Working for a few years in the ever-evolving insurtech industry, I have had the opportunity to simultaneously oversee the technological and business aspects of insurance. This has allowed me to grasp the intricate yet straightforward nature of this industry. After hearing Varun’s honest(podcast linked below) and straight-from-heart reflection on the Insurance landscape which was packed with learnings and insights, I couldn’t hold but start typing my experience about the space less from a technology but more from a product lens which is a birds eye view of my experience.
When a few smart and intelligent folks from the tech industry asked me what did I do, to simplify my job of explaining and still trying to be as close as possible, I explained “We sell insurance as e-commerce products similar to Amazon where the product is a promise summarized in a PDF document” and I got such weird looks that I had to make it savvy and complicated for them to say ‘now it makes sense’.
Insurance is a long-standing industry that emerged when trade ships began transporting people and goods. At that time, the industry was new and fraught with risks, as there was a chance that the ships might not reach their destination. In order to safeguard the families of sailors in such unfortunate events, entire villages would pool together small amounts of money to pay the sailor’s family (now referred to as term insurance).
Insurance is a business of selling promises
Insurance, fundamentally, is a commitment by the company you have entrusted with a portion of your safety funds, known as the premium, to uphold their promise and provide assistance in the event of a negative occurrence, also referred to as a risk. The promise is a contract documented that acts as proof a.k.a policy schedule which you receive when you pay the premium. The key point here is that the premium is a fraction of the benefit you receive i.e. you are expected to pay X as a premium and when something bad happens you receive typically 100 times X. So where does the 99 come from? It comes from a science of probability which is called actuarial science combined with mathematical modeling of what and how many can go wrong, which in itself is a specialized study of the probability of every aspect of risk, i.e. risk in weather, health, vehicle, human behavior, transport, you name it and a probability model either already exists which is constantly updated with real data to enrich the model or is newly created. At this point, it is sufficient to understand actuarial science is at the heart of insurance product pricing.
Since insurance is not a tangible product, like an iPhone or an experience like a concert, or a service like a hotel stay or air travel or a digital product like an app, humans inherently are not attracted to it. In some economies, the government mandates it as it is a humanity or public hazard when something bad happens and welfare or rehabilitation becomes a govt responsibility they don’t have the resources to do it and rather expect the community to help each other, which goes back to the origin of Insurance. Now that was quite a bit of tangent we went into but an important one explaining the cliche of insurance being a push-product vs. a pull-product.
Channel
Insurance is a non-intuitive product or service, and due to the nature of the terms of conditions of the product, note that what you are buying is a contractual promise written on a PDF document which sometimes goes into 2 to 3 dozen pages, there is a need for it to be explained by an expert who is trained in the legal and financial aspect of the product. They are termed agents or brokers, who are expected to protect the interest of the customer. The capitalist nature of every business tends to maximize profits by selling products even though they might not meet the needs of the customer which is rampant in the financial services sector hence there is a lot of regulatory scrutiny in here by making agents or brokers held accountable for miss-selling. I will stop apologizing for digressions, as it seems necessary at some points. So the Agent Channel is the traditional form of selling an insurance product which is still a huge % of the distribution channel.
The rest of the channels being , Direct-to-Consumer (D2C), and Embedded are some of the popular channels of distribution
Distribution is the holy grail of Insurance
We have spoken about how insurance is a push product (although I hate this framing, I am yet to prove it otherwise), hence distribution i.e. appropriate use of a channel to make the buyer pay for the product is critical to the business. One who is able to crack the distribution of insurance is able to survive. Having been deeply involved in business building, I have come to realize this is true for every business, digital or otherwise. I would any day bet on a distribution lead growth than any other strategy out there.
We will focus more on technology lead aka non-traditional distribution channels in this writeup. Digital distribution can be broadly categorized into D2C and embedded.
Embedded insurance
D2C because it takes the core insurance product i.e. actuary-driven pricing risk underwriting directly to the customer bypassing the agent.
Embedded is the flavor of D2C, which means it highly contextualizes the risk with a primary-highly sought-after anchor product or service i.e. hotel booking, flight booking, a gadget, logistics/delivery, a concert or event. Think of all the ways each product or experience can be ruined and they have a focused actuary model to price the risk.
Let us take an example of a concert or paid event. Let’s consider the most talked about event when we are writing this, 25 July 2023. Taylor Swift is performing in Singapore. Tickets will cost an average of S$800. The key assumption here is no one is going to purchase the Taylor Swift concert, paying $800 upfront with the intention of cancelling it. Typical risk associated with that can ruin the experience is; 1. You fall sick, 2. There is a significant event in the family, 3. You stay in Malaysia and travel to SG by road or air and it gets canceled, delayed etc, 4. Bla, 5.Blu, 6. Blee. This is where the creativity of the risk underwrites lies where sometimes the boundary is blurred between which risk event can and cannot be underwritten. I can go on and on on how the risk underwriting can impact the profitability of the book and how the model needs to be real-time pricing the risk eg #1 and #2 are focussed risks to individuals and the probability of claims happening can be a smaller fraction as compared to #3 and hence the cover can be priced accordingly.
