What Wins In Fintech: Distribution Or Data?
The future of innovation is global. We discuss it here.
This piece originally appeared on my Forbes column here.
Coming out of two big fintech conferences this month (Insuretech Connect and Money 2020), it’s clear that fintech is evolving – but it’s not clear which evolutionary approach will dominate. Startup innovation seemingly bifurcates around a choice: to either build towards either a distribution advantage or a data advantage in the insurance industry.
In 2020, I wrote about the unique attributes of a successful fintech company, or the “3 Ds”:– distribution, data, and delivery. I argued that successful startups had at least one of the three, and particularly one of the first two: distribution or data. The best had more than one. Some even had a trifecta of all three.
But which of the Ds is most important? Which will lead to more consistent multi-billion-dollar startup outcomes?
Let’s start with a few considerations. I promise I will provide an answer at the end of the article.
Is the customer hard to reach?
Some customer bases are easy to reach via easily accessible channels like social media or online television. Over 85% of millennials make purchases online and influencers, reviews, and social recommendations are a big driver to decisions. Others are easy to reach via readily built existing channels – think the broker channel for car or auto insurance.
Some audiences are harder to reach. The elderly may not be on social media. Medium-sized businesses face more esoteric risks like climate business insurance. You get the idea.
As a broad simplification: when the customer base is easier to reach (and willing to switch), a long-term distribution advantage matters less. When the customer is hard to reach, a distribution advantage is key.
Is the product specialized or is it a commodity?
Certain products have well understood parameters and dimensions. They are easily comparable between companies. Car insurance and bank accounts are clear examples. These of course also tend to be more easy to distribute (e.g. online or via established channels).
A distribution advantage in commoditized products is harder to achieve. The playing field can be leveled in online acquisition (e.g. bank accounts) or broker channels (e.g. car insurance). That’s why brand matters a lot. No surprise, to get awareness, Geico spends $2 billion on marketing every year.
In more commoditized products, a data advantage can be used to build an edge. For example, companies like Root promised to underwrite based on differentiated data (driving behavior). But unless the new data creates a massive underwriting advantage in commoditized categories, ultimately distribution still matters. This allows specialized players to better price the customers it seeks, and gain market share.
More specialized products will allow providers to exert greater pricing power. No surprise specialty insurance lines have much lower loss ratios and higher profitability.
There are of course several nuances here. Is there a willingness to experiment with new products? What do switching costs look like (e.g. switching bank accounts and credit cards is challenging because of auto-pays creating stickiness)? How important is brand loyalty?
Is the market changing?
In a changing world, new risks and new needs evolve. Some are on the horizon today, notably cyber and climate.
In insurance, new risks lead to new questions: how will losses manifest themselves? How big will they be? Who will be affected? What behaviors today will shift losses in the future?
Unfortunately, these are massive black holes without clear answers.
If the product were available at affordable prices, often customers would clamor for it to mitigate this uncertainty. But if priced incorrectly, they could create huge challenges for the insurer. That’s why data in uncertain situations matters more.
That’s one of the reasons parametric climate is on the rise. As Nick Cavanaugh, the CEO of Sensible Weather explains it: “The availability and fidelity of remote sensed data – increasingly originating from satellites – combined with highly resolved computational models and scalable data processing architectures has made many parametric products feasible for the first time. Purely data-driven risk products can now provide accurate coverages while dramatically increasing cost and operating efficiency." Parametric simplifies and controls the risk equation (e.g. Descartes in the corporate space and Sensible Weather in travel). But ultimately, these companies are built on a data advantage.
Profit margin for product
Some products have low margins. For example, the average loss rates in auto insurance are between 60-70% (and in some cases over 100%). For ACA healthcare plans, it is mandated to be 80%. Other categories like extended warranty insurance are far more lucrative, with 50-60% profit margins, inclusive of loss and also management expense!
When the margin is lower, the margin of safety is as well. As a result, data matters more in underwriting to ensure that in low margins profits can exist.
Conversely, when margins are high there is room for error. There needs to be data, but through distribution, with a large enough margin for error, the data set can be built over time.
The role of regulation
Some products are more or less regulated. For example, in home insurance, there are restrictions on how much an insurer can increase pricing year over year. If you’re in a region with changing weather patterns (e.g. California fires or Florida floods) - or mispriced your policy for any reason - it makes it far harder and costlier to fix the mistake. In ACA plans there is a minimum loss ratio of 80%. If you don’t hit it, you are penalized.
Without diving into the benefits and trade-offs of the regulation (generally speaking, I am for consumer protection), the more limits of regulation there are on pricing and pricing modification, the more data matters.
Embedded financial services
Embedded financial services – by selling a financial product as part of a broader offering – has a built-in distribution advantage. This is the core value proposition. So by nature, the distribution advantage of the original product or company matters the most.
Embedded fintech also has a twist. It can enhance or improve the original product. Spot insurance includes health insurance as part of a ski lift ticket. In the event of an injury, the care experience is smoother and integrated (and free).
And if the embedded insurance offering helps improve sales conversion, the parent business can monetize in different ways (regardless of the profitability of the insurance product). For lending, this is one of the key incentives merchants have to implement buy-now-pay-later.
So which 'D' matters most?
The unsatisfactory answer of course is that it depends.
In my role as a venture capitalist, I gravitate towards businesses with unique distribution advantages, but where a data moat can be built over time through experience and scale. This is one of the advantages of embedded financial services for instance, as well as emerging risk categories with great potential for dislocations (and creation of multi-billion dollar businesses). These include new risk areas (e.g. cyber) or changing ones (e.g. climate).
However, your answer to the same question will depend on your strategy and business model.
Where do you land?
Thanks for reading [99%Tech]! Subscribe for free to receive new posts and support my work.
What Wins In Fintech: Distribution Or Data?