Company

Cape Analytics leverages geospatial imagery, computer vision, and machine learning to provide geospatial structured data and aerial imagery for insurance and property underwriting.

Year founded: 2014, HQ: Palo Alto, CA, https://capeanalytics.com/

Funding: Series A, $14M, 2016.10, Investors: Formation 8 (lead), Data Collective, Lux Capital, Khosla Ventures

Partners: Nearmap (data partner), TigerRisk Partners (reinsurance and capital advisory), Nephila Advisors

Market Opportunity

In 2015, the US home insurance industry delivered $86B in revenue, expecting ~3.4% growth/year. The US P&C insurance industry earned $608Bn premium in 2016, but loss ratio rose to 60.57% from 56.02% in 2013.

Inefficacies in assessing properties: Traditionally, this involved prospects filling out long forms and insurers quoting based on incomplete or outdated information from public records or after conducting expensive on-site property inspections. As a result, there are often discrepancies between the quotes and the charged.

Lack of data-driven dynamic pricing: the residential homeowners underwriters either don’t inspect renewal business at all or on a 3-to-5-year cycle. When situation changes, insurers take long time to react.

Failure in modeling catastrophic events: Recent reinsurance demand growth in the US has been focused in casualty lines. Catastrophic events accounts for huge losses, but insurance models have “short memories”.

Product and Strategy

“We’re absolutely not trying to bill ourselves as a replacement to property inspections. I think our offering is a lot broader that that simple economic measure.” — Ryan Kottenstette, CEO of Cape Analytics

Business model: pull down geo-imagery from partners (outdoor, satellite photos), brings the imagery onto its internal platform (like Google Earth), instantaneously and at scale, for entire property portfolios, and produces structured property for insurance carriers

Products: MS Isolation Platform, Web Isolation Service, Doc Isolation Service, Email Isolation Service

· Cloud-based platform; data integrates into an insurance carrier’s quote engine via API

· Provide near inspection-quality data of high-value property features

· Property data include: roof condition and characteristics, square footage measurements, parcel features such as solar panels and pool enclosures.

Features and Benefits

· Shorten insurers’ application processes, increase price policies accuracy, and reduce post-binding adjustments/cancellations by allowing insurers to instantly pre-fill property information at time of quote

· Reduce the risk (namely, increase potential revenue) for renewal business by reminding insurers to inspect properties that have an extension or a guest house or pool is newly built

· Increase transparency of exposure risk at a portfolio level and provide insights into data quality for specific cedents with new cedent portfolio product offering. The solution drops directly into current reinsurer workflows, offering them instant, accurate, and objective data at scale about each individual risk in their cedent portfolios and across the entire portfolio

· Easy to implement: the data solution format plugs into leading CAT risk models from AIR and RMS without any change to reinsurer workflows

· Improve risk assessment: Reinsurers can submit their portfolio to Cape Analytics, which will send back an augmented data set with delta analysis in under 24 hours, allowing reinsurers to identify exposure changes of up to 15% on a portfolio basis, dramatically improving risk assessment.

· Improve catastrophic event management: In an event like hurricanes, the product can help identify the properties in the insurers’ portfolios most impacted, identify reported and unreported losses, prioritize the efforts and deliver better service.

Customers: Security First Insurance, Cabrillo Coastal, etc

Management team

· Ryan Kottenstette, co-founder, CEO, Principle @ Khosla Ventures, BD director @Amprius, Inc

· Suat Gedikli, co-founder, CTO, Senior scientist @Open Perception, Inc., PhD in image processing

· Olivier Collignon, VP of Data Analytics, VP Business Analytics @LoopNet, Risk Applications director @RMS

Pros

· Promising market: The P&C insurance industry represents a billion-dollar market for data companies.

· Strong value proposition which is well-aware and widely-accepted by the market: increase costs-effectiveness in operation (risk assessment process), open-up potential revenue source in renewal business, and reduce potential lost of capital and customer in catastrophic events

· Experienced founding team with good market-fit: the team is specialized in cloud security/virtualization. Success in co-founding and exit a startup.

· Data partnership with Nearmap, whose imagery provides detailed 2.8” GSD resolution that is updated on a frequent basis, contributes to the competitive advantages

· Highlights of technology

o Clientless: No endpoint software needed, no impact on customers’ native experience

o Compatibility: Ideal compatibility and Integrates with existing infrastructures

o Supporting various vectors including Web content, documents, phishing and endpoints

Cons

· Technology: Its stealth IP is in collaboration from Juniper, Check Point, UC Berkeley, maybe similar to others with similar technology. Adding a layer of abstraction may lower speed.

· Attraction for decision makers: Focus more UX than available policies

Driven by mobile, online sales in the US are expected to increase 56% reaching $523bn in the next five years. Research has shown that higher income online shoppers are more interested in buying something hard to find elsewhere.

Stylewe targets growing middle-class female customers between the ages of 30–50 with an average annual income of $60,000 — in contrast with Wish.com and LightInTheBox on the low end, and Farfetch on the high end. The average order size, $170, has increased significantly since the company started, and fits well with need and consumption power of its target customers.

Stylewe has an experienced management team. Founder & CEO Xingjian Yang has more than 10-year experience in cross-border ecommerce, including as founder of Tidebuy, the first wedding dress crossborder B2C company in China.

The marketing strategy is to help designers build up their brands primarily through social media (Facebook, Snapchat), particularly by recommendations from celebrities and fashion bloggers. These efforts successfully retain valuable designers on the platform.

