Nov 30, 2018 · 24m · a16z
Managing Risk and Uncertainty: The Future of Insurance
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this a16z presentation, General Partner Angela Strange explores the evolution of the $5 trillion insurance industry, explaining how real-time data, modern underwriting, and AI-driven claims processing are overcoming legacy barriers to revolutionize how risk is managed and priced.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 0.0 Guest teaching 5.5 Guest disagreement 0.3 The host pushing back 0.0
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 5:50 Bucket of cold water on startup ease
Angela forcefully dispels the notion that starting an insurance company is simple, highlighting massive incumbent ad budgets and regulatory moats.
Hardest push from the host ▶ 0:00 No host pushback presentThe transcript is a single-speaker keynote monologue with no host interactions or spoken pushback.
Biggest teaching moment ▶ 14:40 Health IQ Facebook mortality trackingAngela educates the audience on an unconventional data source where Health IQ tracks Facebook memorializations to prove lower mortality to reinsurers.
The host holds their own ▶ 0:00 No host expertise demonstratedThe host does not speak in this monologue transcript, preventing any host expertise counter-demonstrations.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| Human Nature Around Risk and the Need for Protection | 0 | 5 | 0 | 0 | This is a solo keynote monologue presentation by Angela Strange without active host dialogue. The host-side scores are set to 0. Angela educates the audience on human nature around risk pooling and the historical coffee shop origins of insurance. | |
| Calamities as Demand Drivers and Early Claims Management | 0 | 5 | 0 | 0 | Monologue continuation detailing how historical calamities like the London Fire created demand for insurance and led to early claims management like private fire fleets. Host is absent from transcript. | |
| High Barriers to Entry: Marketing Costs, Low Engagement, and Adverse Selection | 0 | 6 | 1 | 0 | Angela explains the massive structural barriers to entry in insurance including high marketing spend and regulatory hurdles. Minimal combativeness is present, expressed as humorous critique of incumbent advantages. | |
| Catalysts for Disruption: Evolving Consumer Expectations and Cyber Risks | 0 | 5 | 1 | 0 | Angela highlights rising cyber risks and consumer expectations for on-demand services, sharing a lighthearted story about her personal telematics tracking showing bottom 1 percent driving skills. | |
| Modern Customer Acquisition Strategies: Niche Groups and Specialization | 0 | 6 | 0 | 0 | Monologue focusing on modern acquisition and underwriting strategies, detailing how companies like Health IQ leverage social media data for lower mortality verification. | |
| Innovative Risk Pooling: Peer Networks, Cyber Tools, and Micro-Insurance | 0 | 6 | 0 | 0 | Angela presents examples of peer-to-peer risk pooling, cyber tools freemium distribution, and satellite-based crop insurance for smallholder farmers. | |
| Modernizing Claims: Behavioral Psychology, Charity Givebacks, and AI Fraud Detection | 0 | 6 | 0 | 0 | Angela details behavioral psychology in claims processing like Lemonade charity giveback models and machine learning fraud detection through Shift Technology. | |
| Investment Thesis and the Future of Personalized Insurance | 0 | 5 | 0 | 0 | Angela summarizes the venture capital investment thesis across customer acquisition, underwriting, and claims automation, predicting rapid innovation in personalized risk scoring. |
Statements from this episode (28)
Crop pooling by early farmers is the fundamental concept of insurance
“Farmers used to pool their crops at the end of the season. So they would choose to give up a higher than average yield in good years in order to mitigate loss in down years. And this is the fundamental concept behind insurance.”
The world's first financial insurance market operated out of a coffee shop
“The first insurance market was actually a coffee shop.”
A Scottish clergyman created the first modern insurance fund
“The first insurance fund as we know it today, where we pay differentiated premiums into a pool, and then that pool is reinvested, was developed by a Scottish clergyman.”
The UK government once sold life annuities at uniform prices
“Rewind a few hundred years, and the UK government used to sell life annuities at the same price, whether you were 20 or whether you were 40.”
The global insurance industry is worth $5 trillion
“If I asked you all to guess the size of the insurance industry, you probably wouldn't get up to Five trillion dollars. This is the global industry.”
The United States insurance market is worth nearly $1.5 trillion
“In the US alone, it's almost one and a half trillion dollars.”
