SWOT
Strength
-
- Tea
- Milk
- Coffee
- Tea
- Milk
Weakness
- Palantir faces challenges associated with potential pricing deleverage and converting bookings to cash
- Crowded competitive landscape: More than 100 big and small companies currently offer big data and analytics software products and services. Our survey indicates that the large software tech vendors (including Microsoft, IBM, Oracle, and SAP) are some of the most commonly used IT vendors for BI & Analytics. What makes us marginally more cautious about Palantir’s near-term outlook? Over the past few years, Palantir’s competitive headwinds from large tech companies have strengthened. These headwinds, coupled with a greater focus on AI/ML initiatives are driving BI/analytics products.
- Unclear long-term profitability potential: We are encouraged by recent media reports that Palantir remains on track to become profitable by the end of 2017. However, the long-term cost structure of comparable big data analytics and business intelligence companies remains unproven and debatable. Over the longer-term, as Palantir shifts from a services/consulting-driven sales cycle to a traditional SaaS sales cycle, we expect longer sales cycles, which will likely lead to a structurally higher spend compared to today’s levels.
- Opaque pricing strategy: Unlike some of its peers, Palantir does not publicly disclose its pricing strategy. This lack of transparency leads to an unclear understanding of its offerings and how they compare with the offerings of its peers. Media reports regarding Palantir’s price sheets indicate that Palantir’s offerings may be more expensive than its peers, a situation that could lead to a lack of pricing leverage or to increasing customer acquisition costs.
- Data privacy and security challenges: Since Palantir is involved with government agencies that solve fraud and crime, Palantir’s products interface with a significant amount of sensitive information, both public and private. As such, Palantir has an additional responsibility to be more transparent about how it uses public data and protects public information. Recently, reports of information leaks and hacks across enterprises have put a greater spotlight on companies like Palantir.
- Potential pricing deleverage driven by customer churn: Media reports suggest that Palantir’s pricing is likely leading to elevated customer churn, particularly among those customers who may not need a highly sophisticated analytics platform. While it’s difficult to find publicly available Palantir pricing data, we estimate that a typical Palantir Gotham or Metropolis license, per server core per month, could range between $3,000 to $10,000 per month, with extra charges for support, maintenance, and training.
- Challenge of converting bookings into net revenues: In 2015, Palantir’s customer bookings exceeded an estimated $1.7 billion. However, it is unclear how much of that money translated into revenues for the company. We believe that typical Palantir customers sign three-year contracts, implying a 33% current fiscal year book-to-cash conversion. Investors would view a growing cash conversion ratio as a sign of long-term healthy performance.
- Services revenue weighing on Palantir’s gross margins: Software services tend to have lower margins compared to products and licenses, largely due to the variable personnel component. We believe that Palantir bundles consulting, support, maintenance, training, and other such on-site services with its customer contracts. Over the longer term, we expect the company to “productize” its services, as it has since starting a clear push into enterprises.
- Rising private shareholder activism potential: Palantir’s early investors have been fairly patient for a liquidity event. It will be almost ten years since Palantir first raised institutional capital. Over the past year or so, the company has been embroiled in a two-way legal battle over information access with an early investor (KT4 Partners). Such legal battles highlight the contentious relationships between startup investors and companies that want to maintain tight control of stockholders. As evidenced with recent events at Uber, we would closely monitor the ongoing lawsuit with KT4 partners, and any such other related events.
- Morgan Stanley significantly marks down Palantir's valuation, following Fidelity and BlackRock.
- Potential pricing deleverage driven by customer churn . Media reports suggest that Palantir’s pricing is likely leading to elevated customer churn, particularly among customers who may not need a highly sophisticated analytics platform. While it’s difficult to find publicly available Palantir pricing data, we estimate that a typical Palantir Gotham or Metropolis license, per server core per month, ranges between $3,000 to $10,000, with extra charges for support, maintenance, and training.
- Challenge of converting bookings into net revenues . In 2015, Palantir’s customer bookings exceeded an estimated $1.7 billion. However, it is unclear how much of that money translated into revenues for the company. We believe that typical Palantir customers sign three- year contracts, implying a 33 percent current fiscal year book-to-cash conversion. Investors view a growing cash conversion ratio as a sign of long-term healthy performance.
Opportunities
- IT professionals favor BI and analytics offerings from large tech vendors. A majority of our survey respondents identified Microsoft, IBM, Oracle, and SAP as the vendors currently used at their companies. These selections highlight the competitive landscape that Palantir faces, despite its innovative and unique product offerings.
