The New Age of Insurance Aggregators

Tech innovation is coming to insurance, but where and when it strikes is uneven.  Auto and health insurance have been facing serious disruption, for instance, but for very different reasons (self-driving cars and telematics, vs. the ACA and hospital mega-mergers). Life insurance and commercial P&C are only now feeling it. Reinsurance and annuities are following behind.

To see trends, then, it can be instructive to focus on specific insurance functions rather than the type (market vertical).

Distribution — that is, sales and marketing — is one area that has been especially active compared with other functions such as underwriting, risk, investments, admins/support or claims.

Why disrupt distribution?

 It’s where the money is. In general, when a P&C or Life insurer gets $1 in premium, 40 cents goes to distribution (marketing/sales costs, i.e. the agent) and 50 cents goes to everything else (underwriting, claims, service/support, risk, fraud, product, executives, etc.). Only 10 cents is profit. The largest distribution cost is usually agent commissions, which range from 50% to 130% of a policy’s first year premium.

It’s easy for carriers to work with alternative distribution channels. Insurance carriers are used to third-party distribution. They have been using independent agents, wholesale agents, and affiliates (e.g., sales through AARP) for years. Systems are already in place to easily take on new distribution outlets.

The rise of insurance aggregators

Aggregators are simply comparison shopping sites — like for insurance. They allow consumer to easily compare product features, carriers, coverage and price.  They aren’t the only distribution disruptors, but new developments are making them more potent.

Comparison sites come in three general flavors: Lead generators, call-center based agencies and digital agencies. From their websites, it can be difficult to tell them apart, but they operate differently and appeal to different investors.

Lead generators — such as InsWeb, NetQuote and — use a comparison shopping format to entice insurance shoppers to provide personal information. They then sell these leads, often to traditional brick-and-mortar insurance agencies. Lead generators specialize in either gathering lots of leads cheaply, or curating data to sell fewer but higher-value customer referrals. Lead generation is a specialized technique, an art even. But it’s mostly unrelated to emerging technology. It is difficult for non-lead-gen experts to assess the quality and sustainability of lead-gen platforms.

Call-center agencies can develop leads or purchase them, but their call centers employ licensed insurance agents who make sales.  A classic example: SelectQuote, founded in 1985 and known for its early TV ads, is now the largest direct channel for life insurance. Goji is doing this well in the auto insurance space and now has several hundred licensed call center insurance agents. Call-center agencies are also great businesses; their core competency, however, isn’t technology, but HR hiring and training. Hiring and training a large sales force and managing its churn is growth-limiting. A commissioned call center sales force might reach 300 agents, but it is almost impossible to get to 1,000 or beyond high quality agents. Crucially, too, the model does not leverage technology so margins are reliant on commissions remaining high.

Digital agencies allow customers to shop and buy entirely (or nearly so) online. Without a human sales force, they must create well-thought-out user experiences that make the process of buying insurance transparent and understandable. Typically this requires building sophisticated interfaces into multiple carriers’ systems so that the customer experience is unrelated to the company selected. In general, digital agencies hire developer talent rather than sales talent. Esurance, one of the first successes in the space, was bought by AllState for $1 billion in 2011.  A more recent example (and AXA Strategic Ventures portfolio company) is PolicyGenius, which is bringing this digital model to life and disability insurance.

I believe digital agencies are the future. They focus on technology to sell policies without the aid of a commissioned human agent. This is a crucial distinction because while both call-center agencies and digital agencies generate income from commissions, call-center agencies have to split that commission with their licensed call center agent while digital agencies do not.  This means that at scale, digital agencies should have higher profit margins.  In fact, it is likely that digital agencies will actually look to lower commissions to drive sales and to provide more competitive pricing than human agents or call-center agencies.

Insurance Aggregators 3.3.16.png

Insights from the UK

The UK insurance markets adopted the aggregator model earlier than U.S. carriers, which gives us a window into their potential future. What’s happened there shows us three things.

