You know that feeling — the one where you’re certain you’re overpaying for something and you’re just not sure where the hidden costs lie? At Deft, we make that feeling go away. We meet our clients at their most suspicious, working with them to go through their cloud budget line by line until they understand where their money has been going, what can be reallocated, and how they can take control of their infrastructure costs again.
Cloud computing is a no-brainer for businesses in the early days, when there’s uneven demand and a need for rapid scale. But over time, maybe a few years in, many companies find their needs become more predictable. The economies of cloud hyperscalers don’t necessarily make sense anymore. It’s time to do a deep dive, to consider what could leave the cloud and how much money could be saved by using dedicated servers for predictable loads.
Navigating the cloud repatriation process
Deft knows how to convert complexity into success for companies of every size. It helps that we’ve been in the cloud business — and left it entirely. We know how the cloud works, but we’re not dependent on cloud hyperscalers to send us business. That gives us the rare ability to be thorough in our assessment and neutral in our recommendations.
So how do we know what infrastructure mix makes sense? When we bring on a new client, we start with a high-level overview and discovery phase. Consider it like putting your cloud bill through an x-ray machine. We’ll help you see behind every line item, so you understand whether what you’re getting lines up with what you’re paying.
A full assessment of your company’s cloud architecture, includes but is not limited to:
- Production and development workloads and related infrastructure.
- CI/CD workflows, platforms and practices that are in place and planned.
- Additional cloud services and on-prem applications and environments that may be relevant to developing a comprehensive understanding of your environment.
Often, these assessments are the first time anyone has thoroughly reviewed the infrastructure as a whole. This is generally where we find all the things that someone thought they would go back and clean up sometime in the future — old workloads that were left to balloon over the years, quick fixes that were meant to be solved when they had a little more time. It’s understandable, infrastructure oddities will accrue at a fast-growing organization just like anything else will. But at a certain point, someone’s going to have to go back through and clean it up.
Leaving the cloud with a plan for the future
Once we get the bleeding stopped with discovery and remediation, we can start to proactively design an infrastructure plan that makes sense for your business. Usually, this isn’t going to be all in cloud, or all off cloud, but a balance of what works best from both. Let’s be clear: We are not anti-cloud. We just don’t want you overpaying for cloud services and capacity you don’t need.
As we work to design a solution, we’ll sit with you to:
- Identify pain points related to infrastructure, management and monitoring.
- Identify areas that can be retooled or modernized to meet internal and external security goals and to increase and improve billing transparency.
- Identify areas that can be retooled or modernized to provide automation for system scaling to meet future demands, to optimize deployment efforts that allow for improvements in the speed, execution and debugging of development efforts.
Ultimately, we want to get your business to a point where the quickest infrastructure is also the best infrastructure — so no one spins up more “temporary” solutions that end up sticking around.
IT optimization in practice: A big shift for 37signals
One of our longtime clients, 37signals — the makers of Basecamp and email product, HEY — consistently and constantly questions how it does business. The leaders and operations team there love to look at long-standing business logic in new, better ways.
Initially, the company kept everything in the cloud because it was essential to staying nimble and scaling up when needed. However, as 37signals’ business became more consistent, it still had billowing cloud bills. It was time for a deep dive and reassessment. The result was a big do-over and a shift to owned servers managed by Deft because on-prem colocation is always going to be cheaper than cloud for predictable workloads.
The cost savings have been eye-popping. Transitioning to an owned model was expected to generate operational savings of $7 million over five years, but the savings keep racking up. Co-founder David Heinemeier Hansson (DHH) wrote, “Our cloud spend is down by 60% already, from around $180,000 a month to less than $80,000. We don’t have to squint hard to see the eventual savings climb all the way up to about $2 million a year. That would be $10 million over five years.”
How we helped PicsArt leave huge cloud bills behind
Another case of cloud budget run amok is PicsArt, a photo and image editing app launched in 2012 that included a social media element. The tool lets users edit or add elements to photos without needing Photoshop expertise.
The app was popular from the jump, downloaded 35 million times in its first year. Fast-forward to 2020 and the start of the pandemic: PicsArt traffic shot up 40%, and today the app has more than a billion downloads and more than 150 million monthly active creators. This is great news, right? Yes, but processing millions of photos can quickly overwhelm most infrastructure and cause data transfer costs to surge.
Cloud computing was pinching PicsArt’s ability to grow. The company needed a few things:
- A data center for an always-on, steady-state application environment.
- An object storage platform with no fees for data egress.
- Secure connectivity to public cloud services to handle bursts of activity.
- A trusted vendor that picks up the phone to provide expert support.
That’s when PicsArt met with us. The company received an unexpected bill and an unsatisfactory response from Google — the huge month-over-month surge chalked up on the bill to “data transfer.” The rationale PicsArt received for the big bill was not rational: Google told them it wouldn’t explain the issue unless the company coughed up more money for enterprise-level support.
Totally leaving the cloud wasn’t the answer here. Hyperscale cloud providers still do a great job of delivering content to PicsArt users. Continuing to process images on the cloud, however, was going to make the cost of growth unsustainable. Deft is able to provide PicsArt with physical servers for image processing, connected directly to the cloud so the company can enjoy all of the speed and none of the surprise bills.
Guiding our clients to the ideal infrastructure setup
Deft can help companies when they need to suddenly scale, and we take pride in our role as trusted advisors. We plan for all the possible situations, and we pick up the phone.
Our experts know a lot — a lot about servers, a lot about the cloud, and a lot about making a cloud exit. Migrating out of the cloud is tough to do well unless you’ve done it, and Deft has. You can check out our free IT infrastructure reviews and see how well your cloud and on-prem infrastructure aligns with the industry’s best practices.
We know cloud costs spiral and that there are hidden costs. We also know plenty of vendors and their sales reps won’t tell you about those costs, happy to let you pay more for cloud services that aren’t meeting your needs. At Deft, we’re never quiet about saving you money. And we’re never quiet when you call.