- There is a positive outlook for BI & Analytics spend over the next twelve months. 89% of the surveyed IT professionals indicated that they were planning to increase their spend on BI & Analytics software over the next twelve months.
- Predictive analytics and machine learning integration will be the key areas of focus over the next five years. 49% of the companies using BI and analytics tools identified predictive analytics and integration with machine learning as the major areas of focus over the next five years. These advanced analytics help enterprises to analyze the volumes of data they collect and gain actionable insights from it
- The majority of companies prefer to use consulting services along with BI & Analytics products. 83% of survey respondents indicated that they used professional services along with data analytics products, further reinforcing Palantir’s go-to-market strategy.
- Unstructured data is already widely used in enterprises. 47% of survey respondents indicated that they used unstructured data “all the time” while performing data analytics. Another 43% of respondents said that they used such data “occasionally.”
- Large and growing market opportunity: The demand for big data analytics, visualization, and business intelligence software is estimated to grow from $150 billion in 2017 to $210 billion in 2020, translating to a robust 12% growth (three-year CAGR). While Palantir’s serviceable market opportunity remains subject to debate, our analysis of key market trends and Palantir’s product evolution suggests that the company could target an annual spend of $40–50 billion, primarily in data visualization, predictive analytics, and fraud analytics, implying a fairly robust and sustainable growth trajectory.
- Significant secular trends: Key secular trends, such as mobility, data growth, and movement to the cloud, have increased data analytics needs, triggering the rise of several new analytics vendors. These trends have led to an exponential increase in unstructured data from multiple sources. Our survey of IT decision makers suggests that over 90% of companies today use unstructured data. Palantir, with its advanced data integration technologies, stands to benefit from these secular trends.
- Beneficiary of growing IT spend on cybersecurity, anti-fraud measures, and counter-terrorism: Industry experts predict a continued rise in the IT share of global military and defense budgets. We believe Palantir occupies a unique place within the data analytics domain and benefits from increasing awareness about the global war against crime and fraud. The company’s industry-leading analytics platform makes it the go-to vendor for governments and businesses across the globe.
- Continued significant growth potential in government IT spend: U.S. federal, state, and local IT spending is expected to exceed $200 billion in 2017, with roughly 30–35% of this amount dedicated to IT in defense, public safety, and justice departments. Palantir’s predominant focus on these verticals, coupled with its custom offerings for specific government use cases, gives us greater conviction that Palantir has a pathway to long-term growth within the government.
- Unique, user-friendly customer proposition: Palantir’s first-generation product roadmap was largely driven by government use cases, emphasizing ease of use and actionable dashboards for non-technical end users. We believe that Palantir’s in-house intellectual property, which was built upon early government contracts, has allowed the company to reduce friction around big data analytics. Enterprise companies can now focus on drawing actionable business conclusions rather than implementing data cleanup, parsing, and other input-related activities.
- High-profile customers increase sales and marketing leverage: Palantir’s technology has been associated with prominent global events in disaster relief (Hurricane Sandy), fraud detection (the Bernie Madoff Ponzi scheme), and anti-terrorist activities (the capture of Osama bin Laden). While its peer SaaS companies spend up to 50% of their net revenues on sales and marketing, we believe Palantir benefits from the media coverage of highly visible events as well as customer stories. This publicity likely leads to significant savings on sales and marketing expenses, taking the company a step closer to profitability by the end of 2017
- Attractive revenue mix-shift from government to commercial: We believe Palantir has diversified its customer base by growing its commercial revenues from an estimated 0% of revenue in 2008 to roughly between 50–60% by 2017. Investors will view this shift positively, as it will increase reporting transparency and mitigate investor concerns about Palantir’s reported conflict with certain government agencies.
- Potential acquisition target: With an estimated $3.5 billion in gross bookings this year, industry-leading technology in the data analytics space, and an arguable moat around its government relations, Palantir presents an attractive acquisition target for established enterprises wanting to expand into the predictive analytics, data integration, and data visualization space. This attractiveness puts a reasonable floor under Palantir’s valuation and will help investors frame a downside risk scenario in their valuation frameworks.
- experts expect content and predictive analytics to take the lead in helping businesses integrate multiple data sources and provide actionable insights.
- One of the sub-segments, business intelligence and analytics, has been growing rapidly, thanks to the enormous amounts of data that enterprises collect and need to interpret. Over the last few years, reporting and query tools have been in high demand. However, moving forward, experts expect content and predictive analytics to take the lead in helping businesses integrate multiple data sources and provide actionable insights.
