Digital transformation is an ongoing process. Businesses must continually adapt and grow as technology changes, and customer expectations change along with it. While this can seem intimidating at first, it ultimately opens up a world of possibilities and opportunities, allowing teams to test, fail, iterate, and repeat — while driving your business forward in unprecedented ways. In this series, we’ve explored why digital transformation is essential for companies today, and focused on the importance of performance and responsiveness. In this post, we’ll explore one more crucial element: experimentation.
What do we mean when we talk about experimentation? It goes far beyond throwing things against a wall to see what sticks. To continue to grow your business in a digital world, you need the flexibility to experiment safely and effectively. That means giving your teams the freedom to fail, while ensuring that the inevitable failures still occur in a secure environment.
This also speaks to a philosophical shift: teams need to feel empowered to test, and take ownership of ideation, validation, and iteration. When nothing is off limits (within reason), the ability to test assumptions can open up surprising doors that accelerate businesses forward in innovative ways. And that means setting expectations correctly. When a test is sent out to the world and the results come back, it’s important to embrace transparency of both the good and bad, as there are learnings in both instances. Sharing the results with the organization helps teams across an organization understand why things are working, and what takeaways they can embrace moving forward.
There’s also an aspect of speed here. Tests can’t take too much time to roll out, gather data, and inform next steps. It’s simply not effective. One of the biggest benefits of good experimentation is efficiency, so you can get results and implement changes that drive results, quickly. That requires a thought-out plan, and the right technology.
Reddit is one of today’s leading news aggregators, with an vocal online community that’s active 24/7. Last year, they unveiled sleek new profile pages, and decided to strategically launch it as an opt-in beta. They did so to ensure profile pages rolled out safely, collect early engagement data and feedback, and enable a seamless user experience before transitioning to the new design. But experimentation doesn’t just start and end with features: it applies to all aspects of your technology processes and infrastructure innovations. Which is why their team tested and validated the new architecture in a controlled manner.
Because this was an opt-in feature, they needed to preserve users’ original profiles, and serve the right profile experience based on whether or not a user had decided to participate in the beta. By leveraging Fastly’s edge cloud platform and URL redirects, their development team wrote code at the edge, so they could redirect the profiles of 75,000 users who opted into the new profile. Reddit was able to serve the right content to the right person in a controlled manner, and eventually shift everyone over cleanly to the new experience — all while providing users a seamless experience, without impacting performance.
This experiment was effective because it allowed Reddit to move quickly, and engage with customers iteratively instead of pushing a new experience to everyone overnight, and ensure things rolled out safely. By empowering their own teams at the edge, and encouraging their users to participate, they successfully validated a new experience that’s essential in Reddit’s ecosystem.
Most news organizations know that a presidential election is often the biggest news day. This was true for The New York Times too, and because of those huge traffic spikes, their site went down in the 2012 election. The CTO that joined since that time, Nick Rockwell, wanted to ensure they mitigated those risks as the 2016 election approached. This meant taking stock of how the company viewed failure and risk, and experimenting with a new approach to ensure everything ran smoothly for their users.
As Nick put it, “Risk taking can only occur within the space of the unknown.” The team wasn’t afraid to shake things up and go beyond the status quo. First, they made moves on technical infrastructure, consolidating data centers and nixing the tech they weren’t using. They also integrated with Fastly for better performance, scaled caching, and more protection against DDoS attacks. What seemed like a gargantuan task — shifting their CDN just six weeks before the election — quickly became a validated and essential part of their stack.
With election night approaching and their technology moving into place, Nick decided to conduct a “pre-post-mortem” based on what occurred in 2012. Alongside the team, he defined what failure looked like and how they could address it better. Together they realized responsibility was too diffused across developers, which made it too difficult to take ownership and prevent those failures. After those conversations, they made responsibilities clearer, and used the debates leading up the elections as practice runs.
In 2016, things went off without a hitch, and the team themselves felt more confident, better able to experiment, and comfortable saying what was working and what wasn’t. The New York Times embraced failures of the past and leveraged that in their planning to make quick decisions and deliver incredible results.
Experimentation goes beyond launching new products, and it requires more than reexamining underlying infrastructure. When businesses embrace the culture and spirit of experimentation, there’s a perceptible behavioral shift. Teams feel more confident about thinking big and taking risks, and with the right technology behind them, they can move quickly and securely to test those ideas. Failures can serve as key guiding points, and successes can rapidly move the needle forward.
As technology continues to evolve, it will only become more important for businesses to encourage this kind of experimentation, and ensure their ongoing digital transformation succeeds.