VideoAmp turns 4 years old this month. So many memories, and so much growth. Let’s reflect on some of the worst decisions and associated learnings that have been fundamental to our growth. Why the worst you say? Because when you accelerate your learning rate in a fail-fast environment, it then allows you to make fast decisions vs. the analysis paralysis predicament.
One of my favorite wildcard interview questions for managers is: “Describe the failure you’re most proud of.” This sometimes catches people off guard, so I’ll share mine first to get the conversation going. Inevitably it comes down to choosing because there are so many.
You can’t talk about growth and success without talking about these failures. Unless you’re some kind of hybrid AI-robot-space-alien who’s absolutely perfect, there’s going to be mistakes. What’s important is that as a leader you own your mistakes and be transparent about them.
Finding the right people is the single most important aspect of growing a startup. The biggest mistakes in hiring have been:
- “Hiring out of hand”, or short cutting the normal processes because the candidate is well known by other engineers in the group.
- “Promoting out of hand” is another variation of this where you don’t put someone through the normal due diligence because they already present. In our case, individual contributor (IC) engineers don’t go through the rigorous reference checks that new managers do.
- “Hiring under duress”, or lowering your bar simply to meet headcount goals. I strongly advise to hire fewer “full-full stack” senior engineers a bunch of junior or mid-level career developers who require more time, nurturing and precious attention units from your senior staff.
This is a moving target, and one which has definitely been an ongoing series of blogs. In the early stage, we “over prescribed scrum.” Something that a very early stage company can do, along with too much agile processes too soon.
For example, prototypes and proof of concepts should not have a rigorous definition of done, with endless unit, functional, and end-to-end tests because the product’s requirements will probably zig-zag wildly.
Conversely, having too little process when you are supporting enterprise customers can also be a problem. I spent my prior days as an IC and contractor retrofitting and scaling CI/CD pipelines and training engineers how to write better tests.
The challenge is right-sizing this all along the way, and the trigger points on when to change may not be obvious. I say this because we grew headcount by 22 engineers from Oct 2017 to Feb 2018, and in the process we did not modify our simplified Kanban approach to a prescribed Scrum process quickly enough. Growing pains emerged, to say the least. Now we’re in a spot where we can withstand a magnitude scale of growth with roughly the same squad and tribe-level process.
Riffing off rapid growth of Brains in engineering, we didn’t scale management fast enough. Almost all startup engineering orgs start very flat, with all ICs and no management. You have “tech leads” who may split their time doing light managerial functions, but they all write code and dive into the operations.
The biggest fail here in scaling the ICs was not scaling the org and management structure to follow. At ~44 engineers and data scientists, we have a duty to deliver on our mission to provide an environment where they have the opportunity to do the best work of their lives and be worth more in the marketplace.
Without this vital management structure, there is a vacuum. Do not bolt this on later, build it as you go. We took inspiration from Spotify’s model of engineering organizational scaling.
Growth & Career Pathing
We’ve talked about all of these fails, how about something that has worked well for us?
When doing initial contact with a candidate, I often ask “why are you in market?” I have seen countless folks who are looking because their current management does not have their growth and career pathing in mind. In extreme cases, they can do a day’s work in four hours, feel like they are under-challenged, and have not learned anything in years. They work in an environment like the movie Office Space.
Taking an opposite approach is to engage in the growth and pathing of every individual. We do this by:
- aligning the growth of the individual with the company’s growth.
- having management check in frequently on the success of this, and
- setting up formal quarterly check-ins on measuring these goals. Google adopted this early on from John Doerr in the form of OKRs, and there are great platforms out there which can formally measure and track these objectives.
Career pathing is a longer-term concern. I ask candidates from the start “so what’s the next job after VideoAmp?” This often catches them off guard, then after careful thought most reply with a role 1-2 levels beyond where they’re at now. It’s our goal to help steer them in whatever path they currently see.
Many earlier-career engineers think management is their ultimate path, but I have found that many will stay on a tract of engineering excellence. Whether it’s a Principal Engineer or a VP of Engineering, the goal is to orient the new challenges in a manner which grows them in that direction, even if we can’t fully realize their ultimate pathing goal while at the company.
By paying careful attention to these details, we have found our annual retention rate in the high 90%.