How to (Fail to) Convert From Services to SaaS

Starting in January of 2020, I began to take company strategy more seriously, instead of secondarily. I hired my first boss in my software career as my COO (and effectively a consigliere of sorts), brought on a silent partner as an investor, and together we laid out a 5-year plan. It had a few different parallel strategies in it (which itself is a yellow flag, at least), one of which was a spectacularly fun crypto strategy, which failed (not spectacularly, but fully; I'll share those gory details another time), but the one I want to talk about today was the most ambitious—converting the company from being a service provider to being a SaaS product company. 

I'm not a famous person in software—exactly the opposite, really—but I've been fortunate to rub elbows with some legends. One of my favorites is Joel Spolsky. He was one of the co-founders of my favorite software conference (BoS—Business of Software), which I've attend annually since 2008, so in addition to reading a ton of his work over the years, I've gotten to hear him speak live a few times, which means I've gotten to chat with him a bit at the always-present cocktail hour. One year, I asked him if he had any plans to offering consulting services via his software product company, Fog Creek Software (whose products we used). He answered, without hesitation, "Services don't scale."

Now he tells me! I'm kidding, of course. I'd heard this before, and after delivering services for the first decade or so of my career, I was viscerally aware of how relatively laborious it was to trade time for money. But it's all I knew, so it's all I did.

But something about hearing this from Joel always stuck with me. I was a little slow to make certain decisions, so I sat with this for a while. Not too long—just about 10 years or so—but by the late 2010s, as I was remodeling my company from top to bottom, I was more and more ready to take on the challenge of converting us to a SaaS company. My partners were on board, so after 2 years of crypto shenanigans from 2022 to 2024, we got serious about SaaS. 

Normally, you create a product to solve a problem, and then sell it to others. We sort of did this, but not quite. As I've written about before, in modernizing the company, I'd given the team an assignment to build a product, but more as an exercise in building, rather than because we'd identified a problem that needed solving. However, I did have one strategic eye on the future, so I picked an assignment in an area of the Internet I knew was full of problems—cloud storage. Specifically, I asked the team to build a tool to manage Amazon's S3 offering (Simple Storage Service). 

Even the first version was a great success, especially for a self-funded internal project at a small, bootstrapped company. It was a simple product to start, but the design was top-notch, so we knew we could build a lot of functionality upon that strong foundation. 

Estimates are that there are multiple zettabytes of stuff in S3. I was convinced that customers would pay money to manage all of that elegantly (for the record, we refuse to build software and systems any other way), so we launched the first version in the AWS Marketplace (the "app store" for Amazon's cloud) in 2024. We knew the product needed enhancements beyond the basic first version, so we created a roadmap and got down to the real work of making it an enterprise-grade piece of software.

We built all the stuff you'd need—security (e.g., SSO), speed (our "Fast Buckets" release can search hundreds of millions or even billions of files in seconds), size (need to upload a 30 TB file? No problem...), organization (our Tag Explorer is unparalleled, as is our Advanced Search feature). We've even added natural language processing and some simple AI features, so you can "talk to your files." As I'm writing this, we just shipped some of the most advanced security available for a SaaS product, and we're next working on allowing you to connect the product to your favorite AI (due to ship in the summer of 2026). 

There's a rule of thumb that you're not supposed to fall in love with your own product, because it can cloud your objectivity (see what I did there?). I think I'm still pretty objective about the product's strengths and weaknesses, but I will say this—I am proud of CloudSee Drive (finally!). We try to hold ourselves to the highest possible standards, even for a tiny company, so it took a while for me to get to this point. But it's really excellent sofware. 

But wait...this is supposed to be a story about converting from services to SaaS, so what about, you know, SALES! How is this fabulous technical feat selling? 

Timing is everything, as they say, and I'm not going to beat around the bush—this is terrible time to be selling a new SaaS product. We've been reading the markets for over a year now, and the unavoidable truth is that SaaS is dying. The "-ing" is important—approximately 100% of "X is dead" arguments are wrong because they don't acknowledge it's a process, not an on/off switch. More specifically, certain kinds of SaaS dying FAST—especially products early in a lifecycle, without the moats of an established customer base, long history of customer data, or the other advantages of incumbents. Even these aren't guarantees against a slow death—just look at the stock prices of large, public SaaS companies. It looks like another dot com bubble burst. The software industry is priced lower than the S&P 500 for the first time...ever. As Jason Lemkin summarized, as of 2026, investors would rather own...ANYTHING ELSE, other than software companies. 

So our so-called "failure" to convert to a SaaS company is sort of like failing to be born in a better neighborhood. Not much you can do about it, other than fight your way out, and move on. Which is exactly what we've done. And our SaaS building has given us some of the strongest muscles of any project we've worked on over the years. This makes us very much ready for the AI fight that's waiting for all of us. Bring it!

Post script: We're still selling. (Knowing when to quit is part art, part science.) Just re-allocating staff, and re-prioritizing our roadmap. We've been on a parallel strategy for over a year now, and our SaaS product and our AI-native transformation have been mutually beneficial to each other. While SaaS is dying, the strengths developed from a strong product development cadence will be critical in the AI era. We're excited. 

I am Sigmund. Thanks for your time.