In 11 months of operation, StyleWe has made impressive traction:

400+ licensed designer brands, 300,000+ registered users, and 150,000+ active users.
Monthly GMV exceeding $8M. Vertical integration drives revenue share with designers of over 50%. Delivery is $10/order for 8–10 day delivery and is expected to decrease with scaling.
High customer retention rate: ~30% web and ~60% mobile bolstered by high quality products, responsive customer service, and flash sale model.
Lenovo will support Stylewe with its cloud service and supply chain resources. A Lenovo VP will serve as Stylewe’s chief design advisor

By 2016Q1, Stylewe reached ~$48M annual revenue. With steady demand and improving operational efficiency, it has the potential to reach $100M annual revenue in 2 years.

Source:
CEO interview
http://www.baijingapp.com/article/6984
http://www.uplier.com/xqfs/5515.html

Founded: 2013, Funding: Seed round $1.2M in Aug 2015, Seed round $6.5M in Aug 2016

Seed investors: Michael Antonov, Crosslink Capital, Haystack, Liquid 2 Ventures, WI Harper Group

— — — — — — — — — — — — — — — — — — -

With forthcoming FAA regulations and advancements in intelligent navigation technologies, commercial drones are expected to generate billions dollar of value for the US economy by significantly improving efficiency in various industries such as agriculture, construction, insurance, and film, while enabling data that have never been collected before. Prenav is a drone platform startup which can bring this value to the task of industrial inspection.

Industrial inspection is one of the most promising applications for commercial drones. In the US, over $3bn is spent on building inspection alone, and even more on larger infrastructures such as cell towers, wind turbines, and bridges. Meanwhile, the demand of maintenance is increasing as consumption grows. However, the traditional inspection process has been extremely inefficient, expensive, and dangerous. For example, an annual rope-access inspection at a wind farm can cost $1,500 in addition to the costs incurred from shutting off the turbines for at least half a day. It is also difficult to get consistently good imagery and complete coverage of the structures by climbing or using hand-flown drones.

While fundamental challenges in robotics generally deal with localization, planning, and computer vision, commercializing drones in industrial inspection faces the following major problems:

· Existing navigation systems typically rely on GPS data and collision sensors that result in slow communication between drones, fail to guarantee the precision needed for ultra-close views in maintenance inspection, and cannot support tasks indoor or in congested cities

· Most drones are manually controlled by skilled pilots which incurs greater costs for companies

· Most drone solutions don’t provide a complete workflow, leaving customers responsible for third party tools such as image processing

Prenav’s LiDAR-based 3D mapping and construction data analysis frees human operators from unsafe large infrastructure inspection and enables data-driven asset management. The system combines a drone, a ground-based guidance robot, as well as software to plan the project and analyze the resulting data. The robot performs an initial laser scan of a structure then creates a 3D point cloud to guide the drone through precise routes. This solution provides unprecedented business value in multiple aspects:

· With precise navigation, PreNav’s system allows close-up photographs of structures and provides final output of high-resolution 3D render that can be used for inspection purposes — more valuable for customers than the videos-and-images report provided by current solutions

· With laser-based sensors and guided by computer vision, PreNav drones are able to safely fly in complicated or GPS-denied environments with a high degree of precision and repeatability

· The fully automated system enables any existing workers to use without hundreds of hours of pilot training — an important added-value after FAA’s new rules.

· PreNav’s complete workflow makes it easy for customers to operate the system and utilize the data from the cloud. It also has the potential to retain this data advantage through machine learning

PreNav hasn’t released its pricing yet, but has signed contract with infrastructure companies like Senvion, OWBS and Spectrum services, saving customers an average of $500/inspection. The team has impressive 70+ years of experience in robotics, computer vision and aerospace, starring a few big players including former CTO of Bot & Dolly (acquired by Google) Nathan Schuett. With current funding, PreNav expects to officially launch product in late 2016, and push further into functional indoor jobs.

Founded: 2012, Funding: $8M Series A round in Jan 2016

Investment analysis

Street Contxt is a vertical CRM that manages daily workflow of developing, deploying, and tracking morning notes for the financial services industry. Thus, it significantly improves marketing effectiveness and research team management on the sell-side; and investment analysis efficiency on the buy-side.

Wealth management (WM) firms in the U.S. spend $150bn in annual trading commissions for sell-side professional services and the number has been increasing even during market downturns. Morgan Stanley alone paid $2.3bn in professional fees for its WM business in 2015. However, the sell-side still predominantly distributes thousands of morning notes and research reports via email/pdf attachments, leaving no collections or analyses of associated data. Most banks use circa 1990’s in-house CRMs which require manual data entry and consistently fail to keep up with the fast paced world of trading. Sell-side firms are calling for tech-driven tools to analyze the effectiveness of their morning notes and to quantify the performance of their expensive research teams.

On the other hand, buy-side investors who usually have to spend hours manually going through piles of buy-side notes to filter out noise and make informed decisions also encounter a serious pain point. Bringing data-driven solutions will make the previously opaque and relationship-driven processes more meritocratic, while incentivizing better research and fairer yield.

Street Contxt integrates into the existing email-based workflow of sell-side analysts and tracks their notes to determine if they were opened and for how long. This information is then stored in a CRM, which provides analysts visibility into the effectiveness of their campaigns and commissions made. The information also allows sell-side managers to evaluate their teams and quantify performance. The key feature for buy-side is content discovery. Based on users’ interested topics, Street Contxt’s software automatically builds a profile for readers that can still be used even if that person changes e-mails or firms by scanning wall street reports for key words and themes. Long-tail small firms can also subscribe to get previously exclusive research reports or pay per report.

