For the past year, I've helped a few Private Equity (PE) firms evaluate small companies (around $10 million to $20 million Annual Revenue) for acquisition into one of their portfolio companies. Here's what I've learned, in somewhat long form because I didn't have time to make it short form. These are my thoughts and opinions, yours will likely vary.
First, the cloud is not innovative. Every company today has a cloud-first approach, and their budgets show it. Why did you raise so much money? Because cloud computing is expensive. Every single CTO brags about the discount they negotiated with one of the big four cloud providers (AWS, GCP, Azure, Oracle). In fact, from seeing enough of these, they all basically got the same discount. If your first argument is how much of a discount you got on the cloud, perhaps that implies it is too expensive. Meanwhile, the cloud providers are rolling in profits. The cloud lets you move fast and try things. It also takes serious cloud engineers to make it reliable, high performance, and cost effective. PE is pretty brutal here, because they're considering negotiating in bulk with the cloud providers, and want at least 3-5x the discounts of their portfolio companies. Your $10m budget isn't enough to move a cloud provider to "break out the pencils" and figure out the best deal. PE's $150-300m annual spend across all their portfolios is a big stick.
What is innovative are the companies who run docker/podman/virtual machines across multiple providers like Linode, DigitalOcean, Vultr, and the like. These providers are sometimes referred to as the AltCloud or second-tier clouds. The most innovative companies I've seen did everything with rented dedicated servers at providers like Hetzner, OVH, or FDCServers. Having budget constraints and hardware limits forced them to build just enough system to get the job done. Yes, there are budget constraints in the cloud, but typically they are far higher limits than on AltCloud or dedicated servers.
Docker and Docker Swarm seem to be more aligned with the more innovative companies than does Kubernetes (K8s). The feeling is that when you hire a team of K8s experts, your cost savings is lost to expensive talent. It's great that K8s experts are making $300-500k/yr, but that's an expensive asset to a $10m/yr company.
Second, the companies getting the higher valuations are the ones that are bootstrapped or angel funded and nothing else. This is getting out of my area of expertise, but having a tight budget, sharp focus on profit margins, and building just what's needed seems to work best. They probably aren't going on Hacker News or the conference speaker circuit and bragging about their systems, and the tech debt is large. However, they have exactly what's needed to grow the company to $10-20m/yr in revenue and didn't spend on "potential future growth". The infrastructures, app environment, and entire tech team are lean. They have little debt alongside easy cap tables and balance sheets.
Third, none of these companies follow "best practices". I always consider "best practices" to mean "minimum standards". Yes, they are SOC2/PCIv2.x/ISO2000x compliant, but they don't take a best practice blueprint and just do that. They know their needs deeply and implemented to what they need, not what someone else says is now fashionable. The teams cherry pick the best they see around and/or develop their own to just what they need, not the whole codebase/plan/practice. They are anti-trend. It's not sexy to say, "we run QEMU VMs in production", but it's wildly profitable. Almost all of the companies have a development and production environment. Very few have a staging environment. Hotfixes galore in production happen as well. A "move fast and fix things" sort of approach.
Fourth, it helps to be a company from China, Singapore, Scandinavia, Israel, or the Baltic States. I don't know what makes these countries different, but their people are very smart, very quick to implement something, and iterate quickly once in production. Tell me which pitch sounds typical to investors:
Company One: "As part of the roadmap to triple revenue and expand our systems to scale to handle 5x the transactions, and it involves a fast queue, columnar database, and serverless lambda functions to using (AWS|GCP\Azure) technologies in the next 6-12 months. We're currently hiring some ML, K8s experts but are having a tough time competing against Google, Amazon, Meta, and Apple for candidates."
Company Two: "We migrated to a new database on a new server over two weeks to handle the doubling of customers tied to our advertising campaign. The SQL functions are faster and we hired a contractor to optimize them over a two week period. We hired another contractor to help optimize our queuing code. We migrated our servers at (Hetzner|OVH|Linode) from older Intel Xeons to brand new AMD EPYC-based systems over the past month, which gives us 25% more capacity for surges and future advertising campaigns."
In my experience, American companies are Company One, always selling what they're going to do and how great it will be when done. PE is buying up Company Two type companies left and right because they executed and implemented far faster and with vastly less resources. Generally, Company Two type companies are from China, Singapore, Scandinavia, Israel, or the Baltic States. They all have a pretty sane quality of life, but they deliver quicker and just in time versus over optimizing for a future that will likely never come.
In my view, the unsexy, boring technology is what supports the business. The goal is to make a profit and provide a service. There are lots of SaaS companies in the mix, but even when not, the ones with the highest purchase prices have a mix of boring technology and fantastic customer service. When you have customer reviews which are, "their product isn't exactly what we need, but they respond quickly to feature requests and their sales/product support is great", that's a winning combination for 10-15x multiples of revenue as a valuation for PE.
Another view is that the more automated the sales process, the faster the companies scale. Following up with great after-sale support hooks customers and keeps them around. I've seen company after company hit $10m/yr revenue and switch to an "enterprise sales model", which drastically slows down their growth, dips their profits, and cuts out a huge segment of their market. No one wants to talk to a sales person and go through a multi-step money extraction process before they even get to try the product. Click here, sign up, you get a free trial works wonders. And follow the hell up with empowered people who know the product and you have a goldmine. Smart, technical salespeople are a goldmine, literally.
There is a lot of "cargo cults" in startups or small companies today. They do what they see in the news. I think it's a better indicator when I meet a CTO and they're almost embarrassed at their stack. One quoted to me at the start of their deck, "We couldn't afford a Porsche, so we built the Toyota Yaris of infrastructure." I love that. They also had $8m/yr in revenue and were bought for over 10x valuation. Mr Yaris can afford that Porsche now. It's symptomatic of a successful, lean company that can do that. Some PE want the flash, they want the name brand ML or K8s expert. The vast majority want a profitable company that can ramp up revenue with their money.