What would it mean if anyone could make their product ideas a reality?

People have great ideas for products all the time, but they lack the resources or expertise to actually make that idea into a real product.

It's never been easier to build new things, but despite that advancement, for most people, building things can feel more out of reach than ever. Quirky is an invention engine that enables just that. It turns ideas sourced from an incredibly dedicated community into real products on store shelves.

Everything comes together in a weekly, livestreamed broadcast known as Eval (short for “evaluation”), where the entire company, along with a judging panel of executives and special guests, would gather to discuss, ask questions about, and vote on the community's ideas for new products.

In many cases, those ideas turned out to be great, and we had an exciting relationship with our customers:

One of the biggest challenges of companies that are based on social networks or marketplaces is managing the user/company relationship.

It’s a challenge to effectively communicate and collaborate with a huge, diverse, and geographically distributed network of community members. And since Quirky operated entirely on community-sourced ideas, this was an essential part of our product consideration.

In many ways, the community acts like part of the company. As designers on the Platform team, we were tasked with designing an experience that felt open, with free collaboration and creativity, but that also channeled that crowd enthusiasm productively.

‍

‍

Problem

But we had a significant problem with spam. There was a large volume of irrelevant, inappropriate, or obviously impractical invention suggestions.

These ideas generally fell into a few categories:

• Ideas for products that were illegal or that would make it easier to break the law, such as a device to manufacture illegal drugs (also probably not cost-effective/feasible, I should add)
• Ideas with no technically feasible plan for implementation, in particular at scale, such as real hoverboards or flying cars
• Ideas for software or ideas that would require significant software to make them work (these were some of the toughest rejections since they often weren’t bad ideas—just out of the scope of Quirky's physical product focus)
• Ideas that were unintelligibly misspelled, very incomplete, or clearly not a good-faith submission

Attempts were underway to mitigate this issue with things like artificial intelligence screening, but distinguishing noise from signal still required human attention from the community support team.

‍

Challenge

I was tasked with exploring options to reduce this issue from a design angle:

How could we design the idea submission page to reduce the number of spam submissions while not presenting increased difficulty to those with legitimate ideas?

Optimizing funnels like these are one of the most common design goals, but this situation presented a unique paradox. Unlike with something like checkout, there was no “right” outcome to optimize for—only “wrong” ones to discourage.

Moreover, this distinction relied on the user’s judgement, which is something that typically improves with additional consideration. Unfortunately, more consideration positively correlates with more time and difficulty, increasing the likelihood that a user would bounce.

‍

Process & Exploration

The first step was research, and we started with our coworkers who reviewed ideas every day. Usually, products make it to Eval by being upvoted by the community, but Quirky staff may also move ideas forward as well if they’re exceptional. These staff members are in groups that handle different invention categories. We asked a member of each group to tell us some of the bad ideas they frequently see submitted.

Our success metric was to see fewer ideas submitted (because some users who would have submitted a bad idea would not), but we didn't want to create too much of a barrier to conversion. We decided on a success metric of no less than a 10% decrease, but no more than a 20% decrease.

Initially we considered putting the list of "what not to submit" on one page, but after more iterations on that concept and some feedback, my team and I realized that the page was getting too long and people were unlikely to proactively find the information they needed out of a giant list.

‍

Solution

Ultimately, a simple answer ended up being the most successful. After exploring several implementations, we ended up implementing a checklist requiring community members to acknowledge that their idea didn't fall into one of the above categories.

While we initially considered this problem from the standpoint of policing, this solution was all about education, rather than enforcement. We discovered that a surprising number of users were unaware that Quirky didn’t produce software, for example, and the friction presented by checking a checkbox was enough to discourage a sizable number of “bad” submissions.