Pinterest and the Process of Designing a Multi-Billion Dollar Startup

Some startups recognize that design is an investment rather than a cost, and one that yields better results in the long run if incorporated into a business from inception.

At a recent Kleiner Perkins CEO Workshop, I interviewed Pinterest CEO and co-founder Ben Silbermann to discuss how he came up with the idea, and why he was inspired to build the company.

Since its inception in 2009, Pinterest has raised a total of $762.5 million and has a $11 billion valuation with its most recent round of funding.

Our conversation is captured in KPCB’s Ventured podcast, where Silbermann reveals how he handled the growth of his company, discussed strategies for getting designers to work with engineers, and described how he attracts top talent with generous equity.

Actually talk to your customers (5:15)

Silbermann’s mom was an early Pinterest adopter and these days, she sends him bug reports via the phone. But in most cases, friends and family won’t give the most direct feedback. Silbermann worked in customer support at Google prior to starting Pinterest, so he wanted to speak to many of his first users. It’s a good way to understand what is working and what isn’t.

Words are important to design (13:50)

Language was the original interface for the computer. The language that companies use to introduce their product still dictates a user’s first impression. For example, Pinterest decided to use the vocabulary of pinning in boards to reinforce a physical metaphor.

The idea of pinning a board is something that people have been doing in the physical world, so Pinterest introduced their product using that language to acclimate its users to the idea of collecting digital objects.

Strategies for when engineers and designers clash (15:20)

Engineering in general is a deductive problem solving discipline and engineers tend to want to reduce the uncertainty in code. However, great designers keep a lot of variables open so they can play with all of the possibilities. Therefore, getting engineers and designers to work together in a scaled system is challenging.

One way Pinterest has been able to do this is by setting constraints that both engineers and designers can agree on. A constraint can be a deadline or a constraint can be a set of actions that somebody needs to perform for the benefit of the entire team.

Phasing user introduction (23:16)

It’s really hard to change consumer behavior purely through optimization. Therefore, Silbermann likes to break up the introduction of a product into three phases. In the first stage, he asks: ‘Does it work or not?’ Data is not used in the first phase. Next, the product is put in front of employees. The point of the second phase is to see how sticky the product is: Are people using the product again? The last phase is focused on optimization and scaling.

When data helps design (24:37)

Design is a problem solving discipline just like engineering. Data helps give good direction. Generally before a project starts, there is a statement of the problem and how it will be solved. But data isn’t perfect and doesn’t suit every occasion: Pinterest wouldn’t redesign its logo based on data.

Equity attracts top talent (26:17)

While some people advised against this, Silbermann thinks the whole point is to make top talent feel ownership in the company. Silbermann has never regretted stretching equity to hire a really good person.