Compared to similar firms such as Mail Chimp, an e-mail marketing tool to track open and click rates, or Blue Matrix which tracks readership, Street Contxt goes further to address the industry-wide problem, connecting the two parties by providing valuable services for buy-side. In the long run, Street Contxt has the potential to become the portal of Wall Street information, owning the allocation of billion-dollar trading commissions as well as great volumes of proprietary structured data.

The management team and the advisor group have strong domain expertise and entrepreneurial experience. Founder and CEO Blaire worked at sales and trading at RBC Capital market. Blaire’s brother, Ted Livingston, is CEO of Kik, a Canadian messaging app with $1bn valuation.

Currently, Street Contxt charges users $500-$5,000 per year, and has gained half of top 20 wall street brokerages signed up. In the concentrated WM market where number of users cap around 400k, the company could capture large market share with its vertical-focus approach. It can exit via acquisition by horizontal incumbents that have historically paid more than 5x revenue multiple for similar deals. Alternatively, it can go public in 5–7 years with an expected higher EV/LTV multiple than average for vertical SaaS since it targets the relatively lucrative trillion-dollar financial services industry.

A fundamental strategic advantage of vertical-focused enterprise software comes from improving workflow efficiency in major industries by structuring, integrating, and analyzing big data for knowledge workers — a product strategy that can only be accomplished by being vertical-focused. These companies also have the potential to develop into open infrastructures within verticals that allow for further third party innovation.

The aforementioned industries include energy, finance, healthcare, business services, transportation, and manufacturing; which are normally large in size but haven’t fully utilized their exponentially growing data to disrupt decision-making processes. Therefore, enterprise software companies that dramatically improve productivity in those industries present significant investment opportunities. The value of going-vertical strategy has been proven by successful IPOs of vertical software companies such as Veeva Systems, DealerTrack, and Fleetmatics. Key aspects for investing in these companies are:

· Targeting critical “front-office” workflow: Certain back-office business functions, like payroll and accounting, are similar across industries and therefore appropriately served by horizontal platforms (e.g. Workday, RelateIQ). However, front-office roles tend to be the main driver of business success and require specialized software that standardize best practices from the industry to identify and improve vital workflows.

· Integrating previously disparate metadata: Structuring data and drawing insights is vital for companies, while not all data is the same. A differentiated product is based on deep understanding of workflow and a focus on integrating industry-specific data to empower the workforce, instead of developing generalized black-box algorithms to make human decisions.

· Potential to build open platforms with proprietary data: Platforms typically begin as solutions for single problems. With access to more data, they form a network for third party developers and enable data contributed by additional clients to be used by a broader customer base.

The benefits of going vertical have been proven by public companies’ performances:

· Faster scale with lower costs and higher profitability: On average, EBITDA margin of public vertical SaaS is 20%-25% higher than horizontal SaaS, while SG&A expenses are 35% lower. Industry-focused strategy leads to more referred revenues, bottom-up sales strategies, and smaller sales/marketing teams, thus lower customer acquisition costs (CAC).

· Understand user patterns through deep data: It takes time to understand industry pain points, transform legacy systems, and identify relevant data. High-quality data is vital for machine-learning-based analytics. Industry-focused data with more consistent user patterns enable vertical SaaS companies to generate meaningful insights from behavioral analysis easier.

· Capture more revenue per customer with multiple-layer product strategy: After achieving deep penetration with its core suite, vertical SaaS companies can leverage inherent data to add new layers of functionality for a single customer, such as business intelligence, data management, and mobile solutions.

Successful vertical SaaS companies share hallmarks including: having a clear big problem to solve, an executive team with deep domain expertise and technical strength, and a high percentage of referred revenue. Key metrics for valuation are LTM, EBITDA and growth rate.

Exits opportunities lie in the inevitable trend of consolidation. With the potential for a single company to achieve market shares over 50%, we will see fierce competition with every new entrant, and ultimately a result of less than 10 large public companies per vertical. While public vertical SaaS companies trade at a median EV/LTM revenue multiple of 1.5x lower than the horizontal group, Oracle’s acquisitions of OPower and Textura represent opportunities for exits with higher revenue multiples through M&A.

Email me at monicaxie.aiesec@gmail.com to let me know your thoughts!

**Special thanks to William Fan!

Part one: Why I’m interested in Blockchain-based applications in Internet of Things (IoT)

A blockchain network is a decentralized system that uses a shared ledger (public or private) for recording the history of electronic business transactions that take place in a peer-to-peer (P2P) business network. Major challenges in building a scalable IoT network include traffic constraints in current radio-frequency-based transport media between machines, difficulties in machine authorization and authentication, high costs of centralized cloud infrastructure, and system compatibility. The best way to monetarize IoT is through micro-payments. Blockchain technology will enable large scale machine-to-machine transactions by building trust without middleman costs. Notable cases include Filament (uses wireless sensors to create low-power autonomous mesh networks for data collection and asset monitoring) and P2P energy transactions like Transactive Grid.

Investing in blockchain-based applications is difficult. Unlike traditional internet, the value distribution here is different: the market cap of the “protocol” always grows faster than the combined value of the applications built on top, since the success of the application layer drives further speculation at the protocol layer. While the long time focus has been on payments and the securities industry, early real-world blockchain applications will still be limited to use cases such as trade finance, syndicated loans, or reference data management — all activities that are not dependent on a critical mass of assets already being on blockchain. Real-world use cases for payments and clearing & settlement purposes, however, will take years, and incumbent efforts like R3CEV and Hyperledger will leave thin profit for application-based startups.