First, aggregators have captured a material share of the insurance market and are continuing to grow. A recent Accenture survey found that in the UK, “aggregators account for 60 to 70 percent of new business premiums in the private automobile insurance market” and french aggregators have seen “18% average annual growth for past 5 years“. We’re already starting to see this in the U.S.: Oliver Wyman recently reported that “the number of insurance policies sold online has grown more than 400% over the last eight years.” Swiss Re saidmore than half of consumers say they are likely to use comparison websites to help make purchase decisions about insurance in the future.”

Second, aggregators eventually will compete with one another across all personal lines of insurance. U.S. digital agencies today focus on one type of insurance (life, or auto, or home) though they might claim to offer a few others. But almost all UK aggregators compete across all personal line insurance products.

Third, insurance carriers will get into the aggregator game.   Two of the top three UK aggregators sold to UK carriers; GoCompare was purchased by esure and was purchased by the Admiral Group.  Only MoneySuperMarket remains independent.  And carriers are looking to create aggregators from scratch. Accenture’s report noted that 83% of UK carriers are “considering setting up their own aggregator sites”.

Startups, tech giants and carriers in the ring

Still, it’s not totally clear how digital distribution will play out here. Multiple startup digital agencies have raised significant capital.  PolicyGenius and Coverhound have raised more than $70 million, $50 million of it just in the last six months. This represents a fraction of what a major tech player (say Google or Amazon) could put toward an effort to enter this market. Interesting, Google purchased the UK aggregator BeatThatQuote in 2011 and launched the California auto insurance aggregator Google Compare less than a year ago, but just announced it was shutting it down.  It could be because Google Compare functioned as lead-gen for CoverHound and Google decided the fee they received per lead was cannibalizing their ~$50 per click ad-sense revenue they received from auto insurance search terms.

Long-term success in insurance, requires focus, deep knowledge of the industry, and deep knowledge of the consumer. Insurance is very different from most e-commerce products and Google’s experience could be indicative of the difficulty big digital brands will have trying to crack the insurance aggregator market.

Finally, most large American carriers haven’t decided what to do. Purchasing an aggregator creates strange incentives, potentially driving customers to a competitor. At the same time, it also gives the insurer the opportunity to quietly select the risk they want to keep and pass off the risk they’d prefer to give up. Progressive has had mixed success.

Final thoughts

I think the U.S. will see trends and dynamics similar to the UK’s, and soon. Within three years the major digital agencies will start to compete fiercely, and within five, one or more will have been purchased by a carrier.

More digital agencies also will tackle the complex insurance products: annuities, permanent life insurance and commercial insurance. Right now, startups are trying — Abaris for annuities and  Insureon, CoverWallet for commercial insurance — but their offerings aren’t yet as developed as Policy Genius or CoverHound.

Finally, I think the rise of digital financial advice platforms (a.k.a, robo-advisors or “robos”) give digital insurance agencies an interesting channel to consumers that will help at least one of them mature and grow to an IPO.

I asked a two digital agency CEOs what they thought the future was going to bring

Jennifer Fitzgerald, PolicyGenius’ CEO said “Consumers are much more self-directed in the digital age, so the focus is giving potential insurance clients the tools they need — instant and accurate quotes, transparent product recommendations, educational resources — so they can go through the process at their own pace. Then it’s important be there for them with an intuitive, easy-to-use platform and service when they’re ready to buy.  That’s the basis for the new wave of insurance education tools like the PolicyGenius Insurance Calculator, and is reshaping how consumers look at insurance.

Matt Carey from Abaris said “I think we’ll soon see a new wave of made-for-online products. Carriers have always gone to great lengths to create products that made sense for a specific channel.  The Internet will be no different. In our business, that probably means very simple lifetime income products that are subscription-based and have low minimums. Until then, I don’t think we’ll reach a tipping point in the migration from offline to online

Coming up next: Reader’s choice! Below are three topics. Comment or email your vote!

