The Future is Retro: How One Early Product Decision Made All the Difference at Pendo

Written by Shannon Bauman  | 

3分

 

When building a new product from scratch, it’s often the early decisions that matter most. These are the foundations that you build upon, the choices that you often cannot take back.

As the first employee hired at Pendo, I’ve seen the product change and grow over time. But one thing hasn’t changed: our focus on making Pendo as easy for PMs to use without sucking up precious engineering resources. A prime example can be seen in what may have been the very first, foundational decision here at Pendo—the decision to build retroactive analytics.

It’s a decision that has paid dividends for us as a company, one that most certainly fits the criteria of making product analytics more accessible for product managers. But it was the harder path, the more daunting engineering challenge. The right early product decisions often are.

The choice to go down this path came from a very simple insight. The traditional model for analytics is engineering heavy. The product team tells engineering, “Hey, we want some data on usage of this feature; can you install tracking code on it?”  The PM then slots that work for some future sprint, waits for the deployment, and weeks or even months later the data STARTS flowing in. To actually get significant results they then need to wait weeks and months longer for the data to build up.

This seemed ridiculous to us.

So during the first few breaths of the company, with just a few of us stuffed into a small windowless room, we figured, hey, let’s just collect it all! Todd (Pendo CEO and co-founder) and Erik (Pendo CTO and co-founder) put fingers to keyboard and started to build it the Pendo way.

What’s the Pendo way? Collect everything. We grab information about every click and page load, regardless of whether a PM has identified it as important yet or not. If you never need that data, then no worries. If you decide you want that data at some point, Pendo will go back through all of its historical data collected, and compile the results. A time machine of sorts.

It’s not an overstatement to say that this is life-changing for PMs. Before, they waited months, or more often, just didn’t even bother because it was too heavy of a lift. Now PMs ask any data question they want, and get the answer back in minutes.

Why doesn’t every analytics company do it this way? The quick answer is that it’s really hard. It means collecting a ton of data, and doing so from the messiness that is the internet. People develop applications in thousands of different ways, and users view those applications with dozens of browsers and browser versions, and we have to support them all. We have to take this messiness, and create order from it. Calling it challenging is an understatement. (One of our early engineers has a pin-cushion voodoo doll on his desk with an Internet Explorer 7 logo on it for this reason.)

At a startup—or any company, for that matter—you don’t always make the right choices out of the gate. Fortunately, in this situation we absolutely did.

We chose the harder path. And that has made all the difference.