One in ten Fortune 500 companies is an insurance firm
“In the Fortune 501 in 10 companies is an insurance company.”
Primary insurers offload catastrophic liabilities to reinsurance companies and retrocessionaires
“Insurance companies offload to reinsurance companies. Then this actually keeps going, that offload to retrocessionaires, and it spreads out even further across the private and even backstopped in some of the public markets.”
The US government operates fundamentally as a giant insurance company
“And he stated that the insured, the government can be thought of as a giant insurance company with a sideline business in national defense. And Homeland Security.”
No ground-up insurance startup has reached the top tier in 80 years
“In terms of a ground up new entrance, this hasn't happened in over 80 years.”
Geico spends over $1 billion annually on advertising
“Geico spends more than a billion dollars a year advertising to you.”
Insurance ad spend accounts for $50 billion of Google's market cap
“We did some quick math, and apparently fifty million dollars of Google's, or fifty billion dollars of Google's market cap, can be attributed to insurance.”
Long feedback loops prevent lending-style underwriting experiments in insurance
“In most categories of insurance, the dollar volumes are higher, and the feedback loops are a lot longer, and so you can't use the same technique.”
US insurance lacks a federal regulator, forcing startups into state-by-state licensing
“No surprise, all financial services is regulated, but the special challenge with insurance is that there's no federal regulator. And so by definition, a new company has to go state by state by state, which just costs more time and takes more money.”
Most US insurance outside of auto is sold through brokers
“Lastly, with the exception of auto, most insurance is sold via a broker.”
Navigation apps like Waze replace self-reported commute data for auto insurance
“What's the distance between your home and work? This used to be offline and self-reported. Now, Waze, Google Maps, any number of apps probably know it even better than you do.”
An OBD dongle placed Angela Strange in the bottom 1% of drivers
“I happily stuck mine in, drove around for a couple of weeks, and then I got this email at the end of two weeks which, can you not, said, Angela Strange, you are in the bottom one percent of drivers.”
28 million US small businesses require general liability and workers' compensation insurance
“There are twenty-eight million small to medium businesses in the US that are all gonna need general liability insurance, workers' comp insurance, a whole variety of other types of insurance.”
Traditional home insurers require applicants to answer 40 to 60 questions
“It won't surprise you to learn that, on average, companies ask you 40 to 60 questions, and they go something like, when was your home built? How old is your roof? Have you had it fixed? What's the kind of material?”
Digital insurer Hippo underwrites home policies using a single question
“Now, with all of this data online, a company like Hippo can come in, and they've aggregated dozens of different data sources, and even brought in some new data sources, such as their underwriting for home insurance is not 40 questions, it's one question. What'…”
The majority of commercial property insurance claims involve roof damage
“Wouldn't surprise you to learn that the bulk of claims in commercial property have to do with the roof.”
Health IQ uses Facebook data to prove health-conscious applicants die less often
“Then, what Health IQ does is they've had millions of people that have filled out this quiz, and this is where this gets just a little bit morbid. People who fill out the quiz, it's attached to your Facebook profile, and when you die on Facebook, your status is…”
Two-thirds of all cyber attacks target small and midsize businesses
“What might surprise you to learn, though, is that two thirds of cyber attacks are on small to medium sized businesses.”
An estimated 500 million farmers globally lack crop insurance
“You don't have to go very far afar into many, many other countries in the world such that by some estimates, five hundred million farmers don't have crop insurance.”
Lemonade pools policyholders by charity to deter claims fraud
“Lemonade collects premiums like every other insurance company. It takes a fixed percentage of those premiums, puts them aside, and that's Lemonade's profit. Then the rest of the money is going to go towards claims. But versus most other insurance companies, th…”
Shift Technology uses machine learning to detect anomalous insurance claims
“And here's where companies like Shift Technologies come in. They'll partner with probably many of the insurance companies in this room, suck in all of your claims data, self-reported data, millions of different pictures, and run fairly sophisticated ML algorit…”
Automating claims below fixed dollar thresholds quickly triggers fraud gaming
“Let's say claims under 500 dollars, I decide they're just not really worth the human effort into reviewing them, and I decide to automate everything under 500 dollars. Very quickly, fraudsters will figure out that that's the threshold, and you'll start seeing …”