenses and workforce reductions.[45][47] Cumulative net losses exceed £119 million since inception, reflecting heavy investment in global expansion and technology maintenance.[48] Business sustainability remains challenged by unprofitability after over a decade of operations, with ongoing reliance on investor capital to fund scaling efforts.[49] However, recent cost controls, revenue acceleration from commercial integrations (e.g., automotive and logistics), and loss reductions signal potential viability, though achievement of profitability has not yet materialized.[46][49] The proprietary model's barriers to free alternatives aid retention of paid enterprise clients, but high operational costs for a non-revenue-generating core grid system pose risks to long-term independence without further funding or break-even.[7]Adoption and Applications
Integration in Emergency Services
What3words enables emergency callers to communicate precise locations using three-word addresses, which dispatchers convert to GPS coordinates for rapid response, particularly benefiting areas lacking traditional addresses or where verbal descriptions are imprecise.[50] In Mongolia, the National Emergency Management Agency (NEMA) integrated the system to locate incidents in vast rural and nomadic regions, where over 636,000 users had adopted the app by December 2024, allowing herdsmen and citizens to report exact positions without address ambiguity.[51][52] The United Kingdom saw early and extensive adoption, with 59 emergency services—including police, fire, and ambulance—incorporating what3words by September 2019, rising to over 85% coverage among British services thereafter; this integration connects directly to control room systems, reducing response times in urban and remote incidents like mountain rescues.[53] In the United States, more than 250 public safety agencies, supported by partnerships like RapidSOS, use the system as of February 2025, with documented cases from 2021 where three-word addresses enabled first responders to reach victims in under critical time thresholds, such as vehicle crashes and medical emergencies.[54][55] Other integrations include Da Nang, Vietnam, which embedded what3words into its public safety app on August 28, 2023, for citywide emergency dispatching; Brampton Fire and Emergency Services in Canada; and select agencies in Germany and Ireland, often via mobile apps or Advanced Mobile Location (AML) protocols that automatically share three-word data from smartphones.[56][57] These implementations prioritize offline functionality and multilingual support, with emergency operators trained to verify addresses against potential confusions, though reliance on app downloads or user familiarity remains a deployment factor.[50]Commercial and Navigation Uses
What3words has been integrated into various commercial logistics and delivery platforms to enhance address precision for last-mile operations. For instance, companies such as Metapack, Zenstores, and ShipTheory have adopted the system to streamline e-commerce deliveries by allowing customers to provide exact 3-word addresses, reducing failed attempts and improving efficiency as of December 2022.[58] In freight management, Merzario incorporated the what3words API into its transportation system in March 2024, enabling more accurate tracking and routing for international shipments.[41] Similarly, Optimize integrated what3words in March 2025 to support precise route planning in logistics software, minimizing navigation errors in complex environments.[59] In navigation applications, what3words facilitates direct input of 3-word addresses into vehicle infotainment and GPS systems. Jaguar Land Rover introduced compatibility in its Pivi Pro system on June 9, 2022, allowing drivers to enter addresses via the navigation bar for enhanced accuracy over traditional street-based routing.[60] HERE Technologies expanded its partnership with what3words, initially established in 2020, to provide automotive manufacturers with precise in-car navigation features.[17] IVECO announced integration into its light and heavy commercial vehicles on September 16, 2024, addressing limitations of street addresses in sat-nav systems for professional drivers.[61] Consumer-facing apps like Sygic GPS Navigation added support announced on July 1, 2025, enabling users to navigate to any 3m² location using three words.[23] These integrations underscore what3words' role in sectors requiring sub-meter accuracy, such as urban deliveries and off-road navigation, where ambiguous postal addresses often lead to delays or errors.[39] Adoption has grown in the US market, with multiple logistics and navigation tools incorporating the system by early 2025 to handle location challenges in underserved areas.[62]Global Reach and Specific Case Studies