- Shift from analysts to business users: The primary cause for user growth in this market has been the shift in the usage segment. Usage is shifting from analytical users to business users, increasing the user base by more than 3.5 times. These additional users need more complex and cross-functional analyses. Business users have become the main driver for innovation, as they seek self-service without IT intervention and innovative collaborative tools to share insights across the organization. Organizations are moving from basic data reporting to more advanced analytics and insight-based tools to help make future decisions.
- Decentralization of BI model: Traditional BI platforms have been led by the IT teams within an organization, with the BI analytics performed by a data scientist but the rest of the activities (such as data collection, maintenance, and security) all managed by the IT team. BI is now moving toward a completely decentralized model that does not need IT assistance. In this model, the business user implements the data preparation and visualization by directly accessing different data sources. This approach enables easier, broader use by both small and large organizations. The new entrants in the BI space primarily utilize this model.
- Convergence toward next-gen BI architecture: Currently, no single platform can satisfy all BI needs. Organizations use multiple tools for the different aspects of BI, such as data reporting, logging, visualization, and advanced analytics. Individual component players are now adding more capabilities to transform into platform players. They strive to become the single BI and analytics source for their clients. Even the traditional BI platforms, which are IT-centric, have attempted to offer better analytics and visualization tools, but they have been dragged down by their existing infrastructure. IBM made some progress with its Cognitive Analytics platform but has not yet been embraced by enterprises
- Taking this estimate a step further, we highlight a relevant observation from our proprietary survey. When asked about the type of BI & Analytics tools they currently used, over 40 percent of IT professionals said they use all three major analytics tools. In other words, a clear change has occurred in the way analytics software tools are used today. Their core value proposition has evolved from simple dashboard-style tools to more intelligent software that can help drive business decisions.
- Second, according to the 2016 IDG Research Survey, data analytics is the top application expected to move to the cloud.
- Third, according to our two recent investor sentiment surveys, we learned that big data and analytics represent the most promising growth areas of all the software applications
- Key secular trends, such as mobility, data growth, and movement to the cloud, have increased data analytics needs, triggering the rise of several new analytics vendors. These trends have led to an exponential increase in unstructured data from multiple sources. Our survey of IT decision makers suggests that over 90 percent of companies today use unstructured data. Palantir, with its advanced data integration technologies, stands to benefit from these secular trends.
- Beneficiary of growing spend on cybercrime, anti-fraud measures, and counter- terrorism
- Continued significant growth potential in government IT spend
Threats
- the key issue for investors to monitor over the next eighteen to twenty four months is the effect of competition and the mix-shift from government/services revenues on Palantir’s profit margins
- We have been encouraged by recent media reports that Palantir remains on track to become profitable by the end of 2017. However, the long-term cost structure of comparable companies remains unproven and debatable. Over the longer-term, as Palantir shifts from a services/consulting-driven sales cycle to a traditional SaaS sales cycle, we expect longer sales cycles, which will likely lead to structurally higher spend.
- Finally, the profitability threshold seems to have been pushed back. Smaller, younger companies are not reaching profitability at the levels where their larger, older counterparts were churning cash. A quick comparison of analytics vendors’ margins at increasing revenue milestones shows that the larger, established analytics vendors, such as Adobe, have been able to achieve profitability, while the profitability of smaller vendors, such as Workday, are well under negative 20 percent.
- Data privacy and security challenges
- Since Palantir works with government agencies that solve fraud and crime, Palantir’s products interface with a significant amount of sensitive information, both public and private. As such, Palantir has an additional responsibility to be more transparent about how it uses public data and protects public information. Recently, reports of information leaks and hacks across enterprises have put a greater spotlight on companies like Palantir.
Porter's 5 Forces
Threats of New Entry
- Test
- Test
- Test
- Test
- Test
- Test
Threats of Substitute Products
- Test
- Test
- Test
Competition
- Test
- Test
- Test
Bargaining Power of Customer
- Test
- Test
- Test
Bargaining Power of Supplier
- Test
- Test
- Test
PESTEL
Politic
- Test
- Test
- Test
Economy
- Test
- Test
- Test
Social
- Test
- Test
- Test
Technology
- Test
- Test
Environment
- Test
- Test
Legal
- Test
- Test
Business Model Canvas
Value Proposition
- Palantir Solution p.18
- We promote human-driven synergies between humans and computers by integrating every data store you have – any kind of data and at any scale
- Its unique value proposition was that it provided a single platform for all the analytical needs of enterprises and government agencies, primarily to detect fraud and crime. Palantir developed technologies such as dynamic ontology to integrate all kinds of data sources into a single platform where non-technical users could focus on making investigative queries rather than learning complex query languages
- They focused on one core product – machine augmented6 data analysis – and built an almost-exclusively engineering team to make it happen
- Palantir Gotham: Palantir Gotham, previously known as Palantir Government, was Palantir’s first product, primarily designed for the government to help defense agencies identify potential threats. The platform, whose underlying model is fundamentally a graph, is primarily structured around objects and the links between them. Both structured and unstructured data from multiple sources is entered into the system. The data is then transformed into objects and generated into a model. Users can use the model to describe, explore, and query properties of and relationships between those objects. The multiple applications built over the platform enable users to search, visualize, hypothesize, discover patterns, and share insights with colleagues, all within the platform. By reducing friction between users and their data, Palantir Gotham augments the intelligence of the entire enterprise.