— — — — — — — — — — — —

Part two: Investment opportunity analysis: Filament (https://filament.com)

Founded: 2012

Funding: Seed round $1.5M in Aug 2013, series A $5M in May 2015

Investment Thesis

As billions of devices are expected to connect to internet by 2020, IoT is showing particular promise when it comes to industrial applications, in which significant value is derived from detection/control to optimization/prediction. Global manufacturers are expected to invest $70bn on IoT solutions in 2020, up from $29bn in 2015. This trend is enabled by developments in lithium battery technology and smartphone-powered mobility; in addition to advancements in cost, power, and efficiency of sensor technology. However, the increase in volume of devices results in growing demand for machine-to-machine (M2M) communications that build an IoT system into a “single product.” Major challenges include:

• Security: Many devices use simple processors and operating systems that may not support sophisticated security approaches. The rise of M2M transactions present a greater vulnerability in potential attacks.
• Connectivity: More devices are needed in environments with low connectivity. Using radio frequency (RF) as a transport medium for devices within close proximity could cause frequency interference in certain spaces.
• Infrastructure: Cost of centralized network hardware lead to high logistical costs. Cloud processing and storage capacities could also become network bottlenecks.
• System compatibility: Diversity of device ownership and their supporting cloud infrastructure make M2M communications difficult. Cloud services are not interoperable and cross-system implementations need to balance many conflicting requirements.

Filament solves these problems with long-range wireless sensor networks secured by blockchain technology — a decentralized network in which autonomous endpoints can interact and exchange value without cloud or a central authority management device. Filament’s products include Tap — a hardware device with embedded environmental sensors, and Patch — a surface-mount module for custom solutions.

Key issues for this solution are: identity management, security, discoverability, digital trust, power, and integration of crypto-currencies. Filament’s platform leverages open protocols such as Telehash (private communication), JOSE (smart contracting), TMesh (mesh networking), and BitTorrent (over the air firmware and management updates) to ensure digital trust for frictionless M2M transactions.

Value for customers: The decentralized approach can significantly improve M2M communication security, reduce infrastructure costs, and broaden application cases. Paid-as-use model will also reduce ops expenses while blockchain-based technology will secure and simplify data exchange.

Filament has an impressive management team with rich history in technical entrepreneurship. CTO Jeremie Miller is experienced in decentralized computing by authoring the XMPP protocol and founded instant messenger Jabber. CEO Eric Jennings is a serial entrepreneur and software architect.

No public financial data is available yet.

Exit opportunities include going IPO in 5–7 years and acquisition by incumbent entrants to decentralized platform, such as IBM, Microsoft and Verizon (Verizon ventures invest in Filament).

Description:
Coverwallet is an online enterprise insurance management platform that makes it easy for small businesses to choose and purchase commercial insurance plans.

Funding history: $2 million seed funding in March 2016, unknown pre-money valuation

Investment thesis
· SMB insurance is a billion dollar market ($103Bn direct written premiums in 2013) with strong growth potential and low penetration by technology.
· The founding team has solid backgrounds in consulting as well as product and business development. The advisory board consists of insurance leaders and influential entrepreneurs.
· Competition has not yet become fierce and incumbents are moving slowly towards digital solutions.

Future outlook/Exits
· Precedent transactions suggest positive value appreciation in later rounds if Coverwallet can successfully hit milestones (Next Insurance: $13M of seed funding at $31M pre-money valuation; CoverHound: $33.3 million in Series C at $69.87Mn pre-money valuation)

· Exit strategies:
o Acquisition by insurance broker such as Navion Insurance Associates
o One or two further rounds of VC funding before IPO

Risks/Questions
· Whether the company can control customer acquisition costs (CAC) and retention rate
· Whether the technology-enabled approach can lower expenses in customer services
· Whether the company can maintain strong relationship with insurance companies

Founded: 2012

Headquarter: Denver, CO

Funding: Seed round $875k in Feb 2016

Investment Thesis

Spex (FieldTek, Inc) is a mobile digital inspection platform that brings automation, efficiency, and data to the ecosystem of insurance carriers, independent adjusters (IAs), and construction firms. The value proposition of Spex’s digital solution is threefold:

o For insurance carriers, Spex creates an effective archive of property information and reduces cycle times and associated Loss Adjustment Expenses. It also brings transparency to the performance of service providers.

o For IAs, Spex offers integrated tools (Spikes for laser measurement) that enables accurate data capture in flexible formats such as photos, videos and, audio. It then enhances collaboration in the cloud services it provides.

o For contractors, SpexReports makes report generation automatic and standardized, and can be used for sales purposes, bid/estimate approval, and ongoing collaboration with the IAs.

As an ancillary benefit, large amount of data generated on the platform can further be used in business analytics products, thus providing more value-added services for all parties. Spex brings the same form of SaaS cloud-based workfoce solution that has improved business logistics for many industries, to one that has not yet seen such enhancement. The next funding round is expected in 6–10 months, at which time current trends dictate a series A MRR of $50k (2014) — $160k (early 2016). Based on Spex’s current pricing, the Louisiana flood of 2016 in which nearly 100,000 homeowners were affected could have generated $104k MRR alone (assuming the project takes 24 months to finish, 1 report per homeowner).

