This has been a big year for Nimble Storage: About two months ago, we introduced Adaptive Flash, a platform that delivers adaptable performance and capacity for a wide range of enterprise applications. This week, we added the CS300 and CS500 product families, Adaptive Flash-based arrays that deliver 50 percent more performance than their predecessors.

Adaptive Flash was the second biggest announcement in Nimble history – the first was our original product, a flash-accelerated storage system. The aftermath of the successful launch seems like a good time to reflect on Nimble product management, a process driven by data and intuition.

Here, in Silicon Valley, every startup begins with a great idea – at least that’s how the story goes. In reality, that “original idea” often gets altered, or scrapped, before a product finally makes it to market. The journey from idea to finished product is sometimes messy, and always challenging. And, as a young company, we don’t have the luxury of funding limitless experimentation. Here’s how we make the hard choices that we hope will result in a successful product.

In the Beginning

Like any product company, many ideas come from listening to customers. We try to talk to them directly – the fewer the layers in between, the better are our odds of getting things right. Sometimes, the most revealing insights come from non-customers; prospects who didn’t choose Nimble, even those who chose a competing product. We interview as many prospects as possible to learn what we can do to win in the future.

You might be surprised to hear that some of our best ideas have come from within. Our two founders dreamed up CASL™, Nimble’s unique data layout. InfoSight™, our cloud support and analytics portal, was the brainchild of visionaries in our support organization. Many of our resiliency or usability features originated with engineering, sales, or support. They, in turn, rely on customer feedback, case histories, and/or InfoSight data.

The Power of Why

We try to be very clear about the why – why should we add a new feature, evaluating its relative importance. This helps to crystallize use cases, to design for simplicity, and most importantly, to prioritize.

Our user experience team defines the user experience as early as possible. This allows us to address usability challenges early. It helps engineering to visualize design tradeoffs, and allows QA to develop test plans. When possible, we show mockups to our field sales reps, partners, and customers to validate that we’re on the right track.

We’re somewhat obsessive about analytics like win/loss rates (and the reasons behind them), ASPs, workloads deployed, and InfoSight usage statistics. We slice and dice the data every way imaginable, looking for ways to enhance our products. But numbers aren’t enough. We also look at a lot of qualitative input – notes from customer and partner conversations as well as feedback from our sales team – to better understand the market. And we apply the power of intuition and judgment to make sense of it all.

Unfortunately, there’s only so much that can be done in a given time frame and with fixed resources. Prioritization, the most painful part of product management, is critically important. There’s no magic formula to successful prioritizing. We use the why as our guide.

Get the Big Rocks First

We try hard to get the big ideas right. There’s an old story about filling a jar with rocks: To get the most in, you have to begin with the biggest ones. If the big rocks are arranged very carefully, there will be more room for pebbles and sand. Of course, a single jar can handle only a limited number of big rocks at one time. It’s the same with products.

Nimble’s early history is a fine example. When we started in 2008, we had ideas for two flash-based storage products. The first, an “in-line cacher” (using the NFS protocol), was low risk and could be brought to market quickly. We spent a year on it, an appliance that would sit in front an existing storage system to accelerate its performance. The early reviews were great – beta-testers loved it. But there was a problem: the “available market” was small. Conversely, we knew the market for flash-optimized storage would be huge. We made a tough call: we killed the cacher in favor of the array.

That decision is our version of the Silicon Valley “pivot.” More than an educated guess, it reflected our point-of-view that was arrived at by talking to more than a hundred customers, watching industry trends, and trusting our judgment. We became convinced that a system built to leverage flash and low cost HDDs would provide the proverbial “10x advantage. ” We would need that kind of edge to be successful in a big – but highly competitive – market.

Since then, few of our product decisions have had that “make or break” quality. However, we continue to rely on the same clarity of thought and willingness to change course, traits that served us so well in the beginning.