As you might of picked up from my previous post, “What’s the point of PaaS again?“, I’m a fan of the concept of PaaS, but the current offerings don’t really deliver much for me.
However, a tweet today by @reillyusa got me thinking – he asked “How would you sell the notion of PaaS? Elevator pitch please. Bet you can’t do it”, and I’m always up for that kind of bet, even if I totally fail 🙂
So here it is, my draft version 0.1, ready to be reworked based on what people say, elevator pitch for PaaS:
By enforcing the use of specific technologies and shared services by your developers though an open Platform as a Service, you can deliver a higher level of efficiency and lower on-going support costs than may be achieved through the free choice of technology, thanks to the reduction of technology islands, shared knowledge across all developers in the organisation, and sustainable architectural choices, while allowing structured additions to the platform as needs arise.
So, how bad is it?
Ewan – nice effort, my man.
Some thoughts :
1) Most big company CEOs would think “shared services” were Finance, HR, etc…
2) How would you answer when the CEO said “what is your measure of efficiency & how will you know when you’ve been successful?”
3) What will you say when the CEO asks you how many people will we get rid of (RIF) when you “reduce your technology islands?”
4) What will you say when the CEO asks you how much money you will save by lowering on-going support costs?
1) You’re right yeah, shared services would need to be more explicit “shared databases, and other technology infrastructure” or something like that. But then you’re into technology specifics which I wanted to avoid.
2) Suitable measures might be the speed of project delivery, reduction in delivery costs, and the amount of time spent maintaining applications once they’ve gone live. Success would be some level of improvement over the existing measures, though I’m not really sure what could be achieved – obviously organisations like Google feel it delivers results for them, but there’s little in the way of public before and after information, and I’m loathe to hold up Google as an example of anything, as they’re such an outlier.
3) Most technology islands I come across have already been outsourced, so I can imaging it would be operational expense reduction rather than headcount – get rid of that third party supplier who maintains application X on database Y.
4) Enough to roll around in, just give me 10% of it. That’s the standard sales promise isn’t it? 🙂
PaaS is not worthy of an elevator pitch. It is far more respectable and ‘elevated’. 🙂
Can you say that in a way that a CEO can understand it? 🙂 Make. Short. Sentences. Drop. All. Complicated. Words.
Otherwise, it’s an interesting angle at the value proposition. Most others I’ve heard revolve around things like scaling which you are not mentioning at all?
Scaling and stuff like it are interesting and good, but I left it out for 2 reasons:
1 – Most enterprise applications end up being built for a fixed number of users, e.g. it’ll be used by the whole company, either 100 users or 10,000 users, but it won’t have to jump about that much, so you can design for that right at the start.
2 – If you want to build a scalable database cluster, or message queueing system, or whatever else, it’s not actually that hard – certainly not hard enough to justify on it’s own going down the route of a PaaS, which isn’t going to be a trivial course of action. For example, if you want to build high performance MySQL clusters, you can use Chef or Puppet recipes that already exist and are battle hardened.