For startups, growth still trumps cloud cost control

There’s room for startups to cut their cloud costs, even if they have to balance the implicit costs of doing so, such as the time required and the potential for slower development. The question then becomes: How much of a priority is finding incremental savings for young tech companies?

A recent survey of founders by TechCrunch+ indicates that a change in investor expectations is spurring startups to take a closer look at their cloud spending and move away from a position more focused on speed than cost efficiency — just not too much.

The changing economy and the resulting impact on both venture capital availability and the price of money keeps showing up in our investigative work. Put another way, rising interest rates are having a knock-on effect on cloud spending at tech companies, and therefore, slowing growth at public cloud incumbents.

TechCrunch+ also recently asked startup founders if new startups should pursue a multicloud strategy. They answered mostly in the negative, with some caveats regarding edge cases.

This morning, we have a sheaf of perspectives to digest, building off our work in late 2022 aiming to understand how startups picked their first major cloud provider and why.

Finding fat to trim

Last year, Boldstart Ventures partner Shomik Ghosh told TechCrunch+ that for startups still “in early product or go-to-market stages, optimizing cloud spend should be the last thing on a founder’s mind besides utilizing as much cloud resource credits as possible.”

However, speed to execution can lead to costly bills later on and some startups today are working to curtail their spending. Joe Mainwaring, director of infrastructure at WorkTango, and Tobi Knaup, CEO of D2iQ, estimated that their wasted cloud spend made up only about 10% of their total bills, but other startup founders and execs said they had more room to trim.

“In my experience, most companies do waste quite a bit of money in the cloud,” said Ed Thompson, CTO of Matillion. “Infrastructure and cloud accounts can proliferate very quickly if not controlled, especially in a development-focused company. In a startup, I think it’s important to get a degree of control early, but not at the expense of innovation. It’s more important to keep an eye on cost rather than simply lock everything down. The good news is that when companies achieve the scope and scale that allows them to concentrate on cloud cost management, significant savings can be achieved with little effort.”

Just how little would that effort need to be? Ed Frederici, the CTO of Appfire, says his company’s cloud costs today are “between 20% and 25% more than they need to be,” but the company could save 8% to 12% of its cloud spending “with minimal effort and within the remaining time left in 2023.”

If there was that amount of overspend occurring, why did it persist as long as it seems to have at Appfire? Frederici has a clear answer: “Appfire has prioritized time to market over infrastructure efficiency.” He argued that “organizations shouldn’t over-rotate” on their cloud spend during an economic downturn. At a fundamental level, he said, Appfire “believes that to fuel long-term success, organizations must prioritize innovation over cost optimization, within reason.”

This is all very reasonable. As a startup grows its product and expands to new regions, it will have an increasingly complex and expensive cloud setup. Companies with excess cloud spending will be faced with a choice: Devote resources either to reduce costs or to drive new product development. If there is a lot of stuff to build and demand for it, well, you can see how cost reduction measures might be deferred.

That was especially true when capital was cheap and venture dollars flowed like cheap wine at an undercatered meetup. Today, things are different, and finding ways to reduce costs seems to be a common activity.

Mainwaring said that at WorkTango, they are “very on-top of our cloud cost and optimize quickly when we observe spikes in usage.” Meanwhile, Knaup’s team at D2iQ has “spent a few years optimizing our cloud costs and already cut them down by more than 70%-80%, so we don’t have much left to optimize.”

Siddharth Ram, the CTO of Velocity Global, told TechCrunch+ that one good way to limit cloud costs is to shake up the way that customers purchase computing power. “I think about 6% to 8% of our cloud costs are a waste. A key way we are reducing costs is by moving to usage-based metering models (for example, AWS Lambda) versus time-based models (for example, EC2). Unless a team is operating at a really large scale, usage-based metering is a win.”

Appfire is doing similar work but with a notable focus. “Appfire continues to make adjustments and optimize cloud costs as our clients’ workload and usage patterns become clearer,” Federici said, but with the caveat that his company will forecast cost and is conscious of the impact. “However, Appfire emphasizes speed, agility and innovation above early cost optimization,” he added.

To summarize, it appears that startups big and small are taking a closer look at their cloud costs but are mostly uninterested in slowing down the pace at which they build. Some cost control, then, mixed with sustained product development. But even some pullback from the huge cohort of private tech companies means, in aggregate, slower growth at the big cloud companies, which TechCrunch+ has reported on extensively.

When we polled these founders, I had expected a greater bias toward savings over speed. But I should have remembered that cost cutting has never taken a startup public. It’s rapid growth that makes for an attractive tech IPO, and that means quick development cycles over wringing dollars from an AWS bill.