This is what we have come to understand lately: pay-as-you-go scares customers away.
The assumption behind a pay-as-you-go, or metered pricing plan is that we will think about it as fair. The more we consume, the more we pay. Simple. After all, there are a lot of things, typically utilities, that are metered: electrical power, water, heating gas, phone calls and even Internet usage in some cases. Every month we’re charged a different amount but we regard that as perfectly normal.
The problem with Internet-based services is that they’re never enough. We want even more bandwidth, even more minutes of usage, even more gigabytes of storage. In a word, everything must be unmetered. Or, much better, free.
An then there’s cloud computing, like that offered by Microsoft, Amazon and Google. Companies purchase computing power in 1-hour blocks, storage in GB/month, and so on, to offer a service to end users. It’s a typical metered model, designed to allow flexible and intelligent resource usage. This allows companies to build services on top of cloud computing platforms, and to dynamically scale resources depending on the actual number of users they have at any given moment. Large companies take on the burden of purchasing thousands or millions of servers, acting like buffers, while allowing smaller ones to purchase computing powers at will, without any kind of commitment (except some technological lock-in – but that’s for another post).
So, it seems perfectly fine to apply a similar pricing model to end users, but it’s not, and here is why.
For companies using cloud computing platforms, like ours, estimating their computing power needs is relatively easy. It’s a matter of doing some tests and then some very basic calculations. But when we offer a pay-as-you-go service, not only we have to convince the potential customer that we’re useful to them, but most importantly that they’ll pay a reasonable amount (whose perception varies from one another) for the service. If that amount is unknown or variable, then we’ve lost already.
The conclusion is: a periodic subscription including a given amount of resources is much better, for two main reasons:
- it’s easier to understand because pretty much everything works like that on the Internet
- knowing the amount to pay in advance gives the customer a sense of control which is completely absent in a metered model.
Funnily enough, even if the potential customer has to pay a bigger amount simply because they’re not using all the resource they’re given, knowing it in advance will make a huge difference. I have to admit that I was warned about that, but I grossly underestimated its effects (perhaps because I’m an engineer and calculating things is normal to me – normal people not only look at the bottom line, but they also demand to know it in advance).
There also is another aspect to consider. Metered billing ignites a weird behavior in users: they’ll try saving every single penny they can. It’s mostly irrational because the time they waste doing so is much more valuable than the actual money they’re saving. I believe it’s something that we all have deep inside us. An instinct, if you like.
BTW, I’m not sure which definition feels scariest, metered or pay-as-you-go. I’ll leave you to figure that out.