Lessons hard learnt about the startup design process

A Frontline portfolio dinner discussion with Google Ventures’ Tom Hulme and HELIX Centre’s Maja Kecman



We were lucky to host the Frontline portfolio for a dinner in the beautiful St. Pancras Clock Tower in King’s Cross — a truly unique location for a very interesting conversation about design x startups.

Our founders were joined by Tom Hulme (Google Ventures partner, OpenIDEO founder) and Maja Kecman (HELIX design lead, Balderton associate), who each gave a short design-centric talk.

Below are highlights from each of their talks, as well as great nuggets of information pulled from the dinner discussion.

Tom — 6 things I wish I had known about design when building companies

  1. Do not compromise on the questions you are asking.

Do not underestimate the importance of how much time you should spend just thinking of the questions to ask. Too many teams rush the questions because they are in a hurry to get to the final “answer.” Finding the right questions to ask can save your company immense time and money. Do not underestimate the importance of the the right questions.

2. Empathy is hugely important.

The reason CEOs can often be perceived as arrogant is because they are building products to solve their own problems. They are “scratching their own itch” — it is a problem that they know intimately. Hate it or love it, they have true empathy to the end user (because they are it).

If you are not the end customer, Tom urges your team to co-locate with your customers. You can always relocate engineers to your customers, but never vice versa.

Focus groups are not a place to do this — beware of confirmation bias. Use the Starbucks test: imagine you’re chatting with someone in line at Starbucks without any idea of who you are. Would they understand the product? Would they use this product?

3. Bring in new customers throughout the design/development process.

It is very easy to engineer away from your new customers — by focusing on one customer throughout the development process and forgetting that they are becoming more and more experienced with your product. You will mistakenly assume that their assumptions and education levels are the same as brand new customers.

It often makes sense here to “launch to learn.” You don’t necessarily have to use your brand here, but it is cheaper and easier to launch and learn from how customers are using your product.

According to Tom, you are as good or as bad as the number of iterations your product has gone through.

4. Make creative leaps.

Because there is so much of this “go heads down” mentality in startups, you may find your team grinding away for weeks, only to emerge and realise that you have been incrementally climbing to the top of a hill for the wrong customer. Do not be afraid to make leaps to other hills to see if that could work.

5. The best design removes friction.

Learn to say no throughout the design process. Over-designing is very easy — it is the ability to say no that can help remove the friction in solving the use case.

6. Be as open as possible.

This “openness” can come in two ways.

  • Be open to customers’ feedback. Do not design in a silo. Feedback is essential throughout the design process.
  • Have humility and openness in the sense of your product. Tom is very keen on API businesses — teams voluntarily saying “we are not going to build and design everything.” A good example of this is Twitter, who were initially so open with their product. This allowed other companies to essentially do R&D for them — the retweet, photo posting, and videos were not initially Twitter-built features.

Maja —human-centered design

According to Kenji Ekuan, good design has three pillars — depth, utility, and beauty. Nail all three and you have a great product.

Depth — why are we doing this? Who is it for? What is the problem that we are solving? Are we even solving the right problem? You should reach depth in understanding before starting to even prototype.

You should even question why we do the things we do currently. Oftentimes, there can be a design solution that arises before taking into account the assumptions we make about people’s day-to-day behaviors.

Utility — are we solving this problem well?

It is absolutely necessary for designers to have empathy for the end user. This means getting constant feedback from users — which is thankfully easier and more instantaneous to do in the digital world. This quick user feedback enables rapid iteration on the product.

Do not ask people for what they want — they rarely know themselves.

An example: Maja had worked on creating a surgical tool for knee operations. When asked for what tool they want, surgeons told the group that they just wanted improvements on the existing solution — perhaps something thinner, lighter, etc. But when observing them during surgery, Maja and her colleagues observed that they were ignoring the tool altogether and instead free-handing the process — quite accurately, in fact. Seeing this, the designers created something altogether different from the original tool — a dead simple, easy-to-use tool that would just further aid the surgeons in their already-accurate process.

Form will emerge from function, so understand the use first and design after that. Great design is obvious, but obvious things are the most difficult to create.

Beauty — does this delight users?

This is the “magical” component of design — the element that is oftentimes difficult to describe, but most easy to immediately feel. Therefore, a good designer will be obsessed down to the last pixel.

This element can only be achieved a) when engineers and designers work closely together, and b) if there is true buy-in from the CEO. The company culture must nurture and respect this way of working — protection and advocacy for good design must come from the top.


For user research to be impactful, it needs to be deeply embedded in the product development process. This is only achieved if every member of the team believes in a user-centered design process and is willing to make changes or make decisions based on user research findings.

Here are some companies that have figured out how to shift user behaviour:

  • Airbnb — moving us away from the “stranger in my room” fear
  • Uber — replicating the limousine experience — don’t touch money, it’s dirty
  • Citymapper — “thinking,” and not “processing” — it says “take me somewhere”

All of these companies have the user at the center. Their designers built products that make it easy for the user to interact with the physical world through digital interfaces. They makes things magically appear in front of the user by designing a system around them.

Remember: You are not the user. The CEO is not the user. Shortly after Larry Page was appointed CEO of Google, he said, “A day watching users and talking to them will teach you more than a year of meetings and briefings.”

Maja looks forward to the day when we are no longer talking about design-led companies, but where design is at the core of everything that we do — along with tech, engineering, business, marketing, etc.

The boardroom these days is mainly filled with quants — people talking about hard figures. Lets get more qualitative thinkers in there as well. Lets get designers in the boardroom.


Other golden nuggets from the Frontline portfolio

Build mechanisms for your customers to tell you what they want.

Drop has a Slack integration where all search queries from the site are entered into a Slack channel. This allows them to see, in real time, what customers are actively seeking out/having trouble finding.

Enable people to be more clever than you.

Love & Robots continues to be guided by this ideal — they are building a platform to allow customers to leverage their tech to create bigger, better, and bolder designs.

Being a “design led” company reflects internally as well.

It is not just about what external design looks like to your customers — design thinking should be reflected in internal operations, company-wide empathy, etc. Barricade assigns engineers to incoming Intercom questions, ensuring that customers talk directly to those building the product. This can help close the feedback loop.

Saying is no is not easy.

Pointy gained a lot of traction in Ireland quickly, but that growth came with plenty of questions, feature requests, and other time-consuming asks from small businesses. The team has had to painfully and repeatedly say no to customers, but doing otherwise would shift their focus, decrease productivity, and obscure clarity in vision and design.

Want to learn more about Frontline? Sign up for our quarterly newsletter to receive the latest portfolio news, upcoming events, and team updates!