What3words provides unique three-word addresses for every 3-meter square on Earth's surface, enabling its use across 193 countries with interfaces in over 60 languages.[1] Emergency services worldwide have integrated it to improve location accuracy during calls, particularly in regions lacking precise addressing systems; for instance, Mongolia's National Emergency Management Agency adopted it in March 2020 to expedite responses in vast rural areas where traditional coordinates or descriptions often fail.[52] In the United Kingdom, over 100 emergency services, including police, fire, and ambulance units, accept what3words addresses as of May 2024, facilitating faster dispatch in urban and remote incidents.[50] A prominent case study is Mongolia, where what3words functions as a de facto national addressing standard due to the country's sparse population and limited road infrastructure. Mongol Post implemented it in June 2016 for nationwide deliveries, addressing gaps in conventional postal systems, and by October 2023, the government expanded integrations to e-government portals, cultural sites, tourism, and additional emergency protocols, enabling citizens to reference locations without GPS dependency.[63][64] This rollout has supported logistics in nomadic communities and disaster response, with the technology embedded in public apps for reporting hazards.[65] In navigation, what3words has achieved global scalability through automotive and mapping partnerships. Mercedes-Benz pioneered in-car integration in February 2019, allowing drivers of compatible models to input three-word addresses via voice commands for precise routing, now available in vehicles sold internationally.[66] Similarly, TomTom incorporated it into its GPS systems starting in the second half of 2018, extending support to consumer devices and enterprise fleets across Europe, Asia, and beyond, which has aided applications in dense urban environments like Hong Kong alleys and informal settlements in India.[67][68] These integrations demonstrate what3words' utility in commercial contexts, such as last-mile deliveries via partners like DHL and DPD, where it reduces errors in address verification by up to 42% in tested operations.[1]Criticisms and Limitations
Technical Reliability Issues
What3words' encoding system, which maps 3-meter by 3-meter squares to unique three-word combinations from a 40,000-word dictionary in 50 languages, has been criticized for insufficient error tolerance in transmission and input, leading to potential misdirection in critical scenarios. Independent analysis has identified intrinsic algorithmic flaws, including a higher-than-claimed probability of confusing one address with another due to homophones, near-homophones, and single-character typos generating valid but erroneous locations. For instance, security researcher Graham Gibbons demonstrated that design choices allow single-letter errors to resolve to plausible nearby squares, with a 1-in-10 chance of such errors leading to confusion with adjacent or distant sites, rendering the system unsuitable for safety-critical applications without additional verification.[10] Real-world deployments have exposed reliability gaps, particularly in emergency response. In the UK, mountain rescue teams reported dozens of instances in 2021 where what3words addresses provided by callers resulted in rescuers being dispatched to incorrect locations, often due to spelling variations or auditory miscommunications over phone lines. Similarly, emergency services in multiple regions have noted frequent spelling errors—exacerbated by the system's lack of inherent checksums or redundancy beyond basic Levenshtein distance checks—causing delays or misallocations, with one analysis estimating inaccuracies in up to 30% of transcribed cases without app-based validation.[32][69][70] The system's reliance on device GPS for address generation introduces further vulnerabilities tied to signal quality and environmental factors. GPS inaccuracies exceeding 3 meters—common indoors, in urban canyons, or under tree cover—can assign words to the wrong square, with the app's location display deviating from actual position based on signal strength, yet without explicit warnings of this imprecision in all interfaces. While what3words claims built-in error detection for up to 50% of single-word mistakes via algorithmic suggestions, empirical critiques argue this underestimates compounded risks in verbal or low-literacy contexts, where the absence of phonetic encoding amplifies failure rates compared to numeric coordinates with error-detecting formats like Maidenhead locators.[71][8]Proprietary Nature and Accessibility Barriers