The creativity of the product manager and the actuary is at its best when packing the cover and attaching it at the point of sale of the anchor product.
Note how the push product is converted to a pull product as the cover is highly contextualized and the distribution challenge is fixed as the insurance premium payment is made along with with primary-highly-sought-after anchor product or service.
Traditional D2C
Traditional D2C distribution is basically making the entire buying experience online on a browser or app as the technology stack varies along with complexity, scalability and reusability. Yes, this might sound duhh 🤦 to most, but this is a reality in a large-scale insurance distribution business as the manufacturers and distributors are different and the industry is regulated as compared to other e-commerce products. D2C distribution started with lead generation i.e. getting interest from users from aggregator platforms1 2, financial knowledge & affiliate websites and then directing them to the carriers or insurers whose core business is underwriting the risk and following the regulation. All the “tech” in insurance and finance happens in the distribution layer which tries to innvoate and improve the customer experience while the core regulated bodies like insurerers and banks try to follow the heavy regulation and protect the core business of underwriting risk and loans which is what the regulator mandates it for.
While distributors and carriers aka manufacturers were 2 separate entities traditionally, the lines are blurring with distributors building the risk management muscle with a captive corpus. To keep it simple cash-rich companies set aside some part of their reserves as insurance payouts and pay premiums to themselves. A wider umbrella can be formed for allied businesses. Financial creativity can be at its best sometimes in insurance business1
With distributors solving large-scale, high volume, high-performance application building parts, read silicon valley muscle, they partner with traditional insurers like Chubbs, Zurichs, and Libertys of the world for the risk book building part. While they both need each other, they also try to play the partner’s role in building captives and in-house tech teams. While both need each other commission sharing is a tenet that makes the partnership interesting. Distributors take typically 40-60% of the premium and it goes higher with more traditional insurers which intend to play a negligible role in technology i.e they prefer to run their insurance book with a mainframe computer whose only job is to handle m/billions of daily transactions daily and not fail – which is mainfriam’s only job and rely on technology partners for distribution. They focus on the regulatory and book profitability aspect of the business and hence have the noose of the risk underwriting and have the final word to accept or reject the risk. More tech-savvy carriers eh listed above 😉 have multiple distribution partners as they expose their products with a Policy Administration System(PAS) with SFTP’s imports from the lower end to REST/GraphQL endpoints to the higher end, and hence have good negotiating power with the distribution partners.
Some D2C platforms have a superior customer experience where the quoting, comparing, binding(make payment/commitment) and getting the end product, a bunch of PDFs (lol), is a breeze and equate an e-commerce experience with an excellent chat-bot, read ChatGPT powered, which lucidly explains all the subtle jargons of inclusions, exclusions, asterisk, fine-print, rider, co-pay, deductible, tiered-limit, illness-limit, room-limit and all the neverending caveats insurance companies impose on a product.
Since insurance is a promise aka trust-based business the earlier the trust is earned, smaller is the cycle of closing the transaction and the higher is the chance of return, referral and retention. This is highly applicable in General Insurance where the customer experiences the product at different parts(eg risk cover for different product categories on e-com site) and platforms(same distributor offering bespoke cover on airline, hotel, e-com sites etc) Building trust is just going to make the experience smoother and faster.
There is also a not-so-subtle nuance in offering the insurance product on a digital platform, more so for embedded and less for D2C, about how is the insurance product perceived in comparison to the anchor product, read Liberty insurance providing motor insurance during Tesla checkout, Tesla as a brand is owning the insurance product buying experience while Liberty is playing the part of promise seller, there is so much of brand, UI-UX, claim experience etc that comes into play while building the partnership and integration and imagine the complexity when there is a 3rd party technology distribution partner who carries liberty auto-product and builds partnership with Tesla and other car manufacturers to sell their auto insurance products on distributors platform. An interesting problem to solve.
Book Profitability aka Loss Ratio
This term is the bread and butter of the insurance Industry. As the risk underwriter, you collect premiums from the customer. As discussed the premiums typically are 1/100th of the benefit payout. In the Taylor Swift example above, I can think of $50 as a premium for an $800 ticket (my actuary friends are going to kill me for simplifying this as the number) and imagine the number of claims that will be needed to make the book unprofitable. Simple envelope math, 1000 people buy the risk cover, premium collected is $50,000 – the book, which boils down to ~62 claims made which can wipe the book assuming the payout is $800 for every claim (again the creativity of the product, business and actuary is at its best here to make the claim amount partial to reduce the pain of the customer and not make them whole).
Note that loss ratios aka claim ratio tracking is a core function of the risk businesses that decide to underwrite the risk or not typically in cases where the claims have the potential to wipe the whole book in case of flood, wildfires etc where a large part of the city, village are wiped due to one natural calamity.