  • Disrupting Product: Why a “WebBank” for insurance is needed, and difficult, but doable


  • P2P insurance: Why is everyone drinking the…Lemonade


  • Micro Insurance: Difficult unit economics and blurring the line between Debt and Insurance
The New Age of Insurance Aggregators


The $400 billion legal market has attracted little venture capital investment and remained arguably under-disrupted by technology. Even as other professions — notably financial advisers and marketing professionals — are being transformed by new technologies, tech companies haven’t found as many opportunities in the legal space, where many time-consuming tasks that demand human intelligence have defied automation.

That’s starting to change. Key advances in two technologies are hatching a new Legal Tech market. They are:

MACHINE LEARNING: ML is a form of artificial intelligence that uses algorithms that recognize patterns and adapt in order to create increasingly accurate predictions. Consumers experience basic models of ML when Netflix recommends movies to subscribers, or Pandora selects songs, and it is also at the heart of using sensor data to control self-driving cars.

NATURAL LANGUAGE PROCESSING: NLP systems, another form of artificial intelligence, can understand and use a spoken or written language such as English, rather than a specialized computer language such as C++.

NLP and ML together can understand documents like a human, but with high-speed computational power and vast digital memory. This has been used for years in applications like the iPhone’s Siri and IBM’s Watson.

But now, startups can quickly and cheaply deploy these tools. AWS and other cloud computing platforms have made computational power cheap and ubiquitous. NLP libraries have improved to the point that non-NLP developers can use them.

Already we’re seeing interesting applications from early stage startups, including automated personal assistants (JulieDesk, X.AI), finance tools (Social Alpha, AlphaSense), marketing (Macromeasures, EncoreAlert) and language generation (Easyop).

It’s the big law firms, however, that are poised to be irreversibly changed as startups enter its field:

Text IQ: Using NLP, ML, and social network mapping, Text IQ helps determine privilege in eDiscovery. For example, if a text message between two employees says, “Joe said we can do it, if we get approval”.  That text is privileged and does not need to be turned over to opposing counsel if “Joe” is determined to be that specific Joe from the general counsel’s office.

RossIntelligence: Built on IBM’s hefty Watson platform, Ross does deep legal research. Instead of a team of associates looking through a curated list of precedents, Ross could, in theory, look at all cases related to a particular law and pull out the most relevant passages for their review.  Further, ROSS notifies lawyers of new court decisions that can affect their cases.

Counselytics: This service determines how above/below market a lease agreement is, or how closely a contract compares to historical deals. For instance, it might report, “The lease is out of market as these X terms are greater than 90% of all contracts in the database.”

LitIQ: This program highlights ambiguities within contracts or other legal documents to prevent drafting errors and disagreements. For example, if a contract says “X and Y or Z,” LitIQ asks if the user meant (X and Y) or Z or X and (Y or Z). The company’s goal is to be able to determine if there are logic conflicts between paragraphs within a legal document.

eBrevia: Using technology developed at Columbia University, their software extracts and summarizes key provisions from legal documents. The main uses of the software are for due diligence, contract management, lease abstraction, and document drafting.

Surveying this field, two things stand out. First, even though they all rely on the same core technology and target the same market, there is little to no overlap. Second, most of these companies are targeting the work currently done by high-paid associates at big law firms. Will the upshot be that associates do less mind-numbing work, or that law firms will need significantly fewer associates?

So what does this all mean?

These early stage startups have not yet proved the tech can live up to its potential, but none has flamed out, either.

If these show success, expect more startups to enter the legal space and more focused attention from venture capital to seed them.

Most of these startups must sell to the major law firms, and as a result, those big firms have the power to dramatically retard their growth speed. But if a few firms find the technology provides competitive advantage, the whole market quickly will follow suit (if only to protect themselves against malpractice claims).

My theory is that Big Law is about to change. Firms that are proactive in implementing new technology have an opportunity to increase productivity and market share. Those that resist will face significant pain in five to 10 years, when they will be playing catch-up.

 TLDR: A fusion of two technologies—natural language processing and machine learning—mean computers can understand written documents and analyze them for patterns and inconsistencies. This holds tremendous promise for a burgeoning Legal Tech sector, where a bunch of startups are automating work traditionally done by armies of high-paid associate lawyers.