Cloud pricing has gotten complicated, and this doesn't serve the needs of most users. What does? Platforms with simple, predictable pricing, and good developer productivity.
Right now, if you're using any of the major cloud providers, you generally need a pricing calculator to attempt estimation of what your costs will be. But good luck with this exercise, because the pricing schemes are in terms of such oddly granular units that you often have little idea what numbers to punch in. ("How the heck am I supposed to know what my 'duration of peak read activity' will be?? I'm trying to get things done here, not predict the weather six weeks from now.")
Making matters worse, the default billing relationship with providers like AWS is unfortunately that you write them a blank check and hope that a bug in your code or DDOS attempt on your service doesn't bankrupt your company or send you groveling for a discount on your bill. To work around this, a recommended practice when using any sort of elastic cloud resources is to set up "billing alerts" to notify you when you're in the process of racking up an unexpected mile high cloud bill. So that'sā¦ something?
The setup doesn't serve the needs of most companies, for a couple reasons:
- First, most companies spend far more on developer compensation than on their cloud bill. Hyper-granular pricing that in theory allows optimizing your costs to shave some percentage points off your bill just isn't worth it for many orgs. It's time-consuming and complicated (often requiring external consultants), and when the cloud bill is dwarfed by other factors, time and energy are better spent elsewhere.
- Second, simplicity and predictability of costs are valuable. When the cloud bill and pricing is too complicated to predict and can fluctuate wildly based on access patterns and usage, an org has no choice but to reserve capital to cover such uncertainty. This is inefficient and makes planning and budgeting more difficult. Who wants that?
On this second point, the cloud providers are in the best position to offer greater predictability, since their costs are aggregated over all their customers and they can select prices that reflect these average costs. So why don't AWS and friends offer more reasonable pricing? They certainly could, but they also have so much market power that they aren't really forced to innovate here. Many customers don't like this status quo, but they grudgingly accept it.
What we're doing
For Unison Cloud, we have simple prices that fit on a notecard. We don't pass the bizarrely complicated pricing structure of infra providers on to our customers, since most companies don't want or need that. They want simple and predictable pricing, and good productivity. For a larger enterprise deal, we're of course happy to negotiate a more granular pricing scheme (and we can suggest some options), as long as you aren't asking us to sell you $1 at a 20% discount. If you do need something custom, please get in touch.
To keep our costs under control, we make use of rate limiting so one user can't monopolize our resources or render an entire service unprofitable. But there are a range of limits which still allow us to operate profitably, and for bigger customers looking to optimize their spending, we're again happy to work out some custom deal.
This makes a lot more sense to us than having a complicated default pricing scheme that only serves the needs of the 1%. Big accounts are likely to want a custom deal anyway for their unique needs, so why not keep the default prices simple and leave the complexity for custom deals? Everybody wins.
Besides keeping pricing simple, we are actually serious about improving developer productivity. On Unison Cloud, there's no packaging or building containers, no boilerplate talking between services, no tedious code getting data stashed in durable storage and read back later, and lots more. We think the cloud should be simple and delightful to use, and we're making it happen.
Sound good? Join us.