- Palantir Metropolis: Palantir Metropolis, previously known as Palantir Finance, was primarily designed for the finance industry to help with fraud detection. The platform is structured around time series. Its underlying model is, fundamentally, a stream of events. The platform is popular for analyzing insurance claims data, network traffic flow, and financial trading patterns. It helps users mathematically analyze the behavior of models (e.g., stock ticks) over time. Metropolis primarily supports aggregate analysis: Pick a set of models and a time period and then run sophisticated mathematical calculations over them. Metropolis stands out from the competition in this space because of its rapid iteration and collaboration abilities.
- data visualization, predictive analytics, and geo intelligence spaces
- Palantir, for example, plays only in the data visualization, predictive analytics, and geo intelligence spaces but is the domain expert in those areas. While Palantir has been expanding its portfolio in order to offer other analytics features, it is challenged by heavy competition from the rest of the players in the industry.
- To further illustrate Palantir’s large and growing market opportunity, we highlight a simple takeaway from our survey of IT decision makers. We asked IT professionals to identify their top priorities in data analytics over the next 12–24 months. Our results in Exhibit 8 reveal that the top priorities for businesses today are data integration and visualization, followed by insight generation. The leaders in this space are currently the companies that have better tools for these important features. Thus, Tableau, Qlik, and Splunk have made headway in this market despite competing against larger players, such as Microsoft, IBM, Oracle, and SAP. Palantir, although it is not yet widely used by enterprise customers, possesses excellent tools in integration and visualization, which have been extremely valuable for both government and commercial agencies.
- A quarter of our survey respondents believe that predictive analytics, along with machine learning integration, represents the future of analytics. This makes sense in the world of ever-increasing data volumes, where reporting and visual tools are not sufficient by themselves. Business leaders need help parsing that data to make business decisions.
- Palantir platforms are used across multiple industry sectors to help with data integration, data visualization, and data analytics needs. Palantir excels at integrating data of almost any kind: text, emails, logs, and even images and videos. By creating reference objects to its data, it enables searching for and identifying links between different pieces of data. Analysts in different industry streams then use this data to perform further investigation.
- In the anti-fraud, counter-terrorism, intelligence, and crime prevention spaces, Palantir helps integrate data from random sources such as on-site reports, phone numbers associated in First Information Reports (FIRs), and even low-resolution images from mobile devices. The data is then made available through Palantir Graph or Map, easy-to-view applications that help investigators spot patterns between events and develop hypotheses. In the insurance, finance, and health care industries, where data sources and record keeping is extremely important, Palantir uses its knowledge management application to track and secure every piece of data that comes through its platform. It also ensures the association of appropriate security levels as it indexes the data into the platform for future referencing. Through its collaboration application, Palantir helps analysts and investigators working at different sites to easily share their findings and update data in real time, all on a single platform.
- I’d like to take the position that the real differentiating feature of Palantir is the service component, and the ability of its operating model to execute engineering talent arbitrage
Revenue Stream
- In terms of bookings, we estimate that Palantir is approaching $3.5 billion in 2017, with roughly 40–50% of these bookings coming from government contracts.
- Significant European expansion; revenue from Europe triples since 2014
- Palantir is an enterprise software company that derives its revenue through two standard pricing models: SaaS subscription and professional services. We believe Palantir’s quarterly revenues follow an SaaS-like trend, with sequential growth from quarter to quarter
- A typical Gotham installation consists of: $564K Price for a Palantir server, assuming 4 cores $112K Software Updates and Maintenance for server, per year (first year free) $100K Training, assumes 50 users $600K Engineering Services for integration. Assume 2-3 deployed engineers for 1 year Total: 1.5M for 50 users
- What’s especially confusing about Palantir is that it doesn’t make it entirely clear whether it’s a product or services company. In addition to charging for usage of their actual software product, upon installation they also offer large contract bundles of engineering, deployment, and training services, all of which they charge a premium for.
