Why Your Startup Should Scale Out, Not Up

Think differently about growth as your startup gains its footing: instead of piling users on top of each other, spread them out with customizable platforms.

Why Your Startup Should Scale Out, Not Up
[Image: Flickr user [puamelia]]

You’re not a starving tech startup anymore: You’ve got a solid roster of customers, you’re working in an actual office space, maybe you’re even making some money. Sure you’re still small, but you’re moving in the right direction, and you’re ready to grow your platform.


Now it’s time to think about how you want to grow because, believe it or not, there are other ways than just up.

You’re going to see many companies in your position chasing the massive, social media network model of building a monolithic platform and piling users on top of each other–think one foot wide and one mile high. But it can take some time to fix if something that tall topples over on you.

So instead of scaling up, why not scale out?

At educational technology company 2U, we realized right away that we’d be setting ourselves back if we built a single platform and tried to adapt it to the needs of each university that was utilizing it. We knew the Software-as-a-Service (SaaS) model was the way to go because it would allow us to scale out–meaning we could create a platform that was incredibly easy to replicate and then customize individually to each university’s needs.

Sky High = High Risk

One of the biggest challenges of scaling up is that you don’t have a way of predicting how your system will react under stress. As we’ve learned with social media networks like Twitter, you risk outages, bugs, and other issues because you haven’t been able to test your system with all those users, and you may not fully understand how each of your new arrivals will actually utilize your system.

As a result, you’re forced to learn based on mistakes, which isn’t the best option when your reputation is at stake. From the start, we couldn’t afford this possibility at 2U, because there was already a stigma around online education that the technology just wasn’t capable of facilitating real education outcomes.


Scaling upwards also makes it difficult for your tech support team, because they won’t necessarily have all of the answers at hand. Think about this: your platform gets some great press and goes from 1,000 users to 100,000 users in a day, and immediately it crashes. How will your tech support know what the issue is when your platform has never seen 100,000 users before? Had you scaled out, you would have been able to test your system and put processes in place for when issues eventually occur.

Speaking of your tech team, scaling up can also make it difficult when you try to onboard new employees later on. If you scale up, you’re waiting to hire until the growth of your user base justifies the addition of more employees. So you hire three new engineers, but then what? You still have to train them, and it could be weeks before they are really able to help.

If you scale out, build up your processes, and stick to a software development life cycle, your onboarded employees will fall right into place and have very specific roles. You can have someone start on Monday and be fully up to speed by Wednesday.

Scale Out: Build Your Base

Think of scaling out as building the base of a pyramid: it’s strong, sturdy, and prepares you to build on top of it later on. It’s the foundation upon which everything else can be built, and you know that it will hold. By doing this, you have the assurance that no matter how quickly you may need to grow later on, you’ve built your architecture in a flexible and intelligent way that will allow you to scale without taxing your team or endangering your roadmap.

Scaling out also provides a sense of predictability because your roadmap generally stays intact. When you tell your users a feature is coming out in Q4, you can be sure of keeping that promise because you know your system and are confident you’ll stay the course. With a well-architected platform and a scaled-out team, if there is a bump in the road, you’ll be able to pivot rather than change course completely. This kind of consistency is paramount to any growing businesses.

At 2U, we grew from five virtual servers to 600, but we only went from two engineers to four in order to manage the whole shebang. Why? Because we scaled out, put our processes in place, built our strong base, and created a manageable solution before even thinking about piling on users.


We Did It–You Can Too

A lot of people ask me how we handled going from a startup to a public company in such a short period of time. To be honest, not a lot of things had to change, and it’s because we scaled out early on.

We had our security protocols in place, we had a secure architecture, and we did it all from the start. Yes, we’ve had to hire more people, but we’re doing things the same way as we did them a year and a half ago.

The point is, there is no rush to scale up and grow skyward as quickly as possible. The smart decision is to scale out and set yourself up early to be able to handle anything that may come at you down the road. Maybe it’s a million users. Maybe it’s an IPO. Whatever it is, you’ll be ready.

James Kenigsberg is the chief technology officer and a founding member of 2U, Inc. James leads an experienced team at 2U that has built an industry leading technology platform that transforms online education for students around the globe.