Property-casualty (P&C) insurance is a billion-dollar market in the US with low digital penetration and declining customer satisfaction. In 2015, net premiums totaled $519.8 billion and $15.2 billion were paid out in property losses related to catastrophes. Success in this industry strongly depends on operational performance. Of 2,500 P&C insurance companies in the US, the largest only has less than 10% of market share. With slow growth expectation (3% CAGR from 2014–2020) and a highly-fragmented market, participants in P&C insurance space are now emphasizing technical investment to improve profitability, leaving a huge potential market for productivity software. The founding team of Spex is comprised of experienced entrepreneurs with decades of experience in insurance, tech, and sales poised to fill this critical demand.

Exit Agenda

· IPO

o From 2012–2016, the median amount raised by a SaaS company going public is around $108 million, while LTM revenue at time of IPO is ~$80M (~$6M MRR). To reach this target solely on post-disaster report generation revenue, Spex should generate 266k reports month. Given that over 300 triggered natural disasters are registered in the US each year, Spex only needs to obtain less than 5% market share.

· M&A

o Spex would be a strategic acquisition target for SaaS companies that seek for vertical market opportunities insurance or construction.

Risks and Questions

· Customer acquisition costs (CAC). IAs and construction firms are mostly SMBs, meaning high CAC and low switching costs. Spex needs to develop a proven cost-efficient customer acquisition channel and launch high value-added features to its product.

o As a new and niche market, competition is still nascent. Happy Inspector (https://happyco.com/) provides a software to inspect, monitor, and analyze business assets. It graduated from 500 Startups and has raised $3.7M to date. Although the features are similar, Spex has a clearer value proposition by targeting the P&C insurance industry and fosters a higher barrier by connecting insurance carriers, IAs, and construction companies. Targeted marketing strategy has brought early traction. Current customers include insurance carrier Family Mutual, IA NCC, and restoration institute Claim Assist Canada (CAC).

o Early adopters are already enjoying significant improvements in efficiency. Current customers include insurance carrier Family Mutual, IA NCC, and restoration institute Claim Assist Canada (CAC). CAC has been able to increase its volume of inspections during Fort McMurray from five per day to 25 per day in part through the implementation of Spex.

· SaaS subscription business model relies heavily on demand.

o The model is flexible for adjustment to extract profit targets. Spex allows new customers a 2-month trial, followed with a basic plan that charges $10 per SpexReport, 25 minimum per month. After attracting more customers on board, the product could be expanded to fixed + on-demand pricing with team management and BI features.

· Unproven ability to develop a data analysis product. Spex needs to hire a high quality engineering executive to develop BI features to help retain customers.

Sources:

· http://www.iii.org/fact-statistic/industry-overview

· 2016 global insurance market insights, McKinsey report

· J.D. Power 2016 U.S. Property Claims Satisfaction Study

· http://www.ey.com/Publication/vwLUAssets/ey-2015-global-insurance-outlook/$FILE/ey-2015-global-insurance-outlook.pdf

· https://www.cbinsights.com/blog/insurance-tech-startups-investment-growth

https://blog.intercom.io/a-closer-look-at-saas-valuations

本文是我在量化全球商学院（微信公众号搜索“量化全球商学院”）做的线上分享，整理后发布于此。转载请注明出处。谢谢！

这两年互联网金融在中美两国都如火如荼，借着在美国学习和工作的机会，我有幸跟硅谷的很多从业者和投资机构对美国互联网行业有所涉猎。国内关于美国互联网金融领域的研究已经很多，我才疏学浅，班门弄斧，还请大家不吝赐教。

我今天分享还是以概况为主，分享提纲如下：

一． 美国fintech领域主要分支及关键词

二． 几个主要领域的中美异同

三． 对美国fintech行业常见误解

四． Q&A

一． 美国fintech领域主要分支及关键词

在美国，互联网金融对应的词是fintech. 也就是用技术来提高金融服务的效率。国内谈到硅谷的技术总是觉得比国内粗放式的“创新”更高端，其实之后我会谈到，互联网金融，fintech,本质都是金融。我个人感觉，即使在美国，很多VC衡量fintech创业公司价值，更多还是会从金融的本质出发，尤其是地产、保险这些本身业态和监管体系就非常复杂的分支。

先简单说说美国fintech的整体格局。这是一个我觉得分类比较清晰的图。

根据现在国内行业的发展情况，我选几个跟中国比较有可比性的行业重点说说，通过关键词总结一下各个分支的发展重点：Lending, Personal Finance,保险。

1. Lending: 机构化

在美国，真正意义上的P2P lending,也就是个人对个人的借贷，其实只有大家熟知的Lending club, Prosper两家。其他的平台，都需要你是机构投资者，专业投资机构或者accredited investor,即具备投资资质的个人（不仅是高净值而已），而不是像国内这样面向大众的理财工具。即使是lending Club, 个人vs机构投资者的比例比起IPO时候的3:7，从金额上来说，现在比例已经反转了。有意思的是，所有平台都在努力维持平衡，因为机构投资者单笔投资额较大，单个客户流失对平台影响更大。加上新的平台如Avant等都是balancesheet lending,Fintech领域最顶尖的投资人，Foundation Capital的Charles Moldow就指出，Marketplace lending这个称呼更能体现online lending行业的本质。

机构化的另一个表现是，从2015年的Lendit峰会就看出来，比起“颠覆”银行，lending公司更多寻求与银行的合作。他们的角色也更多是“提高银行某个已有业务的效率”。我觉得这张图是很能说明这个现象的。