- Milk
Cost
- While Palantir might spend a relatively smaller proportion of its revenues on sales and marketing, its peers (such as Splunk, Tableau, and Qlik) spend over 50 percent of their revenues on attracting and retaining clients. We believe one of the primary reasons for these companies’ elevated level of marketing spend is intense competition from large, established enterprise software vendors. Smaller companies tend to spend more to get a foot in the door with enterprise clients, offering deep discounts to persuade customers to change their analytic platforms. However, even after smaller companies take this step, established players can cut prices to retain their customer base. Since Palantir does not have to spend as much of its revenue on sales and marketing, this could result in significant amount of savings, adding to its bottom line.
- he reason Palantir is able to charge such high prices in their contracts is that their services are an incredibly high value add for their customers. Additionally, Palantir’s prices are actually relatively cheaper than similarly priced tech solutions such as Bloomberg terminals or SAP. In a typical general example, a Palantir deployment over 3 years for comes out to $11,000 per user (3), which is less than the $24,000 price for a Bloomberg Terminal (5). Considering that the yearly salary of a typical analyst ranges from $40,000-$90,000, the per user cost over 3 years is a relatively small cost, especially if Palantir can enable those users to be more effective by an order of magnitude.
Key Resources
- In terms of fundraising, Palantir has raised about $2.75 billion in total equity capital since its inception. The company’s most recent funding round was completed in January 2016, when it raised about $880 million, resulting in a post- money valuation of roughly $20 billion
- Palantir developed several applications (such as Graph and Map) and innovative dashboards and technologies (such as Dynamic Ontology and Nexus Peering) to deal with huge volumes of structured and unstructured data.
- We believe Palantir’s employee base has grown rapidly over the past five to seven years. The company was estimated to have fewer than 50 employees in 2009; however, by 2017, the company grew to over 2,000 employees worldwide, establishing a major presence in Washington, D.C., Europe, and Silicon Valley.
Key Partners
- The CIA’s venture firm, In-Q-Tel, a part of the U.S. intelligence community that invests in new technologies for defense operations, was an early investor in Palantir.
Key Activities
- Airbus taps Silicon Valley expertise (Palantir) to speed production of A350.
- Germany’s Merck taps Palantir for big data health initiative
- While it’s difficult to find publicly available Palantir pricing data, we estimate that a typical Palantir Gotham or Metropolis license, per server core per month, could range between $3,000 to $10,000 per month, with extra charges for support, maintenance, and training.
- Software services tend to have lower margins compared to products and licenses, largely due to their variable personnel component. We believe that Palantir bundles consulting, support, maintenance, training, and other such on-site services with its customer contracts. Over the longer term, we expect the company to “productize” its services, as it has since starting a clear push into enterprises
- From encryption and multi-level permissions keeping the bad guys out to audit trails and anonymization keeping the good guys good29, Palantir works to protect freedom
- Next, looking at Palantir’s job listings, it’s even more obscure. They’ve made up their own titles for some positions. Designer - At Palantir Office Software Engineer - At Palantir Office Forward Deployed Software Engineer - Customer Site Deployment Strategist - Customer site Product Expert - Customer site This structure looks more like an Accenture consulting team, rather than a Silicon Valley product company.
- Financially, it’s obvious from the breakdown that the engineering services are the highest margin. Assuming you can pay FDEs $100K/year with lucrative stock options, Palantir has a 60% margin per deployed engineer. Moreover, customers often keep FDEs longer than required because they are miles better than their own employees. Operationally, however, Palantir’s thought-leaders are the product engineers, who come up with advanced data processing algorithms. I speculate that these special titles are a HR move to compete for the best talent in the Valley. Top engineers wouldn’t respect working at a technical services company, so Palantir must re-brand as a product company to attract the best technical talent. This approach has been highly successful as Palantir is seen as a Silicon Valley darling.
- Palantir prices its product by charging the client a large multi-million or even billion dollar deployment contract which includes onsite implementation teams, training services, maintenance, and a number of server cores. Since every client has a completely unique set of data sources, Palantir deployments are incredibly high touch and require a high level of customization.
