Mike Hudack on Product Simplicity, Experimentation and Building Teams

Sling Money co-founder and CEO Mike Hudack joins Harry Stebbings in September 2024 to discuss product lessons from Facebook, Deliveroo and Monzo. His views reflect his experience and his interests as a founder and investor.

Hudack recalls wanting to build digital things from childhood, leaving high school at 15 to work at a startup. On “founder mode,” he says founders have long gone into details and skipped levels, but skill at doing so can vary with a company’s stage.

He does not expect the phrase itself to produce a dramatic rise in bad behavior: a founder who behaves badly may simply use a new label. He regards it as branding around practices that already existed.

At Facebook, he recalls a flat organization in which Mark Zuckerberg could question a small code change directly. Practical support mattered too: equipment was readily available because an engineer’s focused time was worth more than the cost of a replacement headset.

Joining Facebook’s advertising team after the IPO, Hudack looked for a neglected product that a small team could improve quickly. The team rebuilt Page Insights in a few months, showing colleagues that the organization could ship quality work and setting a higher internal bar.

He favors simplicity: understand what a person wants to achieve and minimize the work needed. An advertiser ultimately wants to sell more products, he says, rather than spend time operating targeting and optimization controls for their own sake.

His preferred product team has roughly six to eight people, mostly engineers, with a designer and, where possible, a data scientist; he considers a product manager optional. Teams should own clear outcomes rather than merely a list of features to ship.

Before coding, he wants a theory of the user, the problem and how the product will help. But real usage will expose errors in that theory, so teams must ship, learn and distinguish a partly wrong idea worth improving from one to abandon.

He says founders often know when something is not working but feel pressure to keep going for investors or colleagues. As an angel investor he values an honest discussion of a pivot or shutdown while resources remain, rather than exhausting them without learning.

At Facebook he regrets prioritizing Audience Insights, an ambitious tool for exploring population-level advertising data. It required difficult technology but proved more of a nice-to-have than a core need, in his assessment.

A feature’s cost includes building, maintaining and explaining it, as well as what the team could have done instead. Hudack now asks whether a proposed feature creates meaningful value rather than merely being interesting.

He does not choose between endless discussion and dictatorial product vision. Teams should state an outcome and estimate the impact of a proposed project; putting a number on the hypothesis makes the disagreement concrete and open to later evaluation.

Product is more art than science in his view, but should be managed with data. A team needs room to pursue an intuition it cannot prove beforehand, then measure the result and use what it learns to refine its judgment.

He credits Ribbit Capital’s Mickey with emphasizing distribution: a product cannot create value for users if it never reaches them. Usage and behavior are also tests of whether a product is valuable, although distribution itself still requires work.

At Sling Money, Hudack says the team was considering encouraging referrals and viral growth before spending heavily on paid marketing. That was a plan at the time, not proof of future growth; he has a direct interest in the company.

Stebbings argues that a narrow initial customer profile helps marketing resonate. Hudack partly agrees but says a product with a universal value proposition can begin broadly if it finds communities with a specific pain and grows from those first successes.

He cites WhatsApp’s early appeal to people facing expensive international or domestic messaging. Once a network reaches enough of a community, he argues, its value can extend beyond the people who first felt that pain.

Deliveroo was a different product environment from Facebook: food orders, rider availability, weather and restaurant capacity changed in real time. Hudack says the need to meet a delivery promise created operational urgency as well as product-development pressure.

When he arrived, he recalls late deliveries and manual opening and closing of geographic zones to manage demand. He says around 40% of orders were late at that point; this is his historical recollection, not an independently verified company metric.

The first response was a small engineering team and a model to estimate lateness. More accurate estimates could allow customers to decide whether to order and reduce the gap between the promised and actual arrival time.

His favorite Deliveroo lesson came from Madrid, where customers wanted restaurants farther away than the existing neighborhood-based system allowed. The team had assumed food should travel only as far as it retained an ideal quality; customers sometimes preferred more choice despite a longer trip.

Hudack argues that customers should be allowed to make informed trade-offs, such as accepting a less-crisp burger. But too many options can overwhelm people, so product teams must choose where to sit between total automation and exposing every control.

Data can also mislead without context. He recalls a Facebook surface that appeared widely opened, but says many openings were accidental and immediately closed. A metric needs the surrounding behavior, not just an attractive headline.

He warns against underestimating competitors. Even when an incumbent’s product is frustrating, its team may be capable and its business valuable; a startup should understand why customers use it rather than assume it will fail.

For consumer features, he recommends controlled experiments with a clear outcome metric, adequate sample size and a comparison group. A feature may improve the target, have no detectable effect or make it worse, and each result calls for a different response.

Moving to Monzo required understanding local financial habits and the responsibilities of a regulated bank. A risk officer helped him see that banking rules protect people and the stability of the financial system, even when they constrain rapid experimentation.

He would have liked Monzo to expand internationally more aggressively and considers the US a substantial opportunity. But he cautions that banking needs, payment patterns and regulation differ by country, so local teams must adapt a product rather than simply transplant it.

Asked about Revolut shipping more features, Hudack says the companies had different philosophies. He opposed crypto products for Monzo’s audience at the time and wanted the bank to strengthen trust and become customers’ primary account; this was his judgment, not a universal conclusion about crypto.

The team first had to decide how to measure whether Monzo was a primary bank. It then built targeted features around that goal, including Flex, which he describes as a credit product integrating installment-style payments with the account.

Hudack prefers a theory about what might work over shipping a mass of unrelated features. Experiments remain useful, but they should advance a coherent understanding of the customer and business.

He tells teams not to assume a launch will immediately work. The point is not to abandon ambition: set goals and do strong work, but be emotionally ready to investigate whether a failure reflects a minor detail or a fundamental misunderstanding.

A leader can be proud and optimistic while acknowledging imperfections, he says. To improve poor morale he looks for an early win and a clear explanation of why the work matters to users.

At Facebook he describes reframing advertising as helping people discover products and businesses find customers, and sharing as helping people maintain relationships. These are his interpretations of the work, not proof that every ad or social interaction is beneficial.

He similarly frames banking around helping people manage money, avoid excessive debt and save for goals. Connecting daily work to a purpose and visible progress can help a team regain momentum.

On family and work, Hudack says time with his children helps him reset. He describes integrating family, gardening and work through his day, while recognizing that this is his own arrangement rather than a schedule everyone can follow.

He distinguishes productive work from repeatedly refreshing growth dashboards. Obsessing over numbers can increase stress without producing the thoughtful analysis and changes needed to improve a product.

In quick-fire advice, he recommends angel investors assess whether someone can build from zero to one, rather than equate success at a large company with startup ability. Product managers should solve important problems instead of focusing primarily on promotion.

He advises founders to stay involved in product and be careful when hiring senior leaders whose experience may not translate to a smaller company. At the same time he rejects the rule that any doubt about a person means the hire must be wrong; context and expectations matter.

For a new product leader, he recommends moving quickly toward an early win. He values user and market intuition, product taste, the ability to work with engineers, understanding company context and intellectual honesty about what one does not know.

He says it is easy to rationalize disappointing data when attached to a product. He praises Monzo’s recent execution and experiments with AI, while acknowledging that future outcomes and the technology cycle are uncertain; his Monzo comments reflect his prior role.