Lending公司与银行的合作，与国内打包和证券化银行资产不同，主要还是获取客户。因为他们获取客户成本主要是邮件（不是电子邮件！）成本很高。而通过与银行，尤其是社区银行的合作，获客成本就会降低很多。资产端却并没有“机构化”：美国的online lending还是以个人贷款为主（消费贷款、学生贷款、mortgage等），而不像国内有很多变相处理银行信贷资产或者企业债权的、本质是“影子银行”的业务。主要原因我觉得是美国银行证券化业务已经发展得很成熟。之后谈到中美异同的时候会多说一些。总之，lending这个领域在美国已经有consolidate的趋势，不论是subprime, student loan还是传统消费贷款的分支，不论是资产段还是投资端，个人认为，做平台/lender的创业公司大的机会并不多，但是像lendingrobot, Orchard, PeerIQ这样为marketplace lending提供跨平台交易、数据分析、交易支持等服务的公司还是有一些机会。

2. Personal Finance: HENRY

Personal finance包括自动化投资平台/robot-advisor，如betterman, wealthfront, 等， 也包括个人财富管理，比如credit karma, nerdwallet (类似credit Karma, 但是针对更高净值客户），online broker 如robinhood。HENRY是Highly educated, not rich yet的人群的简称。可以理解为毕业不久的青年才俊，未来的中产阶级。这是这个行业几乎所有创业公司target的客户。总体而言，这部分人群的财富目前占所有财富管理行业比重还并不大。大多数中产阶级的财富还是在传统的银行和资产管理公司、投资顾问公司，比如Vangard。虽然行业leader是有（从管理资产规模来说，wealthfront和betterman，personal finance三家算是遥遥领先），但是现在也还比较分散。

坦率来说，虽然量化投资平台号称以技术取胜，wealthfront还有重金聘请了华尔街和学术界很有盛名的泰斗级人物，但是出发点和战略都并无二致：低门槛（免费管理一定数额，总体费率低），针对年轻人群，操作界面容易，提供增值服务（主要是税收优化）。

目前，只有wealthfront小范围提供直接投资股票的服务，其他的自动投资平台都是投资ETF等基金产品。motif investment走的是社交投资，但是最近增长没有W，B家快。我个人认为这个模式本身就不太符合美国这边HENRY人群追求简单直接的投资偏好。相对于国内，他们对收益要求并不很高，而且这一批未达到五百万美元以上可投资净资产的人群，本身积极投资者就很少，在美国投资ETF已经很成熟的背景下，wealthfront这种消极投资（就是说，长期来看，没有人能战胜市场，wealthfront的目标就是30年内的年均收益与大市持平）为基本理念的自动投资工具，消费者教育成本还是可以接受的。下图是wealthfront和国内的真融宝投资组合对比：

3. 保险：Millennialing （80/90后“化”？）

保险可以说是在前两个领域涌现了一批独角兽、热潮逐渐过去之后，最近fintech领域最受关注的分支。millennials可以理解为中国的八零后九零后群体。millennialing是说，把传统复杂陈旧的保险系统，变得更符合80,90后人群的使用需求。美国是一个保险行业极为发达的国家，不仅是个人保险（健康险，寿险），财产险（车险，房屋保险），还是企业保险等等，已经成为美国人生活不可获取的一部分。但是这个领域门槛也非常高，所以保险的创新即使在美国也是雷声大雨点小。总体来看，我跟很多这边的投资人交流过，大家都觉得，目前这个领域还远没有当年lending club刚出现的时候那种真正结构性的，颠覆性的创新。很多创业公司并不具备保险资质，都是通过与保险公司合作，优化用户体验和underwriting model. 但是仅仅是把保险变得更user friendly，产品更年轻化，市场就已经很大。足见这个领域空间之大。

保险领域的创新主要是三个方面：distribution/销售渠道, 产品和结构。

销售渠道比较好理解，越是标准化的产品，如车险，越容易线上销售，如美国的车险比价平台就有Coverhound等。

从产品来说，还是在于如何扩大数据来源，优化定价模型。优化数据来源的公司就很多了。车险：这个是最为成熟的，比如OBD就有metromile, Automatic等。但是空间不会太大，因为车险本身很标准化，创业公司没有自己设计产品的能力。 健康险：很多保险公司和wearable企业，比如fitbit, misfit合作，甚至自己推出wearables (humanus+fitbit)。还有之前我看过一个中国人创业公司，做智能坐垫的，也有保险公司主动寻求合作。但是都是小打小闹，也许会一直以合作方式或者被保险公司收购，但是独角兽级的机会我目前还没有看到，难点在于保险产品的设计需要深厚的专业知识，这本身就成为创业公司的门槛。

结构性创新：P2P insurance,比如产品还没发布就融资千万的lemonade，算是为数不多的一个“结构性”上创新的模式，但是我个人不是特别看好，因为真正有需求且能够“组团”的群体有限，怎样“组团”的标准/算法还在探索阶段，而且这个模式似乎不符合简单化、便捷化的初衷。

保险领域我个人对企业保险很感兴趣。险种创新上比如近年兴起的网络安全险，还有国内悟空保针对新的互联网商业模式的创新险种，做得很赞。渠道上有为中小企业提供保险规划服务的平台比如next insurance, coverwallet。但是总体而言，保险领域的“令人激动的颠覆性创新模式”，我还没有发现。这一点我是很想了解国内的情况的。

