This article has been migrated to our new blog. Find it at https://building.nubank.com.br/products-are-never-really-done/
At Nubank, we want to deliver the best possible products to our customers. Period. This aspiration guides our product, engineering and design decisions from the moment we conceive of an idea to the first version of a product we put in a user’s hands to the endless iterations we make thereafter.
But what does “best possible product” really mean? How do we decide what features ultimately make it into a product, and when? How do we determine a product is ready for testing with real customers?
As we begin testing “Conta PJ”, our very first product aimed at small business owners, we wanted to explain more about how product development at Nubank works and why we think our approach is core to our values and critical for the creation of the best possible products for our customers.
Where do we even start?
Impact and our values
Product development at Nubank is fundamentally about impact, and this is where we start in evaluating new ideas. We want to have meaningful, long-term value for our customers by providing products that are truly empowering and life-improving. Or as one of our company values says: “We want our customers to love us fanatically”.
Having an impact means we do not look for opportunities that are small or use cases where good solutions already exist. We look for big problems that cause people (and companies) enormous amounts of frustration and pain. One of our other company values emphasizes this point: “we want to be bold and challenge the status quo”. In financial services, this boldness often means targeting the complexity endemic to financial products.
Technology that is high quality, scalable, and delightful
If we think a particular opportunity is worth pursuing, we try to assess how well we can use technology, design and data science to create a high-quality product at scale. We look for ways we can simplify experiences using software, usually by enabling customers to do things themselves in a way that costs them far less time, money, and effort.
We try to rethink conventional wisdom about how financial products are offered currently, taking a first principles view based on the underlying customer need. From a design perspective, we think deeply about how we can make this experience easy and intuitive (and hopefully delightful in some way). And the experience does not end with the app: we also evaluate whether we can support the product operationally with our world-class customer service.
The business matters
Lastly, our product decisions are bound by financial, legal, regulatory and compliance restrictions that we take very seriously. We believe our technology-first approach positions us well here — our lean cost structure enables us to give more value back to our customers, we have invested heavily in security, and we have developed numerous integrated solutions related to compliance, reporting and other back-office requirements that ensure we can deliver our products safely and according to rules and regulations.
Data, judgment, and time
The challenge of product development is that there is an infinite combination of features that will move the above dimensions in different ways, making solving for the “best possible product” a difficult proposition. For example, a particular grouping of high-impact but high-cost features might render a product nonviable economically. Or a flow designed for one customer segment could render the product confusing or unusable for another if our technology does not adequately support customized experiences for each.
Further complicating this process is optionality and information asymmetry. As we develop a product, we want to eliminate the largest risks to success as early as possible, but this is also when we have the least information. Our product teams, therefore, try as hard as they can — using data we already have, market and user research, test prototypes, their guts, and any other means necessary — to make smart decisions early so that we do not waste time building something that our customers will not love.
Sometimes these decisions can be easy and do not require a lot of research. For example, it did not take us long to decide that NuConta should be free of monthly fees and offer a phenomenal rate of return for our customers.
Other decisions are more difficult. For our credit card rewards program, for example, we decided to use a simple 1:1 ratio for cash to point conversion, as we felt a program structure that was more transparent and less complex for customers was more aligned to our values. However, this choice also meant we had to work harder to educate customers who struggled to compare our program to others (as few other programs are structured this way).
Even the best data and intuition, however, is no substitute for real usage. Customers react to real products in ways that can be difficult to predict with prototypes, interviews, or our best guesses as to their needs. For this reason, we believe strongly that once we have an early version of a product with a strong enough value proposition for a group of test users, it is better to get that product out into the market, even while we know there are many areas where it can be improved. We do this because the value of the information we gain from real usage — seeing how customers respond to the experience, what features or tools they use, and how they feel about using the product — is critical to making that product better, faster.
A live product also enables us to validate our assumptions about our ability to deliver a high-quality experience (and to fix bugs and improve flows). Furthermore, it allows us to co-create with our customer support teams any additional supporting processes our customers might require in the event they are confused or have issues we should be solving for them. This is a key reason why we use gradual roll-outs and waitlists, as they enable us to maintain a high-quality experience while increasing access to the product over time.
Why our PJ test in Brazil is special
A few months ago, we formed a team to evaluate products we could build to serve the needs of business owners. A few weeks ago, we announced the first test roll-out of our product “Conta PJ”: a digital account for “pessoa jurídica”, aimed initially at entrepreneurs and small business owners.
We think this account product solves several pain points; in particular, ease of account opening and money movements. But we know it can be better, and we hope to get there faster with the help of real PJ users. As we have only a handful of Nubankers who have tested earlier versions, these first users will be of even greater help in validating some of our hypotheses and evolving the product together.
If you are a new PJ user and are frustrated you cannot share the experience with all of your colleagues yet, we’re sorry about that! But that’s a great sign our early choices around impact have been the right ones.
If you think we’re onto something but we really need to add feature A or B before you can leave your existing bank or make Nubank your primary account, that’s also incredibly useful feedback at this early stage and will help us tailor the roadmap of this product such that it meets your needs more fully. The only feedback that really concerns us is a shoulder shrug.
Launching products at Nubank means we are never really done. As we say at Nubank: it is always day one. The moment we think a product is “good enough” is the moment we start losing relevance to our customers.
Want to join our team?
Check out all the positions at Nubank's offices in Berlin, Mexico City, Buenos Aires and São Paulo.