Software people don’t just code, it’s a problem solving mindset. The proof is in the cloud. Every part of the IT infrastructure has moved to the cloud because software people have made it reliable, scalable, and secure.
Jeff Lawson, who developed “software people” chops at Amazon, StubHub and others, saw an opportunity to move one of the few remaining industries to the cloud: telecommunications, where installing boxes in closets, and upgrading them every 10 years is still the norm.
In 2008, he and co-founder Evan Cooke started Twilio – a solution that took the communications out of the closet and onto the cloud for others to create new and innovative services. The Smarter Planet blog caught up with Lawson recently to get his views on scaling audio and video on the cloud, its impact on the industry, and innovation.
Why is Twilio moving Telco to the cloud?
As we’ve all observed over the last 10 years, software as a service and infrastructure have moved to the cloud to deliver web services. Telco vendors are the last companies still trying to put hotboxes in your closet.
But today, it makes sense to run one multi-tenant Customer Relationship Management product in the cloud, as opposed to having 100,000 customers running their own installation – each customizing their own instance (making it near-impossible to repair or maintain). In the cloud, this one CRM product can get better everyday, for everyone.
At Twilio we want to move communications from the closet to the cloud, and move its legacy in hardware to cloud-based software.
What are the challenges in moving communications to the cloud?
Think about running CRM in the cloud: there’s nothing natural about companies running it outside their own firewall. But with cloud, they’re being asked to trust a cloud vendor with their data – and that’s exactly what they’re doing.
Communications, on the other hand, begins and ends outside the firewall. It’s about calling, texting, or videoconferencing with someone else. Yet, Telcos still put this media inside a customer firewall (that box in the closet). This doesn’t make sense because real-time streaming audio and video media is the hardest thing to scale.
But it’s exactly what can be done on the cloud. Scaling audio and video is just the beginning. We can use multi-tenant cloud software to help customers communicate better because we can make what they do more reliable, more secure, and more scalable.
For example, when our customers create an application they use us to deploy it on a pay-as-you-go basis. In the legacy world, they would have had to buy channels, and capacity (thinking about what their peak capacity would be and when) before they do anything.
Political campaigns use Twilio to ramp up an enormous telecommunications infrastructure for about two months every two to four years. This “on demand” ramp up (and down) is something they couldn’t do before – or had to spend millions of dollars, and months in advance to set up.
We also have customers like Home Depot using us for their call centers in the cloud. They told as that when investigating a traditional set up, they could not have gotten anything done for months, while paying lots of money; and at the end of that entire process, it still wouldn’t have done precisely what we were looking for. They needed something in composable building blocks.
This notion of a composable enterprise is interesting because, sticking with Home Depot as an example, they got exactly what they wanted in two month, didn’t spend anything up front, and best of all, they own the roadmap. This means that their engineers can ask their call center agents, “What’s your biggest pain point?” Then, the engineers can the turn out a solution in a couple of weeks. You can’t bang a box into the shape you want that fast.
What made Twilio possible for you, and not one of the current heavyweights of the industry?
Part of the answer is in the “innovator’s dilemma.” When building a business on the notion of doing millions of dollars in sales by installing a box into a closet, it’s hard to imagine making a wholesale change over to the cloud. Doing so would mean selling at a fraction of the price, as your customers use it.
Twilio sells at a penny-per-minute, per message. Our customers sign up and get started right away – the antithesis to installing hardware with pres-set software.
But the biggest difference is that Twilio is a software company. We happen to make money because our software makes phones ring. We’re using the power and flexibility of software, with the nearly unlimited scalability of the cloud, to solve problems for our customers. And compared to the box in the closet, that just sits there, we can ship updates to our production services about 20 times per day.
That software thinking, no matter what the area, will outmaneuver the other way of thinking because of this ability to operate quickly.
Our “draw the owl” value actually started as a meme in the early days of the company. We started emailing it to each other, and even printed it out and taped it up around the office. Unlike a typical meme that goes away after a few days, “draw the owl” wouldn’t die.
So, we asked ourselves: “why do we love this message so much?” And we realized that it’s how we view our jobs. We didn’t want to work where there’s already an instruction book. Instead, we thrive in situations where we have to figure out what’s next. We love taking a vague situation and figuring out how to solve it.
What is Twilio’s next “owl” to draw?
We have hundreds of thousands of developers who use Twilio every day. As we see these different kinds of customers – from dorm room hackers, to developers at a Fortune 500 company – we need to be able to speak (to their needs) at the same time. How do we make the simplest, most-accessible communications technology possible, especially to those who historically never thought telecommunications would be accessible to them?
We’re not just solving problems because something’s broken. We’re able to come up with solutions to make things easier, like reducing the number of clicks in a process that the call center agent has to do everyday. So, we need to continue to better-understand their decision-making processes so that we can make Twilio easier for them to use. That’s our challenge.