二. 几个主要领域的中美异同

刚才其实已经陆续涉及了一些，这里再补充几点吧：

1. 借贷领域

这个领域大家可能都比较熟悉。差异主要在

• 资产：美国资产主要以个人贷款（包括学生贷款）为主，中小企业贷款一方面因为银行体系相对发达，另一方面，数据unstructured,我跟做中小企业贷款的ondeck，以及做供应链金融的fundbox, C2FO聊过，非结构化数据的模型也仍在发展之中。而Kabbage，自己没有Amazon，阿里巴巴，京东这样的电商体系，所以发展瓶颈也非常明显。学生贷款好像国内也有公司在做，但是美国这块市场的确太大。因为资本主义学费、申请费都贵得惊人。一般中产阶级都是负担不起的。学生贷款公司一般的路径是：refinance传统渠道的学生贷款（利率7–13%，类似lending club一开始refinance信用卡），然后通过资产证券化等低成本的融资方式，从利率差和手续费赚钱。数据模型：这个我想放到后面的“常见误解”中讨论，企业贷款信用模型我研究不多，但是个人信用模型，我个人感觉美国的lending公司远没有大家想象的那么data-driven…
• 融资：大家都说中国p2p lending是影子银行，这点没有错，也没啥不好。因为中国的信贷资产证券化发展还很慢，银行很多优质信贷资产都还没有最大效率的盘活。个人投资者的风险承担能力是一回事，但是证券化需求是真实的，这个商业模式就没有错，只是还没有成熟。其实通过资产证券化融资对于个人贷款来说，利差并不够大。因为美国个人消费贷款其实违约率并不低。

2. 互联网金融创新与银行的关系

我觉得美国产品形式上的创新并不多，基本都是银行现有产品的“查漏补缺”，锦上添花。前面的那个unbundling banks的图也说明这个问题。比如payday loan,就是一种针对subprime人群的紧急贷款。他们遇到疾病或者临时的重大开支，支付不起就申请这种贷款，在下个月拿到薪水以后还钱。这种贷款在银行APR能达到200甚至300%。fintech公司做这种贷款，其实并没有更大创新，只是可以以比银行更低的运营成本，加上一些灵活的差异化定价，把总体利率降低一些。比如一家叫ellevate的，去年本来要IPO，但是因为市场等原因推迟了。从收入上看规模发展很快，但是$300M的贷款就要将近150M charge-offs，18M provision。另一个角度，APR降到了150%这样，需求还是这么大。可见美国成熟的银行体系下，underserved的人群还是很多。但是这部分人群的风险很高，最终利润率并没有比其他贷款种类有明显优势。

3. 传统金融体系

我觉得中美最大的不同，也是中国最大的机会，就是传统金融体系的力量。有几个对比很能说明问题：一个是信用卡渗透率，一个是保险渗透率。

数据表明，中国消费者对于网络投资和宣传的接受程度是最高的。正如印度跨过了固定电话阶段直接进入手机时代，中国互联网金融也跨越了PC时代，在移动互联网的发展上一骑绝尘。

而在美国，不论是传统理财产品，中小企业借贷，还是保险，都已经有极为成熟的巨头和极为稳固的消费习惯。要想改变这些习惯，成本远比什么发实体邮件的获客成本高多了。而且HENRY群体真正成为最具有消费和投资能力的中产阶级家庭，也需要至少五年的时间。这个窗口，我觉得也正是中国互联网金融的机会所在。

三. 对美国fintech行业常见误解

时间原因，我就说一点。

美国fintech公司真的那么tech-driven吗？

我最常用的例子还是lending。跟中国不一样，美国很多个人信息都是比较完善的，银行更是信息的最大拥有者。大多数lending公司的模型还是极为依赖FICO score，比如lending club就明确了借款人FICO的底线（虽然已经从IPO时候的680一路降到620……）。而有FICO score的人群，本身已经不属于最underserved的群体了。模型真的是online lending核心竞争力吗？是，也不是。监管红利的确是一大因素。比如美国的银行，credit model里面，很多“有歧视”的指标都不能用，（比如哪个学校的，性别，种族等等），而且模型的重大改动还要通过监管机构的审查。这个灵活度，对于fintech公司，已经是很大的红利。稍微小一些的student loan公司，Upstart, Earnest, Commonbond，声称自己还纳入Linkedin profile, GPA, social network等等，但是扒开了看，FICO还是权重最大的。前面提到的Elevate，还有跟京东合作的Zestfinance，是真正在数据模型上的创新者。因为他们target的客户，真的是高风险的，银行不愿接触的客户，信息也少。但是从Elevate IPO的推迟，我们也不难看到风控之难。

风控上的“非数据”创新靠什么呢？Sofi是一个最好的例子。他们现在完全没有违约。首先，从选择消费者开始，他们只给最top的学校的最好找工作的专业（MBA, Law, engineering等，人文社会学科免谈……）的学生贷款，这个pool真的不大，现在他们借款人数还不到3万。其次，他们构建一个社区，增加借款人的线下曝光度和依赖度。从alumni network，到求职帮助，最近开展了创业支持项目，等等。你说你接受了alumni的工作推荐或者创业指导，还好意思不还钱么？所以sofi现在不需要多么强大的engineering团队（从他们招聘重点就看出来了），只要把线下社区运营好，抓住最顶尖，风险最小的用户群，延伸产品线，从一开始的student loan开始，到mortgage, consumer loan, auto loan…最大化每个用户的价值。

