Elasticity: One of the core promises of cloud. The ideal cloud workload is supposed to flex, allowing customers to better align IT expense with revenue or changing performance demands. And for in-house developed workloads, this is certainly achievable.

Developing and maintaining software and infrastructure comes with a lot of overhead though, hence one of my favorite sayings- “The first rule of engineering: Don’t build it if you can buy it.” This is where Software as a Service (SaaS) comes into play.

I’ve been wrangling cloud solutions for nearly 15 years, from multiple datacenter sunsets to multi-million-dollar software deals, and one of the many truths I’ve learned over that time is simply this:

SaaS often isn’t as scalable as it appears to be.

SaaS is an attractive solution on the surface, and is certainly a great fit in many scenarios. If a workload isn’t core to your business, then why invest more human capital into it than you absolutely needed? As with everything, there are tradeoffs. In the case of SaaS, the reality is that many vendors design their subscription model to scale in one direction—up. Scaling down? You’re stuck until the contract renews.

If you want bi-directional scalability, then you’re stuck with their most expensive, monthly model. And that’s okay- If monthly scalability is important, it has value, so why not attribute a cost premium to it? Unfortunately, SaaS vendors also tend to limit your options based on consumption as well. Only 50 widgets? Sure, you can pay month-to-month and scale up & down all you want. 100 widgets though? 1000+? At that point, you’re likely out of luck and will be forced into a “premium” or “enterprise” plan.

Yes, you might snag extra features and better support in those higher tiers, but you’ll lose the ability to quickly scale down. Worse, once you’re tied into a long-term contract, they’ve got your cash- And your problems don’t matter until renewal rolls around. This is where a real partnership shows its worth.

Take DataDog- Fantastic product with tons of features, but over several years, our subscription costs had grown beyond what we were comfortable with. Over a seven year period, our environment hadn’t really grown, but our DataDog spend had tripled. Disappointingly, over the course of an entire year asking for help, they just weren’t interested in helping us optimize our usage.

A great tool’s no good if the partner’s a brick wall, in the end, we walked.

Here’s the lesson after too many vendor dances: If the value’s not crystal clear, and they won’t partner with you, don’t just grin and bear it- Do the math, shop around and end the relationship if you have to. Sometimes the only leverage you’ve got is the door. It’s not a fairy-tale fix, but it’s real, and it sends a strong message.

SaaS should bend for you, not break you.