The Marshmallow founding team and insurance designed around migration

A specific underserved customer problem can support a defensible insurance business when it changes underwriting, product design and distribution together. Marshmallow did not merely market ordinary motor cover with welcoming language. Its founding proposition was that people who had moved to the UK could be misunderstood when a conventional process placed too much weight on a short UK record and too little on relevant experience elsewhere.

That focus gave Alexander Kent-Braham, Oliver Kent-Braham and David Goate a clear place to begin. It did not make the risk easy. The company still had to determine which overseas evidence was reliable, price uncertain claims exposure, explain cover, serve customers and operate within insurance regulation. The founder lesson is therefore narrower than "find a neglected audience". It is to identify exactly where the standard assessment loses information, then build the operating capability required to use better information responsibly.

The problem was a mismatch between a customer and a proxy

Marshmallow's official company history says the three founders started the business in 2017 after observing that people who move to the UK can pay more for car insurance. The company's stated premise is that relevant driving experience may exist even when it does not appear in a conventional UK history.

The current company page describes Marshmallow as UK-based, with offices in London and Budapest. London is therefore one operating base; the customer proposition and insurance structure are not London-only.

That is an important distinction. A newly arrived driver is not the same thing as a new driver, but a dataset designed around domestic records can make the two look similar. When a useful fact is hard to observe, an insurer may rely on a simpler proxy. The proxy can be defensible at portfolio level and still describe an individual poorly.

The commercial opening sits in that gap. A specialist can ask for evidence that a general process does not collect, design verification around it and communicate clearly to the people who recognise the problem. The insurer gains a potential information advantage, while the customer gains a route to present a fuller history. Neither benefit is automatic. An overseas licence, claims record or residency history varies in format and verifiability, so every additional input creates operational and fraud questions as well as insight.

Better underwriting means making foreign context usable

Insurance underwriting turns available evidence into an estimate of future loss. For customers who have crossed borders, the difficult work is not simply acquiring more data. It is deciding whether different records are comparable, current and attributable to the applicant.

A credible migration-aware model needs documented rules for accepted evidence, missing information and conflicts between sources. It must separate facts that predict insured risk from facts that merely correlate with nationality or income. It also needs monitoring after policies are written. If a new variable appears useful during development but does not predict claims in live cohorts, the insurer has to revise the rule rather than defend the original narrative.

The same discipline applies to customer communication. Applicants should know what evidence can support their application and what happens if it cannot be verified. A product built for newcomers should not quietly penalise the people whose home-country records are least machine-readable. Manual review can help, but only if the review standard is consistent and its effect is measured.

Becoming a carrier increased both control and responsibility

Marshmallow says it became a licensed insurance carrier in 2021. Its current company explanation says Marshmallow Financial Services Ltd arranges and sells policies and is regulated by the Financial Conduct Authority, while Marshmallow Insurance Ltd underwrites policies and handles claims. The 2025 solvency and financial condition report identifies Marshmallow Insurance Ltd as supervised by the Gibraltar Financial Services Commission and permitted to underwrite specified classes of business in the United Kingdom.

Moving further into the insurance chain can strengthen the original proposition. A distributor can optimise an application and find an insurer willing to accept the risk. A carrier can have more direct control over underwriting appetite, pricing, claims learning and product changes. That can shorten the loop between a hypothesis about overlooked evidence and the actual performance of policies.

Control brings exposure. Premium has to be adequate for claims, expenses and capital needs. Reserving assumptions can prove wrong years after a slick customer journey launches. Claims handling is the moment at which policy wording becomes a real service rather than a sales promise. Capital and reinsurance arrangements can constrain growth even when demand is strong.

This is why the founding insight should not be assessed only through acquisition or quotation metrics. The evidence eventually has to include loss performance, complaints, renewals, cancellations and claims experience across well-defined cohorts. Growth can reveal a real market, but it can also postpone the moment at which inadequate pricing becomes visible.

A precise audience made distribution intelligible

"People who have moved to the UK" is a recognisable audience with a recognisable frustration. That makes the proposition easier to explain than a generic promise to reinvent insurance. Search terms, referrals and partnerships can connect the company with customers at the moment they are trying to establish ordinary life in a new country.

Specificity can improve product decisions too. Documents may need explanation for people unfamiliar with UK insurance conventions. Support needs to distinguish language accessibility from regulatory disclosure. Claims processes must work for customers whose supporting records or repair history span jurisdictions. These are not brand accessories. They determine whether the advertised focus survives contact with the customer journey.

The risk is that an initially clear audience becomes a slogan as the company expands. A broader product range can create useful cross-selling and reduce dependence on one line, but each new line has different underwriting evidence and claims behaviour. The migration lens should continue only where it produces a concrete product or assessment difference. Otherwise, the business should describe the product on its own merits.

The current team record requires careful wording

Marshmallow's current careers site lists Alexander and Oliver Kent-Braham as co-chief executives. Its official history names David Goate alongside them as a founder but does not assign him a present executive role. Companies House separately records Alexander and Oliver as active directors of the original technology company.

These sources support a precise description: all three are co-founders; Alexander and Oliver are currently presented by the company as co-chief executives. They do not support guessing a present title for Goate. Group structures can also change, and a role at the parent is not identical to a directorship at every regulated subsidiary. This draft therefore avoids using one entity's officer record as a shortcut for the entire operating group.

The distinction matters because founder profiles often freeze an early team chart. Readers deciding whom to hire, fund or contact need current responsibilities rather than an origin story copied forward. The roles should be checked again immediately before publication.

What another founder can take from the model

Start with the information failure, not the demographic label. Write down what the standard process cannot see, why the missing fact matters to risk and how it can be verified. Then test whether the proposed signal improves decisions after accounting for fraud, missing data, servicing cost and claims.

Next, align the operating model with the promise. If the business needs underwriting control to use its insight, a pure distribution relationship may eventually constrain it. If it becomes a carrier, capital, governance and claims expertise need to grow before volume, not after it. Finally, measure inclusion as an outcome. A targeted proposition should be assessed by who can obtain appropriate cover, how customers are treated through claims and whether pricing remains sustainable.

Marshmallow's story is useful because migration was not an abstract mission statement. It pointed to a specific weakness in the information used to assess a customer. The durable advantage, if it exists, will come from repeatedly turning that observation into sound insured outcomes.

Limitations and questions for reporting

This draft relies on current company pages, public corporate records and an insurer disclosure. It does not include interviews with the founders, policyholders, claimants, brokers, reinsurers or regulators. It does not independently test quotation prices, underwriting variables, claims acceptance, loss ratios or whether different groups receive comparable outcomes. Company explanations of its origin and customer purpose are primary evidence of strategy, not independent proof of performance.

Before publication, Shoreditch Talk should ask the company which overseas evidence it currently uses, how unsupported or incomplete histories are treated and how it tests for unfair outcomes. It should seek cohort-level retention, complaint and claims information with definitions, and compare sample quotations only under a reviewed methodology. Regulatory roles and the legal entities responsible for broking, underwriting and claims should be reconfirmed. No conclusion about fairness or superior pricing should be added without that evidence.