- VALUE CHAIN. Big data analytics involves examining large amounts of structured and unstructured data to uncover hidden patterns, trends, and correlations that can help businesses make informed decisions about their products or services. Businesses use the tools and applications that perform data extraction, curation, analysis, and visualization to support strategic decision making. While the term “big data analytics” describes a wide variety of functions and tasks, as illustrated in Exhibit 5, industry experts believe the broader software analytics space includes three broad categories of functions: (1) Business Intelligence and Analytics Tools/Platforms; (2) Analytic and Performance Management Applications; and (3) Analytics Data Management and Integration Platforms.
-
Customer Segments
- Palantir’s early clients included a fairly long list of federal, state, and local agencies and departments, such as the Department of Defense, the CIA, the FBI, Homeland Security, the LAPD, the Chicago PD and the NSA.
- The company worked with different government agencies (such as the FBI, the CBI, the Department of Defense, and the Department of Justice) to integrate all of their data into a single platform. This step was extremely useful in solving complicated cases, as analysts were able to identify patterns in the different data sources.
- Palantir loses a key cybersecurity client: The Home Depot <
Channels
- Finally, according to Palantir’s publicly traded peers, market opportunity is growing at a fairly robust clip. Tableau estimates that the available market for its products is going to grow from about $15 billion in 2013 to $26.6 billion by 2019, primarily through new customer acquisitions and increasing channel bookings. The company believes its international growth will contribute about one-third of its overall revenues by the end of 2017.
- Palantir developed several applications (such as Graph and Map) and innovative dashboards and technologies (such as Dynamic Ontology and Nexus Peering) to deal with huge volumes of structured and unstructured data.
- Word of mouth
Customer Relationship
- intelligence, law enforcement, and homeland security customers.
- SaaS subscription and professional services
- Data analytics vendors price their products in different ways. Some price based on amount of data used; some price per user; and others use a combination of both user number and data-based pricing. Although some of the companies offer perpetual licensing options, most use a subscription-based model for their clients. The pricing models can vary from very simple (e.g., Tableau’s single pricing model) to very complex (e.g., Palantir’s pricing, which is not publicly disclosed).
- A quick look at Palantir’s pricing chart reveals that, although its licenses cost around $140,000–$150,000 per server, its services can add up to anywhere between $350,000 and $750,000 per year. Given the complexity involved in integrating diverse data sources, some of these services are practically mandatory for businesses planning to use Palantir’s products
- As far Palantir is concerned, two open questions remain: What proportion of Palantir’s revenues are derived from perpetual licenses? What is the average duration of Palantir’s government and commercial contracts for revenues derived from SaaS licenses? b2b company like Palantir
- The company states that they practice value-based pricing, which allows them to maximize WTP for each customer. For example, on the government side, Gotham’s only comparable product is the U.S. Army’s in-house Distributed Common Ground System that cost of $2.3 billion. A leaked document cites a 2012 study where 96% of the surveyed war fighters in Afghanistan preferred Palantir. So we can assume that Palantir can charge up to and over 2.3B for their product - to the Army alone
- Today, Palantir’s expanding to retail/CPG companies as they look for new-market opportunities with shorter selling cycles than government. Unlike the government, these companies do have comparison points to IBM BI, SAP, and in-house solutions. As they enter this industry, Palantir should expect to see its margins decrease due to increased competition, especially since IBM BI may even underbid just to keep out Palantir. It will be interesting to see if these lower margins will eventually spillover to the government side, as greater transparency should identity the true value for data analytics products.
- Once Palantir has built their employee base, they can charge their clients a premium for the engineering talent they are desperate for. This is not to downplay their technology stack – but given how much customization it requires and that Palantir’s business model sells its product not by software packages but in the form of large deployment contracts per its business model, the way Palantir’s operating model attracts, invests in and manages its labor capital is one of the primary reasons for its current success!
Monopoly Characteristic
Proprietary Tech
- Palantir is able to integrate all forms of data, including real-time logs, emails, chat conversations, social media content, and even images and video content. Engineers, as they help install the platform, pool data from all sources, ingesting them into the Palantir platform and creating objects that can be referenced during an investigation. The Palantir platform, unlike other major analytic platforms, can be installed and become fully operational in a matter of weeks, instead of months or years. Over the years, Palantir has added several new technologies to its two product platforms, based on its learning in the defense and finance domains, and it has proliferated them to other industry domains. Palantir is very popular among customers primarily because of its user-friendly applications, which help identify unusual patterns. Analysts can then research those patterns further to develop investigative hypotheses while maintaining data privacy.
