February 20th, 2012
Paul deGrandis: 5 tips for navigating startup life
[Editor: Thanks to Paul deGrandis, Vice President of Engineering at Tutorspree, for sharing his wisdom in this guest post. Interested in more insights? He's speaking on Tuesday night.]
I have been working in tech startups since the age of fifteen and the first “.com” bubble. I have had the pleasure of working with some of the smartest and most influential people in software, building up some of the most successful companies, and helping others move their ideas and dreams forward.
Recently, this wealth of experience has started to coalesce into tangible nuggets of advice; easily consumed heuristics of how to navigate startup life on the frontline. I was fortunate enough to get in touch with Silicon Florist during my most recent stint in Portland. And we both though it’d be cool to pull together a guest post.
So, I’d like to give you five tidbits to take with you. As with any rule or heuristic, adapt it so it makes sense for you, or break it entirely!
1) Don’t forget to take a break
My brother and I used to joke about getting tattoos on each of our forearms, “Take a break.” But there’s a careful and subtle lesson in that tattoo. Taking a hiatus or jumping into a completely different industry has been great for expanding my toolbox. Every unique experience provides one with a new set of inspirations, approaches, and inputs on how people function.
When I was working in advanced software research, I witnessed the incredible possibilities when your entire team is constantly engaged in research and development—if you breed a culture of relentless innovation and constantly questioning the limits.
When I briefly worked in the game industry, I discovered that software development in a purely entertainment-focused industry takes a backseat, but with good reason. The entire industry is bent on finding new hooks and designs that engage and entertain gamers. You only need software in that process to carry out the final vision. I now take those lessons of carefully threading entertainment interactions throughout the user experience into every product I build.
These new experiences are similar to when I learned new programming paradigms. They forced me to think in a completely different way, discovering new solutions to problems. I try my best to make use of every skill I acquire, whether directly (like working on search or recommendation engines for the past few years) or indirectly (like my experience working in social gaming).
2) Make something people want
I wrote a blog post awhile back about the heart of engineering: Deliver well-founded solutions that say something.
Startups are no different. You should be providing a product, that’s built on the best technology and knowledge possible, and that solves a real problem in a measurable way.
When working on OurShelf, an advisor asked, “are you building a vitamin, an aspirin, or a pain killer? The latter is the only one someone absolutely needs.” The analogy has always stuck with me, but I’d also like to add “Snickers” to the list.
Some products aren’t even a vitamin, they’re just something people crave every so often—something sweet and fun, worth indulging. And when I’m hungover, there’s no better medicine than a Snickers.
Take Pinterest—it’s a Snickers. But when you’re planning a wedding, it’s insanely helpful for coordinating. It also follows another heuristic in this category: digitize the analog. Pinterest is no different than the coupon-clipping-scrap-collecting behaviors of yesteryear, pushed into the online and social world.
This is also why I like startups centered on marketplaces- you increase the odds of offering something that people want. Etsy uses a range of diverse goods to appeal to the greatest possible audience. They digitized the analog. They took smaller communities of crafters, brought them online, and networked them together in a single marketplace. By moving an entire lifestyle (DIY) online, Etsy had an immediate audience.
Focusing on what people want becomes easier if you’re actively listening to your users and engaging them for more data points.
3) Pay attention to your data
I recently gave a talk at Meetup.com about how Tutorspree makes use of every possible piece of data. The takeaway was simple: Elevate data as a first class citizen in your application. Structure features around using and producing data and use the combination of your data as guardrails for making business decisions and validating assumptions.
True validation is hard, but it will always lead to success. The simplest validation starts with you exercising your data and asking why not what. This focus on data and validation can be woven into all parts of your company culture and your application, from system architecture to new product launches.
4) There’s always something you can be learning
It’s naive to think you’re producing something completely new that’s never been attempted before. I spend at least a day searching on scholar.google.com for the latest knowledge and approaches to solving my problem or problems similar to mine, before I start working on any piece of technology. Increase your chances of success and use your time more efficiently by standing on the shoulders of giants.
Tutorspree opened up new problem domains for me: learning and teaching styles, evaluating lesson quality, semantic indexing and matching, and contextual ranking to name a few. Taking time read the most influential papers in all of these areas and produce some R&D prototypes has resulted in technology that puts us well-ahead of our competitors.
5) Never stop moving forward
If you’re applying smart research and development, analyzing your data, producing new data, and learning something about your problem domain, then you’re moving forward. If you’re moving forward, then you’re certainly not dead in the water. If you’re working on something and you can’t answer “why,” then your effort should be focused towards shining more lights on the dark alleyways.
Your current momentum encourages your future execution. We look at engineering metrics weekly at Tutorspree. Interestingly enough we found that team output was a function of my managerial output. My momentum set the teams momentum, and the teams momentum allowed me to focus more on managerial tasks. Once we stabilized that rhythm, we were unstoppable in meeting all of our goals.
Like I said above, take this knowledge, adapt it, enhance it, or ignore it completely. Find out what works best for you, isolate it, and repeat it.
Thanks for all the love Portland, keep being awesome.
Paul deGrandis (@OhPauleez) lives for magnificent engineering. Elegant, well-founded, useful solutions to problems that say something about engineering’s beauty. He loves metrics, taking on the impossible, and making lives better through technology. Currently he is the VP of Engineering at Tutorspree (YC). Previously he worked at PushButton Labs, Etsy.com, OurShelf (DreamIt), and SilverCloud Software as well as working in advanced research (DARPA). He’s also contributed, time, money, and effort to Code for America, PyPy, and Clojure.
What’s more, he’s speaking here in Portland on Tuesday, February 21. Feel free to join him at Urban Airship around 6PM.
(Image courtesy sochacki.info. Used under Creative Commons.)