There exist Re-insurers for this purpose, where there is an insurer for an insurer. More on this later.
Productization of Insurance
Traditional Product Management principles are rooted in making sure the metrics like engagement – DAU, MAU, retention, conversion funnel, user experience – NPS, feedback, pain points and most importantly an experimentation mindset to make sure the key metrics are tracked and goals achieved. Facebook, Dropbox and Airbnbs of the world (Product Marketing is a great move by Airbnb and makes sense for all consumer-facing products) have mastered it and bringing them to the insurance world and blending them with aggressive actuarial teams to build the right kind of innovative product for the most-sought-after anchor product is a masterstroke for digital distribution. The mindset and skills needed for embedded and D2C channels are slightly different as the former needs more collaboration and negotiation skills to engage with the product and engineering world of the anchor product while the latter has better control over the customer journey until payment is made. The leading metrics of a D2C funnel are
- lead-to-app (L2A), where the metric tracks the performance of cost of leads generation either by SEO/SEM CPC to landing page to exact type of insurance product. It all depends on how deep is the start of application. Think start of app as a first CTA of intended product
- app-to-quote (A2Q), this is the next part of the funnel where all the required steps of the application are performed to get the quote. This is usually the most important part where the price of the insurance product is acquired and usually, the customer is shopping for the best price for the best coverage. Depending on the price and type of coverage a customer usually takes 1 to 8 days to come back and complete the payment if the product is rightly priced for the coverage the customer is looking for. Again this is where the product and actuary can be creative.
- quote-to-bind (Q2B), this is where the rubber meets the road payment is made and key metric is fulfilled. A supreme online payment experience is key here. Credit/debit cards, local bank transfers, mobile wallets, installments, cheque payments (customers come in all forms and ages) and the lesser the reasons to not make payments better it is for a push product business(not again..lol). A digital transaction is complete when the policy documents, mandated by regulation, are provided to the customer which concludes the transaction.
For an e-commerce store, you can break down the funnel similarly eg. lead2cat1, cat12catt2, cat22cat3, search2skuN, skuN2chkout where cat1,2,3 are categories or hierarchies of your inventory aka sku, and skuN2chkout being the critical funnel goal. Experimentation starts to optimize this funnel by understanding the persona of your customers and how they interact with your platform. A good visualization of a generic funnel can be seen below.
Mostly, distribution-focused businesses emphasize new business conversion, while a customer-focused carrier/insurer also prioritizes the claims and payout experience, which serves as a crucial differentiator in the insurance industry.
A typical d2c metric-driven business would look something like this: The business does a competitive landscape study of which products are in demand or required or mandated by a new regulation or is due renewal to poach from the competitor, the list is long. Once the product is identified say a travel-trip protection which pays 80% of the claim where the premium is decided by the capping of the 100% claim amount i.e I buy a $5000 product where the claims will be paid up to $4000 for a trip cancelation due to standard critical & verifiable reasons then you get 80% of the claim up to $4000 with a max premium of $80.
Now the business has decided to do a $5M GWP(gross written premium) from this product in year 1, which comes to the sale of 62,500 policies. Now we need to make rough industry estimates of funnel conversion of l2a of ~70%, a2q of ~45% and q2b of ~30%, They are standard insurance industry benchmarks that vary on product, region, market, time of the year etc. Now doing the math, 15% of bind equates to 62,500, hence ~ 208,000 customers need to quote, and 462,000 customers need to be warm leads for which we need 661,000 leads. Now comes the strategy of how good the lead generation strategy is and the quality of the lead. Getting leads from big tech would be foolhardy as the CPC for insurance is the highest! In simple words, you need to bid $54.91 to rank on the first page of google search with a mediocre SEM strategy that relies on ‘Insurance’ as the keyword to get leads, and they are not even
CPC affects your return on investment (ROI)
Funnel optimization is a key factor that makes or breaks digital distribution channels. I remember when we were building a pet insurance proposition for a developing Asian market, we brainstormed for a week and broke down a 20-questionnaire pre-quote form into 4 questions to improve our a2q metric. We distilled the conversion problem into smaller chunks and fought with the traditional “regulation doesn’t allow us” mindset to think out of the box.
It costs 7 time more to attract a new customer than it does to retain an existing one
Neil Patel
GWP is a vanity metric, typically for a distributor’s business whose only job is to distribute and not care about loss-ratio or renewals. Real business is made in top-line revenue and book profitability which should be the bedrock of a distributor-carrier partnership.
A significant portion of the businesses also relies on efficient renewals where there is regulatory and risk governance needed to capture updates to re-underwrite the risk during renewals. While 40-60% of the first distribution sale commission is shared with the distribution partner, renewal commisions are a fraction of it in most insurance lines except health and auto for which reasons are not fully known to me. Real money is made in claim-free renewals where the insurer rewards no claims.
If you liked what you read, do listen to this excellent interview with Varun Dua, founder of Acko Insurance.