- The Palantir platform merges human-based algorithms and a powerful engine that can scan several databases at once on an incredibly fine level. The company’s main sources of revenue are its two large analytics platforms: Palantir Gotham and Palantir Metropolis. Integrated into these two platforms are technology solutions addressing different analytic segments. The system accepts huge databases and allows users to slice information in seemingly innumerable ways, with appropriate sensitivity to all the necessary security needs
- Handling unstructured data is one of their core competencies today
- In the startup community, we love talking about the importance of fast iteration, but that’s easy to do when pushing code live is as simple as hitting refresh on a server. In the enterprise world, computers are often not connected to the internet, so pushing code live means an engineer has to fly to the company’s physical location and upgrade the software in-person. The way that Palantir made fast iteration a priority despite the extra challenges helped it to develop a battle-tested, world-class tool in just a few busy years. Porter: Bargaining Power of Customer
Network Effect
- the large number of enterprise and government agencies that continue to adopt the platform
- Palantir introduces Horizon technology to drive interactive workflows on large amounts of data
- High-profile customers providing sales and marketing leverage
Econ. of Scale
- We believe the company will continue to maintain its valuation as it adds larger clients such as Airbus and Merck and expedites cost-cutting through automation
Branding
- For Palantir to remain competitive, it needs to continuously reinvent itself to stay relevant as the competition expands its offerings. This reinvention could require significant investments that are cost-effective only with scale. With the numerous vendors in the analytics space, Palantir’s investments in R&D and sales are expected to increase, contributing to the uncertainty around its long-term margin profile. Additionally, Palantir may find it difficult to compete against companies with deeper pockets and wider brand recognition, since these larger companies can heavily discount emerging products in order to gain traction or retain customers.
7 Questions of Business
Engineering Question
- Palantir develops data analytics software that addresses fraud prevention, counter- terrorism, and other business intelligence tasks.
- Today, Palantir’s software works seamlessly by importing reams of structured data (such as spreadsheets) and unstructured data (such as images and social media posts) into one centralized database, where all of the information can be visualized and analyzed.
- The Palantir platform merges human-based algorithms and a powerful engine that can scan several databases at once on an incredibly fine level
- Problem: SQL has a frustrating learning curve. It’s just hard enough that mostly only analysts know how to use it today
- Solution: Palantir’s data analysis runs on natural language querying, so ordinary people can easily use it 7. This has opened up huge opportunities to get more eyes on important data. BOS: REDUCE
- Problem: Data is scattered everywhere. For instance, if the FBI had a case file on someone, there’d be a good chance the CIA wouldn’t even know about it. Similarly officers in LAPD would have to search one database for robberies but run a different search to check for murders.
- Solution: Through talking with potential users, Palantir realized that most data is fragmented and much of it is also unstructured. So they built their tools to combine multiple unstructured sources. BOS: REDUCE
- Problem: Queries are slow. Even as the computing industry has kept pace with Moore’s Law, speed is still an issue for data heavy tasks like analysis. When you have a small database of 10k rows, you can run a query pretty quickly, but once you’re handling many gigabytes of information, you might easily wait half an hour for the results… if the results don’t timeout or hit memory problems
- Solution: Palantir optimized for speed both in run time and in iteration. BOS: REDUCE
-
- After investing years of thought and programming into Palantir Gotham (the government product), Palantir realized that many of the same core competencies – handling unstructured data, combining multiple data sets, querying speed – could be used to great effect in the business world
- Palantir’s technology is its competitive advantage. No other system allows users to draw associations between disparate data sets and to visualize the connections as easily or as quickly. On the back end, Palantir’s infrastructure uses cutting-edge data processing algorithms to search through huge data sets at lightning speed.
Timing
- Our investment thesis is driven by the growing need for: advanced yet end-user-friendly data analytics, tracing intellectual property back to its government relations, higher spend on big data analytics and cybersecurity.
Monopoly
- Palantir, for example, plays only in the data visualization, predictive analytics, and geo intelligence spaces but is the domain expert in those areas. While Palantir has been expanding its portfolio in order to offer other analytics features, it is challenged by heavy competition from the rest of the players in the industry.
- Overall, we believe the market size for Palantir is between $40–$50 billion, primarily in the data visualization, predictive analytics, and fraud analytics spaces.
- While Palantir’s approach to massive data integration can sound similar to the work of established technology companies such as IBM, however, it takes a pointedly different approach to its work on the ground.
- On a new project, the company usually sends in a small team of no more than three engineers with a mission to come up with the outline of a possible solution within days. Palantir engineers relish their role as the technology world’s outsiders: they do not stand on ceremony and are often deeply unpopular with the in-house IT experts whose work they second-guess.
