Startup-specific examples Flashcards
What did you learn from establishing your own startup?
- It is very difficult to easily articulate innovative concepts in a way that people understand what you’re selling (VC’s, customers). Then switching to technical detail for insurers
- You need to be open to completely changing what you’re building (not be wed to your concept)
- Building something that people will buy B2C is exponentially more difficult when you’re going from zero to one and have no corporate backing. You’re most likely to be successful through partnering
- Getting validation of problem and solution early without expending capital
What do you think the main challenges Genworth will face in starting new businesses?
- Validating early and building a culture that’s OK with learning from failure, constantly changing, etc
- Defining a repeatable model so that we can place a number of bets and not start from scratch each time (e.g. Eucalyptus)
- Genworth has had success B2B but looking at B2C is very different (benefit of CBA etc..)
How did you experiment/prove demand? How did you pivot?
Important to spend a lot of time validating both the problem and solution. PRODUCT MARKET FIT.
Spent most time validating the problem, through market research including a lot of customer interviews and surveys, as well as a lot of indirect research.
Solution validation included:
- overall demand through a landing page and wait-list which had a few hundred registrations in the month that it was up
- detailed solution design example: the MVP target market was going to be cyclists, and one of the features we were testing was how to reduce cost through having on-demand cover. The proposal was that it could be triggered by turning on a Strava ride for example and you only pay an insurance premium when you’re on your ride. We just did basic prototyping of a screen and walked through the process and found that people thought they were more likely to be injured in a car accident, so were worried about the period they weren’t covered. They also didn’t like the friction of an additional acceptance in the process.
What are some blockers and set-backs you had with the startup?
- Rejection from an underwriting partner
My startup concept was double-sided in that I was designing and distributing embedded insurance products but wasn’t the underwriter (similar to Open or Cover Genius or Honey). I had a prioritised list of preferred underwriters to work with. My number one was a large reinsurer and I had managed to pitch to a few of their senior leaders who were then working on reviewing my commercials. Unfortunately during this time two things happened - 1) they found that they had an internal issue where their life insurer couldn’t form distribution agreements outside another branch of their business, and 2) that other branch of their business change their strategy to focus on their existing customer relationships for 12 months
- Biggest blocker I had was finding a technical co-founder. I met a number of potential co-founders but none of them could code the MVP
Give an example of something you had to influence on
One of the problems with insurers that we’re all aware of is customers don’t trust them to pay claims - Deloitte 68% of people said insurers will do whatever they can to avoid paying. The cause of this is misaligned interests.
I had tried to tackle this by researching commercial models where insurers don’t profit from declining claims. There are a few of these - Lemonade and Laka.
The biggest hurdle was trying to convince underwriters that it’s possible to generate substantial sustainable profit without relying on claims experience as a factor. I did 3 things:
- I built out the commercial model myself, using a lot of assumptions, and sent it to them to assess and get their head around.
- The model was dependent on achieving certain sales targets and operating costs. So I also surveyed a lot of people to understand whether they were likely to take out cover with an insurer that was not benefiting from a declined claim.
- I also researched other international startups with these models and referred to their success
Offered an alternative that MyParachute would donate any profit margin from claims experience but that insurer didn’t have to, they would just have to forego any right to decision on declines
How do you go about assessing the viability of a strategic partnership?
Obviously Genworth is looking at various approaches, including strategic partnerships but also potentially building solutions in-house.
I think the approach to take is largely dependent on the problems and opportunities we’re assessing.
In terms of the solution space, is this something we actually could build in-house from a capability perspective, would we want to have full ownership, is speed to market going to be an issue?
In terms of assessing a startup as a solution - team, product, market, business model. Also, how aligned are they with Genworth on things like the value of the proposition, how open to collaboration are they, how feasible is potential integration in future?
Specific example of assessing potential partnership for me was actually the underwriter - company/reputation, AFSL, global presence, product scope, relationships, support startups?
How would you approach getting from zero to one
I think Lean startup methodologies are really useful in this stage to make sure you don’t spend too much capital building something customers don’t actually want. So starting with your business model canvas and prioritising the validation of the problem, solution and riskiest assumptions, before getting into agile development and iterative MVP build.
