How to Double Your Company's Growth Every Year

January 23, 2013 Farhan Thawar

People have inquired: How is it that Xtreme Labs has achieved a 100 percent annual growth rate every year since starting in 2007? Does it negatively impact our culture?

Our bout of growth is driven by the many referrals and resumes we receive, partnered with a low attrition rate and staff who love working here. We owe a large part of our growth to the sharing nature of the entrepreneurship community, both online and in Toronto. It’s our turn to pay it back.

Here are some of the methods we implement to accomplish successful growth:

Hire fast, fire faster

Thanks, Mark Suster, for sharing your mantra with the world. We’re running with it.

Interviews are a terrible predictor of performance. Candidates often test as false positives and false negatives. For example, a candidate could have the ideal skill-set for the position, but may not have put her best foot forward in a two-hour interview. When companies ask coding questions, expect answers on whiteboards, and disqualify for giving the wrong answer, they are forgetting this:

Most people don’t write code on whiteboards

In other words, this scenario sets the candidate up to fail because it forces her to use a method that few engineers use in their regular workday.

The opposite may also happen: she may have mastered the interview, but is an incompatible fit in skills or culture. How do we counteract these scenarios?

We have a defined interview process, but recognize that interviews aren’t perfect. While we ask challenging questions, we are trying to establish how the candidate thinks, and won’t disqualify someone for giving the wrong answer. Interviews are nerve-wracking, and people often aren’t at their best in that type of scenario.To quote Joel Spolsky, we’re looking for someone who is “smart and gets things done.”

We’ve set up Xtreme Labs’ work environment to be a good indicator of whether or not an individual is a good fit relatively quickly. If they’re not, then we’ll both know early on. Instead of using the interview as the filter, we’ve built the filter directly into the real work experience.

This is a better predictor of performance because we use real in-work data instead of interview data. Instead of trying to create questions that elicit a symptom of fit, we build performance indicators directly into the culture and work experience.

The reason most companies don’t do this is because they don’t have a really tough post-hiring process to filter people out after they’ve joined. A lot of companies follow a “hire slow, fire fast” mentality, which involves an extremely rigorous interview process but often results in throwing the baby out with the bathwater. Once someone is hired, though, it’s extremely difficult to get them out. (Maybe not so fast after all).

At Xtreme Labs, instead of using the interview as the proxy for performance, we simply choose to use an internship or a few months of work experience, which is much more valid real-world data. We recognize that “the map is not the territory”: a great interviewee is not equivalent to a great employee.

Monotasking = efficiency

Over the past few years, it has become normal for people to do many things at the same time. Are we going overboard with it, though?

We emphasize monotasking: doing one thing at a time. In multitasking, the time it takes to switch between tasks, also known as the context switch, becomes extremely expensive as more and more tasks are attempted simultaneously. When you consider context switching costs in aggregate, it’s clear that multitasking wastes a lot of time.

This is why our process is designed for engineers to monotask. They work in pairs, which keeps them on task. They are accountable to each other and are less likely to get distracted (by e-mail, by the internet, by social networks, etc.). We also work in an open, agile team room, where everyone can see what others are up to.

Don’t misunderstand: we don’t discourage breaks. They’re extremely important to renewal and creative output. However, we believe that if our staff want breaks, they should get off their seats and take a real one. Go downstairs and play pool, or go upstairs and check Facebook at the computer station. Go all-in on the break; don’t do it half-baked.

People often think the end-goal is to be the most efficient multitasker possible. They take joy, and relief, in catching up on their e-mails during a meeting. However, multitasking often leads to messing two things up simultaneously. Should they be rewarded for this behavior?

Don’t let the calendar dictate when to meet

It’s extremely difficult to have an effective 1:1 meeting: people are busy and often don’t prepare for it properly. An effective 1:1 requires both sides to be in the right mindset and to queue up topics, which requires regular bookkeeping. However, these factors usually don’t coincide in the real world, and the answer why is simple:

A career reflection moment is unlikely to happen during a scheduled 1:1

When an issue arises at Xtreme Labs, we want to fix it right away. Being available to host an unscheduled 1:1 meeting means that we set time aside from a regular workday for people to grab us for a chat, while the topic and context are fresh. It’s also possible to combine the merits of an unscheduled 1:1 with a 15 or 20-minute lag time in order for both sides to prepare properly.

We’re fortunate enough to be at a size where we can still handle unscheduled 1:1s. However, we do acknowledge that as companies grow in size, their co-ordination costs increase. More people require more communication. As we grow, one potential alternative to unscheduled 1:1s is to host scheduled 1:1s, and have people cancel the meetings if neither sides need it. When it boils down, everyone has their own management style, and it is up to whatever works for the individual manager.

Final thoughts

Though our processes works in practice, we’re not sure they work in theory. We don’t rely on research; instead, we’ve implemented these processes through heuristics, and our success is a testament to our methodologies.

Filtering out candidates post-hire, monotasking, and unscheduled 1:1s, are just a few of the principles we’ve used to keep up our growth for the past five years. If you’re managing a fast-growing company, we hope they come in handy for you. If you’d like to learn more, or if you’re curious to swap tips, we’d love to chat.

Connect with Farhan Thawar, VP of Engineering at Xtreme Labs, on Twitter and LinkedIn.

About the Author


Scaling Real-time Apps on Cloud Foundry Using Node.js and Redis
Scaling Real-time Apps on Cloud Foundry Using Node.js and Redis

Common applications being built on Node.js like social networking and chat require real-time scaling capabi...

Visualizing the Economic Health of the Nation, One State at a Time
Visualizing the Economic Health of the Nation, One State at a Time

State by State is an ongoing project by Bloomberg's visual data team that offers a comprehensive look at th...

SpringOne. Catch all the highlights

Watch now