- Rival artificial intelligence companies, meanwhile, snipe that Palantir has traded on clever self-promotion and that its data integration technology is far from unique in the industry. They also say that its approach can sometimes be slower and more expensive than pure AI companies, since it uses teams of engineers to devise on-the-fly responses to problems rather than simply applying existing algorithms.
- But as the company’s valuation rises to $20bn, the mystique it has succeeded in creating around its approach to data analytics has turned out to be highly valuable
People
- Founded in 2004 by former PayPal executive Peter Thiel, along with Alex Karp, Joe Lonsdale, and Stephen Cohen from Stanford University,
- Palantir Technologies is a data analytics software provider founded in 2004 by former PayPal executive Peter Thiel, along with Alex Karp, Joe Lonsdale, and Stephen Cohen from Stanford University
- CNN names Director Shyam Shankar one of the world's top 10 thinkers
- The hiring process was equal parts talent and culture. Particularly on the engineering side, talent is a great rational measure (“can you write a program in C+ to do x?”), but assessing someone for culture fit is always a bit more tentative. In Palantir’s case, they settled for a single measure that factored in all cultural elements: “is this a person you’d like to work with?”
Distribution
- Joe Biden, the Vice President of the U.S. at that time, recognized the company’s fraud-fighting capabilities. This honor was followed by several other recognitions, such as, being include in the list of Best 50 Tech Startups by the BusinessWeek, CNBC and Chase Hall of Innovation Award
- Although a May 2016 BuzzFeed article revealed some of the tensions between Palantir and the government agencies it served, the company’s bookings continued to grow steadily, to over $1.7 billion in 2015, thanks to its diversified client base in the nongovernment sector. Palantir’s involvement in social causes – such as its joint effort with Santa Clara County to tackle homelessness – enables the company to display its platform’s wide capabilities and also gain publicity through word of mouth.
Durability
- The demand for big data analytics, data visualization, and business intelligence software is estimated to grow from $150 billion in 2017 to $210 billion in 2020, translating to a robust 12-percent growth (three-year CAGR). While Palantir’s serviceable market opportunity remains subject to debate, our analysis of key market trends and Palantir’s product evolution suggests that the company could target an annual spend of $40–50 billion, primarily in the data visualization, predictive analytics, and fraud analytics space, implying a fairly robust and sustainable growth.
- Palantir’s high touch installation also serve to help build long term momentum since it’s hard to remove a deployment once it has become the new standard for the client and years have been spent building it. This stickiness has allowed Palantir to quickly grow to half a billion dollars of revenue in such a short period of time, with no signs of slowing down
- Anonymous is correct. Palantir is not in the AI space, and uses machine learning only to support its primary data analytics focus. Examples:
- Pattern matching. In the banking industry, Palantir helps analysts investigate fraud by surfacing details to analysts for them to investigate interactively. When analysts identify a fraudulent transaction, Palantir trains models to represent and identify that form of fraud, which are then used to prioritize similar cases for analyst investigation.
- Predictive analytics: e.g., Given time series data about how an industrial part will perform over its entire life and what it looks like before it starts to break, Palantir will attempt to identify parts *before* they break so they can be serviced or replaced during a scheduled maintenance window instead of requiring an emergency outage, potentially saving LOTS of money.
- So yeah, Palantir uses machine learning to augment the effectiveness of its analytic platform. But again, they're not involved in AI whatsoever
Secret
- Unique, user-friendly customer proposition
- Palantir’s key differentiator is its unique ability to integrate volumes of structured data (database tables, spreadsheets, etc.) and unstructured data (images, videos, reports, etc.) into a user-friendly analytics platform. Anyone who deals with searching and sorting through large volumes of data understands that it can be frustrating to use database- querying languages. Typically, only dedicated analysts can effectively use these query languages. Palantir, by using natural-language querying, has turned this problem upside down, so that anyone – even individuals with no knowledge of programming language – can easily conduct insightful searches through huge volumes of data. As a result, Palantir’s platform has become well-accepted among a wide variety of users across multiple government organizations. Due to the software’s ease of use, we believe end users in businesses can now focus their energy on solving cases, rather than wasting time searching through different databases.
- Palantir possesses three key features that make it a unique and go-to data analytics product: Palantir Graph, Palantir Map, and Palantir’s mobile application
January 3, 2018 Notes
Peter Thiel&Maria Bartimoro
- Palantir= cybersecurity, government around the world
- asymetric in cyber; attack very easy; defend very hard
March 8, 2018 Notes
GovCon7: Intro to Palantir
- Problem of Analysis in a world of big data
- Paypal: Analysis problem