What3words maintains a fully proprietary model for its geocoding system, with the algorithm for generating and mapping three-word addresses protected as intellectual property and not publicly disclosed.[72] The company holds patents on the core technology and trademarks on all three-word combinations, preventing third-party replication without licensing.[73] Access to essential functions, such as converting words to coordinates or vice versa, is restricted to the company's APIs, which require adherence to their terms and often involve paid subscriptions for commercial or high-volume use.[38] This closed-source structure limits opportunities for independent auditing or community-driven improvements, as the word list and mapping data remain under exclusive corporate control.[8] The proprietary framework creates significant accessibility barriers, particularly for non-commercial or resource-constrained users. Resolution of addresses depends on querying what3words' servers, necessitating internet connectivity and compatible devices, which excludes offline scenarios or areas with poor network coverage common in remote or emergency contexts.[72] Developers integrating the system into applications must obtain API keys and comply with usage limits, with free tiers restricted to low-volume personal use, while enterprise features demand contractual fees that can deter small organizations, nonprofits, or open-source initiatives.[74] In 2021, what3words issued legal threats to a security researcher offering an open-source alternative implementation, citing trademark and licensing violations, which underscored efforts to suppress competing, freely accessible tools.[11] These barriers extend to systemic risks, including vendor lock-in for adopters like emergency services, where shifting to alternatives would require retraining and data migration without backward compatibility due to the non-standard, company-specific format.[75] Unlike open alternatives such as Google's Plus Codes, which provide freely downloadable libraries and no dependency on a single provider, what3words' model fosters reliance on the company's ongoing availability and policy decisions, potentially compromising long-term equity in global addressing access.[75] Critics argue this centralization prioritizes revenue over universal utility, as proprietary controls hinder widespread, cost-free adoption in developing regions or low-tech environments.[73]Legal Challenges and Ethical Concerns
In April 2021, What3words issued a legal threat to security researcher Ryan Toponce for developing and offering to share an open-source alternative called WhatFreeWords, which aimed to replicate the system's functionality without proprietary restrictions.[11] The company's legal correspondence claimed that the project infringed on its intellectual property, including the copyrighted wordlist and algorithmic mappings, and demanded disclosure of any individuals who had accessed the code, despite Toponce asserting that his implementation independently generated a similar encoding without copying protected elements.[76] What3words holds multiple patents on its core addressing algorithm, such as those covering multi-word resolution to unique locations, which underpin its enforcement actions against perceived IP violations.[77] No formal lawsuit ensued from this incident, but it highlighted the company's aggressive stance on protecting its proprietary technology, potentially discouraging open-source alternatives that could enhance system resilience or accessibility.[78] Ethically, What3words' closed-source model raises concerns about dependency on a single for-profit entity for critical applications like emergency response, creating risks of service disruption if the company alters terms, faces financial instability, or ceases operations.[74] This proprietary control contrasts with open standards in geolocation, limiting independent verification of the algorithm's robustness and fostering vendor lock-in for adopters, including governments and public safety agencies.[79] Data privacy issues emerge from the system's requirement to query What3words' servers for location resolution, potentially exposing user coordinates to centralized logging, though the company maintains it does not retain precise location data without consent.[80] Additionally, the English-centric wordlist has drawn criticism for cultural insensitivity in non-English-speaking regions, where word meanings or profanities may vary, complicating global ethical deployment without localized adaptations.[81] These factors underscore broader ethical tensions between innovation and public reliance on private infrastructure for essential services.Overall Assessment
Empirical Evidence of Effectiveness