现在尚未被大规模影响的金融领域，都是硬骨头。比如保险，比如房地产这些行业还是需要深刻的行业经验和理解，用户体验的痛点大家都知道，但是整个生态的复杂性远不是一个所谓颠覆性的算法或者模型能够解决的。我们跟创业者聊，如果团队里没有金融从业者，对市场痛点都抓不到，技术真的都是其次了。不厌其烦地说，fintech本质还是金融。这一点，在哪里都是一样的。

金融本身就是一个数据driven的行业。诚然，这么多年的领先和积累下，美国整体数据条件是比中国优越很多，不论是可以用于个人贷款的三大信用信息机构和FICO，还是传统金融体系（银行、保险等）里大量的业务数据，都比中国发达很多，为互联网金融创业公司提供了很好的数据基础。但是在技术和消费习惯等大小环境都加速变化的今天，“前互联网时代”积累的数据价值或许被高估了。那些曾经被认为很先进的定价模型远没有挖掘出数据真正的价值，在今天看来显得这么迟钝粗糙。我们看到，没有这些历史legacy限制的中国企业，反而能更快地适应互联网时代的需要，更灵活更大胆地创新。有投机有急躁，但是这种高速的试错迭代，才是真正创新的姿态。

四. Q&A

Q1: 不过随着e租宝这些重大非法集资案发，互联网金融名声也收到很大影响，在美国，这种P2P跑路的现象概率多大？

A:坦率说，这个情况并不多。我觉得原因是多方面的，并不是美国人多高尚这么简单的问题。首先，我记得数据上看，大多数跑路都是蓄意欺诈，而不是平台运营管理问题，所以出问题的并不是这个商业模式; 其次，美国金融投资品种多，整体环境成熟，投资者不会蜂拥到一个新的投资产品上; 而且，这跟前面说的投资者结构有很大关系。美国投资者相对比较成熟，尤其是accredited investors。这就意味着，不会有那么多傻瓜式的钱一下子涌进来，投机成本大; 当然还有就是美国整体个人信息体系还是比国内完善很多，对个人而言跑路风险太高了。

中国这些资产还是很不透明。在美国，做真正面向个人投资者的P2P平台成本非常高,披露要求也很高。你去lending club, prosper网站上，可以看到他们所有贷款人的数据，每个人至少40项。因为P2P产品被认定为证券，平台每周都要给SEC上报loan sales report。然后还要各个州的license,不像国内拿一个牌照，全国都能做生意。

Q2:美国互联网金融是由证监会监管，但严格来说lending club 或prosper只是信息中介撮合平台，并非证券发行人（即便以广义的证券定义），那么SEC也应该是监管放款人吧，请问SEC监管lending club 或prosper的逻辑是什么？

A: 这个问题当年监管机构是让两个平台暂停业务来探讨的。可以把这些平台理解为证券交易所, 比如NASDAQ。SEC认定，在这些平台上发行的贷款，已经具备了证券的性质，所以应该按照证券交易所的规定来监管。但是并不是所有fintech都是SEC监管的，很多是消费者保护机构来监管，从保护投资者和借款人的角度出发。核心是，面向所有个人投资者（就是散户）开放了。而且，这种方式的借贷需要在每个州分别取得借贷业务牌照（中央ZF好！）对于创业公司成本很高。这就是为什么新的marketplace lending公司都不走这个路线。

Q3: 有研究过真融宝吗？这种把货币基金、债券、债权、股票型基金等资产动态配置在一起，设计成活期理财的模式，在国外是否有类似的平台？国外监管政策对资金池是如何管理的？

A: 这个我觉得很有意思的，投资跟社交领域很像，美国人似乎不大喜欢大而全的平台。一方面，涉及这样多样性的资产管理，还是有很多纯技术无法解决的，所以互联网公司比起传统的理财顾问公司，vanguard这样的，优势并不大，而且如果你自己设计产品组合来卖，就需要资产管理的牌照，每年交License fee，交report，成本还是高。此外，这个量级的财富管理，一般中产及以上才有需求了，银行已经很好的覆盖了。

Q4: wealthfront这种智能投顾，实际上也是在做多资产动态配置，更加个性化些，大众实际积极性如何？还是更多是概念？

A: wealthfront的确发展很快，跟他们用户抓的准，定价策略好有关。申请的时候会问你各种问题确定你的风险偏好和承受能力，通过这些信息自动生成个性化投资策略。推广方式是，5000美元以下免费管理，推荐一个人就免费多管理5000美元。我觉得大众积极性跟投资习惯有很大关系。wealthfront理论核心是消极投资，正如前面提到，就是长期来看，没人能战胜市场。他们做到的就是30年内的收益贴合market return,这就需要消费者也有比较成熟的长期投资理念。我听说，最近整体市场不好，他们平台收益也不好，所以一些客户如果投资资产不多，很容易就withdraw了。不管怎样，对于刚工作不久的人群，这个程度的服务似乎能够满足要求了。这个领域已经进入品牌竞争阶段，不过金融服务本质也是这样的。

结语：

最后，关于互联网金融创新的未来，某次我听到一个略有激进的说法，供诸君探讨：只要使用fintech服务还需要连接银行账户，互联网金融创新就还没有结束。

今天跟大家交流很开心，谢谢大家！

欢迎对互联网金融感兴趣的同行一起讨论

Email: monicaxie.aiesec@gmail.com

wechat: monicaxie

Originally published at

https://www.linkedin.com/pulse/fintech-across-us-china-overview-why-you-should-scared-monica-xie