Culture, culture, culture. Every tech CEO and founder, fiddling in their proverbial garage, dreams of building a company founded on principles partially informed by childhood readings of Alexandre Dumas: All for one, one for all. And, by the way, letâ€™s change the world. This team-focused, ideal-driven mantra flies in the face of everything corporate culture stands for, and thatâ€™s exactly how storybook Silicon Valley/Silicon Alley like it. If thereâ€™s one word that murders the mojo of this daydream, however, itâ€™s this: budgets. Worse than even 360 performance reviews, budgets reek of a decidedly corporate stink. For starry-eyed founders, the thought of hiring someone to sit in your company and focus on pennies spent (how anti-big picture), five-year plans (didnâ€™t Stalin come up with those?) and getting your receipts together to start preparing for your first audit (donâ€™t even get me started) almost induces one to self-inflict extreme physical pain.
And yet, at some point in the life cycle of a company, hiring a finance professional is inevitable. The good news? The financial role is evolving, from helping to optimize cloud costs to useful pricing analyses for new and existing products. The not-so-good news? Budgets, well, theyâ€™re not going anywhere. But itâ€™s a necessary medicine, and, if the right finance professional is hired, shouldnâ€™t be a painful, arm-wrenching process. For many companies, depending on size and stage, cloud spend is the second-highest cost after payroll. Of course, the current slew of cloud wars between the industryâ€™s best and brightest â€” Amazon, Google and Microsoft â€” are helping the costs to reduce themselves, but this sort of passive maintenance is far from ideal. By working in conjunction with the Systems team, a plugged-in finance teammate can reduce cloud costs by another 25 percent (at least).
Letâ€™s assume your company is on AWS, which is an easy example. By working with the Systems team and analyzing instance usage data from the last few months (if the Systems team uses Splunk or similar software, this is a straightforward exercise), prepaying for instance usage reduces the hourly cost significantly, and optimizes instance type for necessary memory and CPU usage (many engineers can grossly overestimate, launching a c3.4xlarge when a c3.2xlarge or even c3.xlarge would have sufficed). Alternate solutions â€” like Mesos clusters â€” live more in the Systems domain, but a clued-in finance person could be very helpful in optimizing the make-up of these clusters from a cost perspective.