CTO Corner #7: When to build a car and not a faster horse
In this week’s CTO Corner, I spoke with Kevin McCarthy of VentureApp. We discussed everything from finding the right solution at the right time to knowing when to build for the future and when to build what the customer wants in the moment.
How did you get into engineering?
Senior year of college, I started working with Chase Garbarino and Gregory Rogan on Campus Word, a digital media publication solely comprised of student writers. I got fascinated with how our CMS (Joomla!) at the time worked and just started building stuff for it.
I was fascinated with how everything was put together and started doing more and more of the technical stuff with that first business. It happened naturally. I didn’t do any CS in college, which would have been a huge help. But I had a knack for looking to see how stuff worked.
My co-worker Jared has better stories about getting started with engineering and entrepreneurship. When in high school, he and his brother would go war driving with a Wi-Fi antenna, looking for open Wi-Fi networks, then approaching the home and offer to protect their Wi-Fi for $40. Apparently, this is a federal crime now, but it’s a funny visual thinking of them two with a little antenna on their car.
After nine years at Streetwise, what drove you to found VentureApp?
Lessons we’ve learned along the way, the opportunity to build something new, and working with awesome people that I learn from every day. When we were first getting going at Streetwise, we lost a lot of time talking to the wrong vendors at the wrong time. It’s not just the wasted money picking the wrong solution, it’s also the time.
What are the lessons you learned?
We made a lot of mistakes. We walked away with some funny stories, but ultimately we wasted money and more importantly time focusing on the wrong things at the wrong time. But we had the opportunity of building something new and working with awesome people that I learn from every day.
On the product side, we learned that talking to your end users is way more important than quantitative data in the onset. On the tech side, we learned that building something people love is way more important than building something that will scale (you can worry about optimizations down the road), and using software/technology that is five years old is better than playing with the latest and greatest is cool (which causes headaches for production). On the hiring side of things, a great attitude and a desire to learn trump great talent every day of the week. Also, hire real slow, but fire real fast. Cultural debt is a real thing.
What are the challenges you currently face as a CTO of an early startup?
Outside of problems any startup faces, we’re moving very quickly through product features, testing, iterating. One challenge we have is keeping up with code/feature deprecation and building as fast as possible while also making sure code documentation is in place. It’s something I have been thinking a lot about.
How do you balance between building for your company vision and what customers ask for?
The biggest challenge is knowing what areas you should trust your gut on, and what areas you should trust the customer’s feedback. It’s generally a case-by-case example. You’ve got to obviously take in what the customer says because you are building for them, not your grandiose vision. But the question hints at thinking about when you need to build the feature you think will lay the brickwork for the future and building something that a customer requested. I think trying to determine which one of these you should follow is the toughest thing to balance. Henry Ford said if you asked people if they wanted a car, they’d say they just want a faster horse.
But there are times to build for you and times to 100% listen to customer feedback. Knowing when to do which is very tough, but very necessary.
What is your take on AI and business solutions services?
There is almost an infinite of permutations in making business decisions, so building the perfect AI without a massive data set is next to impossible. Designing AI to accommodate for 80% of business decision-making, then pairing it with content and the ability to talk with someone is very effective.
I see AI as just matchmaking. For us, it’s currently just the intelligence to match the right provider with the right business and vice versa. We spend a decent amount of time creating qualifying questions for our buyers.
What I am more interested in long term is figuring out trends based on user activity. If we know a business closed funding in January, what are the types of request that they say they need and what don’t they realize they need? Then, how does that change moving forward?