Examples of having done this in practice are obviously MyParachute. I spent a lot of time validating the problem and fortunately a lot of work had been done in this space. When it came to validating the solution - EXAMPLE (prototype of “turn it on” cover). Learnt 1) too much friction and 2) people believe they’re more likely to be injured in a car accident.
In terms of proving traction I’ve done things like landing pages, wait lists, surveys, interviews.
The next stage of actually starting build of the MVP - the main examples that come to mind are when building the AI-enabled claims management system as ResLife. I had general targets and problems to address and paramaters to work within - broadly reducing operating costs. First step was to prioritise the most impact, identify generally what we might build and establish the capabilities to get going. Prototypes AI models that categorised claims, provided rehab recommendations and extracted relevant information to then make liability decisions. Learnt a lot about the most important data and the investment to get that data was substantial.
What experience do you have in setting up new digital channels/G2M?
On establishing new digital experiences:
- Claims transformation at ResLife, building an entirely new digital front end. This is for customers going through some of the hardest times in their lives, who already don’t trust you.
Important to understand their needs and behaviours and we did a lot of prototyping and customer testing including A/B to understand this.
- The biggest insights were: 1) if customers thought the same question was being asked multiple times they thought we were trying to trick them, 2) if they didnt know the exact answer they’d drop out. 3) It was really difficult to get all the data our AI models needed up front to automate the process, while not asking too much. Compared to a person speaking in multiple conversations.
On go to market:
- Winning an award for the Life Insurance Framework reforms. Built a digital product that enabled advisers to understand the impact of reduced commission on their business
- Note that at the same time we had built a new very unique insurnace product and I thought that was the award we were getting.
Tell me about a time you innovated to come up with a unique concept
Business model for MyParachute - solving the problem that people don’t trust insurers. How might we remove the inherent misalignment? How might I do this without being the underwriter?
Need to keep marginal operating costs low enough to build the full profit margin into the price, and not rely on the claims experience. Can do this through RPA, digitisation and natural language processing/AI
Tell me the ways you have used emerging technology
The main emerging technology I have worked with is machine learning models and natural language understanding.
Set up the full transformation delivery to build 3 unique models from scratch using a combination of natural language processing, understanding and machine learning models. Worked with both internal teams and vendors to do this.
Biggest challenges were:
- dragging people on the journey (there is apetite - ANZIIF)
- integrating with legacy tech
- access to data that was both accurate and effective
What are some examples of insurtechs/fintechs/startups you admire?
Lemonade. Three reasons:
- that they’ve attempted to build a purpose-driven business model
- they’ve attempted to disrupt customers perception of what insurers should be - their insta account is great it’s just artwork
- they’re AI-enabled and largely automated
A startup I admire for the way they’ve built a repeatable business model is Eucalyptus. I strongly believe in the value of repeatablility and re-use of core capabilities.
Big believe in the power of creating shared value so I love Who gives a crap
Within Proptech - interesting startups:
- Brickfloor - market price guarantee - buys property at guaranteed price. Uses data - “automated valuation model”
- 2Be - 55-75yo’s - 5 year fixed rate lon for customers wanting to help kids into property
Why this product type/problem?
Problem space:
- I’m the target market and it’s something our friendship group have grappled with a lot over the last few years
- I think there’s so much opportunity. The research I’ve done hasn’t shown a lot of disruption in the space so I think it’s ripe for it
- On a purely selfish note I’m also keen to move away from life insurance. I fell into it years ago and it’s something you can become an expert in and be trapped. My roles for the last 4 years couldn’t have been any product and I’m keen to leverage that
How would you go about validating problems and solutions?
Interesting example of validating the problem - level premium rates. Ethnographic research. Customers believed they were something else. Solving it was as simple as educating advisers.
MyParachute example: validating solutions. Too much friction when introducing on demand cover with Strava. Also found that people are more concerned about car accidents
Main trends in fintech/insurtech
- Open data/CDR
- Embedded insurance/banking - API’s - integration
- Web3 and crypto OR AI/automation
