<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1310905&amp;fmt=gif">

How 2020 Turned Cloud Unit Economics Upside Down

|November 3, 2020|

Do you know how much it costs to operate your software per customer, at each pricing tier?

The past six months have been highly unusual to say the least.

For some companies, the disruption and shift to work from home meant 100x in usage almost overnight; for others usage completely collapsed in the same time period. These sharp swings in utilization have been disruptive to cloud unit economics.

Both scenarios are risky and require extreme elasticity from the technical infrastructure.

The good news, for companies who already built highly scalable cloud-native infrastructure, is that this is exactly what the cloud was designed for — scaling up quickly and scaling down just as fast.

The bad news is that many companies were equally prepared to take advantage of the cloud’s elasticity, and not all companies fully understand which cloud costs are inelastic, both generally and in their specific use case.

Now, more than ever, getting insight into how your cloud costs relate to usage — and what to do about it if they don’t scale down enough — is key to building a resilient digital company.

Here’s how understanding cloud unit economics can help companies weather the extremes of 2020.

Growth Is Good for the Bottom Line, Except When It’s Not

Let’s address the less intuitive problem first: In two weeks, usage has increased by 100x. Here are the questions business leaders and engineering teams are going to need answers to:

  • What percentage of new users are using the free tier? Is that out of whack with the usual user breakdown?
  • Are users behaving differently than before? If so, has the cost of operating the software increased, decreased or stayed the same?
  • How much does it cost per user at each price tier? Have costs increased in any way that seems unusual?
  • Have there been any cost spikes? If so, what caused them?
  • How resilient has the system been technically? Has the team had to invest more time than usual on incident management?
  • If users increased by 100x, have profits increased as well?

Increasing the number of users is only good news if each new user can be served profitably. Companies don’t want to get into a situation where they are scaling up but losing more money the more customers they have or scrambling to address the tech debt from an inefficient system in the midst of a usage explosion.

Scaling Costs Down Requires Elasticity

If your customers evaporate overnight, ideally your cloud costs will too. If they don’t, it’s because you have some fixed capacity, often related to storage, that does not scale down automatically.

While scalability and elasticity are core advantages of using the public cloud, it’s also a mistake to think that they are out-of-the-box features.

The reality is that cost elasticity has to be a part of the architectural design — it's not an automatic result of moving to the cloud. In most cases, though, as long as you have the system visibility to see where the costs are coming from, you can scale down manually.

If users decline, a detailed understanding of your unit economics, your fixed costs and elastic costs and how to reduce costs as much as possible during the downturn can be the difference between scaling back up in six months or closing your doors forever.

Understanding Your Free Tier Is Essential

From a costs perspective, a free tier can be the most dangerous, because the company isn’t generating any revenue as a result.

In March of 2020, many companies saw their number of free tier users increase faster than the paid tiers. Drift, for example, saw free-tier users of their chatbot increase significant — and suddenly the unit economics were totally off. Drift already had visibility into their costs, so they were able to:

  • Recognize that having so many free tier users was a problem for the company’s margin immediately, rather than at the end of the month when the AWS bill showed up
  • See where the problem was — in the chatbot tool — and have specific conversations about what to do about it
  • Re-architect the chatbot (instead of eliminating it from the free tier, which could have alienated users), reducing the cost of running the chatbot by 80% and making it viable to offer for free

Too many free tier users can tank the company if cloud costs increase unchecked.

Building a Resilient Business Means Understanding Cloud Unit Economics

Visibility into cloud costs is critical for companies whose primary product is software to navigate rapid changes in the marketplace. Whether you’re a video conference solution or an online restaurant reservation platform, 2020 has brought major challenges and a need to adapt quickly.

Given the economic climate, it’s important to spot cost bugs immediately, before they add up to thousands of extra dollars in cloud costs. As companies develop new features as quickly as possible to match changes in their users’ needs, they need to make sure they are innovating profitably and that the unit cost of operating the feature is lower than the amount they can charge for it.

CloudZero Can Help

With CloudZero, companies can automatically measure cloud cost across flexible dimensions, including cost per customer, product feature, dev team and more — all without manual tagging effort. Try it out now.




Subscribe to blog updates