A 2023 independent analysis published in PLOS ONE examined the What3Words algorithm's reliability by simulating common input errors such as homophones, single-character substitutions, omissions, and adjacent-key typos on keyboards. It found that approximately 33% of word triples exhibit no confusions under these conditions, while two-thirds have at least one confusable pair globally, with a mean of 0.73 confusions per word.[8] Around 20-25% of addresses have more than three confusions, potentially overwhelming the system's Autosuggest feature, which limits suggestions to three.[8] Locally, confusable pairs frequently occur within distances of about 10 km—far closer than the company's claim of a 1 in 2.5 million probability for nearby confusions—due to algorithmic patterns in word assignments relative to grid positions.[8] These findings contradict What3Words' assertions of minimal error risk, implying elevated chances of misidentification in time-sensitive scenarios like emergency responses, where selecting the wrong nearby location could delay aid or lead responders astray.[8] A 2023 case study on an app integrating What3Words for first responders compared its performance to GPS and traditional addressing schemes, evaluating location accuracy in reaching emergency sites. The analysis indicated that alternate addressing systems like What3Words enhance precision over vague verbal descriptions, potentially reducing response times by providing verifiable 3-meter grid references, though it did not quantify error rates or conduct large-scale field tests.[82] Real-world data on success rates remains anecdotal and company-reported, with limited independent verification. For instance, in South Africa, Vodacom integrated What3Words, resulting in 14,620 emergency callouts using the system over one year as of April 2021, including cases credited with saving lives, such as a COVID-19 patient location.[83] However, no public datasets track overall success or failure rates across integrations in over 100 emergency services worldwide, and rescuer feedback has raised concerns about practical confusions, such as ambiguous word interpretations under stress.[32] What3Words claims that 56% of adopting services handle daily location-challenged calls more effectively via the system, but this lacks third-party auditing.[84] Broader empirical scrutiny is scarce, with no peer-reviewed studies measuring net improvements in response times or survival outcomes attributable to What3Words versus alternatives like Plus Codes or latitude/longitude. The available evidence thus supports utility for memorable location sharing in low-stakes contexts but underscores algorithmic vulnerabilities that could compromise effectiveness in high-reliability demands, necessitating further independent field evaluations.[8]Comparisons to Alternative Systems
What3words employs a proprietary algorithm to assign unique combinations of three words from a fixed dictionary of 40,000 English terms to each 3-by-3-meter square on Earth's surface, aiming for verbal communicability over numerical systems like latitude and longitude. Traditional latitude-longitude coordinates, in contrast, offer arbitrary precision (e.g., to six decimal places for sub-meter accuracy) without fixed grid sizes, relying on decimal degrees readable globally but prone to transcription errors in verbal exchanges due to numerical complexity.[85] Google's Plus Codes (Open Location Code), an open-source system released in 2015, generates short alphanumeric strings (e.g., "7J7P+2M" for a Berlin location) based on a hierarchical grid starting at 20-by-20-meter cells and refining to 3-by-3 meters with additional characters, enabling free, customizable implementation without licensing fees.[86] Unlike What3words' flat structure, Plus Codes support hierarchical shortening—local codes relative to a known area code reduce length while maintaining context for larger regions, facilitating easier navigation from broad to precise locations.[87]| Aspect | What3words | Plus Codes | Latitude/Longitude |
|---|---|---|---|
| Precision | Fixed 3x3 m squares | Configurable; default ~14 m, extendable to 3x3 m[87] | Arbitrary (e.g., 0.000001° ~0.1 m)[85] |
| Format | 3 pronounceable words (e.g., "filled.count.soap")[88] | Alphanumeric (e.g., 7J7P+2M)[86] | Decimal degrees (e.g., 52.5200° N, 13.4050° E)[85] |
| Openness | Proprietary; API integration requires licensing[86] | Open-source, no fees[87] | Universal standard, no proprietary restrictions[85] |
| Usability (Verbal/Memory) | High; words are memorable and distinct-sounding, reducing errors in phone/radio reports per developer claims, though empirical tests show confusion risks from homophones (e.g., 1 in 2,500 chance of adjacent square mix-up)[88][8] | Moderate; shorter than full lat/long but less intuitive than words, with hierarchy aiding recall in context[87] | Low; numbers hard to remember/transmit accurately without aids[85] |
| Hierarchy | None; each code independent, no inherent zooming[87] | Yes; area + local codes for scalable precision[86] | Partial; truncating digits coarsens uniformly but lacks semantic grouping[85] |
| Adoption/Cost | Licensed for commercial use; adopted by some emergency services (e.g., UK, Mongolia as of 2019) but criticized for vendor lock-in[86] | Free; integrated in Google Maps, used in areas without addresses like India | Free universal; standard in GPS devices, no app dependency for decoding[85] |