Should You Optimize Costs or Not?
In honor of Benn Stancil: It's Wed on Twitter, Let's Fight
A few weeks ago Anna from Techcrunch included my thoughts on cloud infrastructure and cloud costs. You can read the post here.
What followed was a very interesting discussion brought about by Ben, Founder of Vantage, an awesome company helping customers manage their cloud costs. I encourage you to click on the thread to read through the responses between me, Ben, and Anshu Sharma, Founder of Skyflow.
After some great discussion on various topics, here is a conclusion that I drew from it (whether correct or not is for others to decide).
So Cost Optimization - What’s The Deal
As always it’s a nuanced topic that depends a lot on each company. But let’s take the comment I made specifically above, “In early product or go-to-market stages, optimizing cloud spend should be the last thing on a founder’s mind”.
At the early product stages, cloud cost optimization can literally kill a startup outside of utilizing cloud provider credits. Reason is any bandwidth spent not building product, talking with customers, and iterating rapidly delays product market fit at a critical time for the capital. Resources are scarce, both in terms cash but also people and time. While cloud cost optimization may buy 1-2 months of extremely valuable runway, in the early product days, the time and mindshare spent on it is highly unlikely to make those extra months valuable (perhaps a vendor may help here, more on that later).
In the early go-to-market stages, this is equally true. Cost optimization at this point encompasses much more than cloud spend, it also deals with gross margins. The cost of goods sold to sign an early customer is invaluable. The first 5 customers will help the startup grow for a long time in the future. So early optimization on anything pertaining to gross margins is a risk. Many of the most successful startups have started out with their first customers using the product for free (i.e. negative gross margin) or very small ACVs. Not only are we talking hosting costs here but also professional services, customer support, etc from the early team. The learnings from these early customers are going to set the stage for building a product that 100s of other customers with the same pain point will benefit from.
A far more useful an exercise is focusing on the profile of that customer.
Do they have the pain point the product is solving
Are they using the product ingrained in their daily workflow
Is there a close feedback loop with the customer
Does the customer understand the current maturity of the startups’ org (i.e. not asking for RBAC, SSO, SOC2 too soon)
Will the customer be a valuable reference for other customers down the road
If the customer fits the above criteria, I can list countless examples of startups that literally give these customers product for free and even as they scale still have these customers on tiny ACVs mainly because they provide more value than just the revenue.
Is Using a Vendor for Cloud Cost Management Useful
The answer is a resounding yes if the vendor can abstract away the mindshare and time needed to optimize cloud costs. Then it becomes a no brainer for customers to use because who doesn’t like saving money! The only reason why I still may caution against it is because it becomes an easy fallback for teams during hard times.
I’m seen it many times. A startup with less than $1M ARR is not able to find customers and scale efficiently. So they turn to what can be immediately controlled (reducing costs to service those customers) vs trying to figure out the much harder questions. This tends to lead to those startups achieving a local maxima.
I am a stanche believer that some products can be so good as to fade into the background while helping customers with the pain points. I think founders like Ben at Vantage are working on exactly that. And the analogy to draw is also to security. Very few founders care about security in the early days because if someone hacks the startup, they probably don’t have much sensitive data/access to hack anyway! However, startups like Snyk, Jit and more have built successful businesses by abstracting away the need for developers to focus on security, fitting into their core workflow and providing identification, remediation and controls automatically. They make a core level of security “easy” for customers to achieve and that’s what great cloud cost management products can do as well.
Local Maxima
In the end, every company needs to be aware of where the big problems are. As companies scale past ~$5M ARR, a lot of those problems switch from just figure out how to get more customers (still a huge focus!) to making sure that you’re earning a good enough return on current customers. This is where cost optimization across the board starts to become a big and necessary focus. Not just cloud costs, but gtm costs, tech debt costs, and plenty more.
But at each stage, teams need to focus on what moves the needle the most. Optimizing costs outside of headcount too early can take valuable mindshare and iteration from building a product that clearly resonates with customers and matching it to the gtm. It can take mindshare away from building a new feature or taking away a feature. It is an easy fallback for startups to say things aren’t going well but we’re extending our runway by optimizing our costs. Outside of headcount, until a certain scale, the most important problems are more around how to generate customer urgency, how to make the product sticky, what features will move the needle for customers, product differentiation and countless other things.
Stuff I Enjoyed This Week:
A great explanation of the benefits of Transformer Neural Network Architectures
The US Strategic Petroleum Reserve is at some potentially depressed levels. Will be interesting to see how this plays out if oil is needed in the future
Love this SaaS Funding Napkin from Point Nine Capital h/t Sheel
Masterclass discussion on the dynamics of the Cable industry (the OG infrastructure for the Internet)
Jeff Bezos’ recommendation for consumers in this economy, notable given Amazon probably has the best dataset in this area
Bezos recommended American households delay big-ticket purchases such as new TVs, refrigerators, and cars, given the risk that economic conditions worsen. Similarly, he suggested small-business owners consider holding off on investments in new equipment, and build their cash reserves instead. "Take as much risk off the table as you can," he said. "Hope for the best, but prepare for the worst.”
