I ran across a fantastic PDF on HackerNews by Henrik Kniberg that pulls back the covers on How Spotify Builds Products. It's such a good article -- and so different how many companies actually execute on building products -- that I wanted to highlight a few of the best parts.
Firstly, Spotify starts with a a strong, concise vision (with a singular focus): "Spotify’s vision is to bring you the right music for every moment." This has proven to be a fantastic guide as they iterate on their platform. In a similar way to Dropbox, which I described recently, everything Spotify does furthers that singular vision -- every new feature, every performance enhancement. "Starting as a music player a few years ago, their products are now evolving into a ubiquitous platform for discovering new music and connecting artists with their fans directly." This allows Spotify to expend all its energy on optimizing for product/market fit around that vision, and that's what creates tremendous shareholder value in product companies.
The article sets up, in simple terms, the product prioritization dilemma that trips many companies:
"Here’s the paradox though: Successful companies like Spotify only want to deliver products that people love. But they don’t know if people love it until they’ve delivered it. So how do they do it?"
Spotify's core product philosophy is this (bold emphasis is mine):
They use a four-step process to product iteration:
Here's why they follow these four stages:
"The biggest risk is building the wrong product - a product that doesn’t delight our users, or doesn’t improve success metrics such as user acquisition, user retention, etc. We call this “product risk”. The 4-stage model helps us drive down risk and get products out the door quickly."
The article goes into deep detail on each of the steps and how they are executed on. Here's a part that a product guru friend of mine says he especially likes:
"In addition, the Think It squad builds lots of different prototypes to experiment with the look & feel of this product – both “lo-fi” paper prototypes and “hi-fi” runnable prototypes (but with fake data sources and such). Internal focus groups are used to help figure out which prototypes best convey the narrative, until we’ve narrowed down to just a few winning candidates.
This is an iterative process with no deadline. The product is simply not worth building until we can show a compelling narrative and a runnable prototype that fulfills it, and we can’t decide upfront how long that will take."
Hope you enjoy this PDF as much as I did!
The four-step process that Spotify uses sounds like it was modeled on the FDA approval process for new drugs/medical devices. Phase I of the FDA approval process is to show that the drug is safe in humans using healthy volunteers--pretty much to "figure out what type of product, and why". Phase II is also to "create a minimum viable product that is ready for real users" by testing the efficacy in a small number of people with the actual disease the drug is supposed to treat. Phase III is when the drug is given to a much larger group of people with the disease (not 100% of people because it's still not approved, but usually several thousand people) and everything is closely measured and able to be improved. Then when the drug is approved, Phase IV begins which is similarly the end state. The drug is monitored indefinitely for long-term effectiveness and any side effects that develop way down the road. So all drugs on the market are said to be in Phase IV trials...technically. Just thought I'd point that out :)
Also, here's a great quote from AirBnB's story:
"Individual team members at Airbnb make small bets on new features, and then measure if there’s a meaningful return on the bet. If there’s a payoff, they send more pirates in that direction. This structure encourages employees to take measured, productive risks on behalf of the company that can lead to the development of major new features. It allows Airbnb to move quickly and continually find new opportunities.We’re trying to create an environment where people can see a glimmer of something and basically throw dynamite on it and blow it up to become something bigger than anyone could have ever imagined."
We've reorganized our team(s) within Marketing here at Quicken Loans and now literally have "The Product Team". We're the singular voice of "The Business" in regard to conveying to IT/engineering what to build. One thing we've struggled with, and are looking at ways to resolve, is "What comes before the Think It phase?".
We're not exactly the most Agile organization here (though we're working on that too) so even just getting an MVP shipped may take too long to catch/keep up to a rapidly-changing market (our mobile apps for example).
Knowing that, we're attempting to seek LOTS of data and primary/secondary research to help guide the inception of ideas from the start. We can sit around and "Think It" all we want, but we run the risk of being so insulated to outside realities that we miss the REAL market opportunity that users are begging for (whether they consciously realize it or not).
All this is to say that this part: "Think It = figure out what type of product we are building and why." can be the step that requires the most time/resources to do "right" so that Steps 2-4 can be dedicated to the ideas that have the best chance at success.
Jeremy the only real solution to the insular group-think issue is to test early and let the market give you direct feedback. This is incredibly hard for big brands to do because they feel like they have brand equity to protect. I.e., "we can't release something until it's perfect." The cost of moving slowly (and in the wrong direction) is very hard to calculate up-front, so the conservative approach wins. This is the classic innovator's dilemma -- "why mess with success." Big companies are loathe to cannibalize a cash cow.
This is why, as I pointed out in this blog about Coin's successful launch, it's typically the small companies that take risks and do innovative things.
A few ways to get around this are:
1) To create a sub-brand within the organization. Then, the pressure is off -- you can launch quickly, break things, etc., and nobody's worried that you'll kill the golden goose. You can take what works and fold it back into the larger organization. This is why Toyota has brands like Scion, for example -- they can be more aggressive with styling, etc., w/o freaking the Toyota heads out.
2) Segment your users into subsets: Figure out who the bleeding-edge fast-movers are vs. the conservative users who don't want anything to break. Use that subset as your beta testing group. They'll give you great feedback and will appreciate being the first to know about what's coming.
Back in January of 2010, my co-founders Sean, Isaac and I created one of the first mobile app creation platforms, AppMakr.
At the time, we had a thriving mobile app consulting business called PointAbout, and we were building high-end (and expensive) apps for large brands. Our team made the iPhone app for The Washington Post and Cars.com. We built the Newsweek iPad app and an iPad app for Disney, along with apps for clients like General Motors, US Army, the Entertainment Software Rating Board and others.
Making custom apps was really expensive -- especially in those early days. We had a dream of democratizing app creation so it was accessible to anyone. From that idea, AppMakr was born.
The day before we launched AppMakr, our team took bets on how many apps would be made in AppMakr's first month. Some people guessed 10, others 100. We had no idea what were about to unleash: In AppMakr's first 3 months, users made many thousands of apps. We had to scramble to support the growth. We even got angry calls from Apple's app review team who were overwhelmed by the number of apps being submitted; that's how our App Quality Index came to be, as a way to turn their frown upside down.
Joshua Gross has been interested in the intersection of art and computers since he was nine years old -- and it shows. He's regularly brought into companies at incredibly high rates to improve their front-end design and user experience, and consult with management over ensuring their products connect with customers. This interview is key for startup founders, programmers, product managers, and people in venture capital: if you're needing to understand what makes a product good for an end user, Joshua Gross is illuminating.
Joshua is currently offering a GiveGetWin deal "Get User-Centric to Win" -- perfect for startup founders, investors, those working on UI/UX, and other front-end developers, product managers, and creatives.
Design Drives Everything, by Joshua Gross, as told to Sebastian Marshall
Design really is the fundamental way a product works and interacts with the end user, the person using the product. It's more than how something looks or how it feels. It's how it looks, feels, works, and you could even go as far as to say it's why it works the way it does.
A chair is designed to fit the human form, for instance. Imagine you didn't take people into consideration when designing a chair, and only took into account making it look good. It could be too skinny, or have a bad angle, or otherwise be wildly uncomfortable. It doesn't make sense to make a chair to look pretty, you do it for people to sit on.