There's an interesting discussion going on about multi-tenancy and SAP.
Let me be clear on one thing: SaaS can't be anything else but multi-tenant and opt-out, meaning that there is a single code base for all customers, with regular upgrades for everyone at the same time
But what is only natural for SaaS "pure players" will be impossible to realise for SAP - a small step for one can be a giant leap for others
From the SAP Management Report in the 2010 annual report:
our core business is selling licenses for software solutions and related services to deliver a broad range of choices fitting the varying functional needs of our customers
Well they say it like it is, isn't it? Can't call them shy for sure. Let's take 2010, and see what came from where:
- Software revenue: 3.2 billion euro
- Support revenue: 6.1 billion euro
- Professional services revenue: 2.7 billion euro
That means that half of SAP's revenue comes from support fees. So take that core business mission statement with a grain of salt: not selling licenses, but cashing in on support fees is SAP's core business.
So, where do they get support fees from? Easy: customers pay 22% of their license fees for support. That used to be 17%, but in July 2008 SAP decided to increase that by 30%.
Why not? It wasn't like a crisis just hit us like a truck or anything, was it?
SAP currently is on-premise, and single-tenancy - of course. Customers need a lot of support, maintenance, and help with upgrades. Naturally, customers are forced to pay for support and maintenance, because if they wouldn't, no one would upgrade.
By making customers pay (dearly, I might add) for future upgrades, SAP makes fairly sure that the customers will in fact upgrade as well because they paid for it already.
Surely the actual upgrades themselves will cost even more money but that's just a speck on the map compared to the amounts already put down on the table - and so the lock-in continues, with the customer endebting himself to SAP for future upgrades.
Payback period for SAP products? Not in your wildest dreams! Hotel California man, hotel California...
Then, the future itself knocks on SAP's doors. "Hello, SaaS's here!"
"Damn", Waldorf must be thinking - "there goes the neighbourhood". The joke's on SAP now:
- Customers will not be willing to pay 22% support fees for future upgrades. As everyone shares the same install base, they'll use the argument that they don't want to pay for their competitor's support fee
- Customers will be very inclined to pay much less than 22% support fee, if any
- However, with SaaS upgrades become stealth, and a thing of the past: everyone moves to the next new version at the same time, and doesn't need to do anything for it - ask salesforce.com
That doesn't look much like a win-win situation, does it? And yet SAP has to go SaaS, because everyone else will, so on top of that they even have to invest in developing new existing offerings so they can be offered via the SaaS model, which means throwing even more money to it.
To make matters worse, customers will want their on-premise functionality in the Cloud as well, so SAP will have to offer their existing service offerings off-premise as well as on-premise - more money please!
Last but not least, of course this will cannibalise their current service offerings so that means less revenue - even more more money
So, would I like to be SAP? No, not really. It will be a slow process, of years, but if not killing it will at least inflict quite a few mortal wounds. One remedy (or rather, quick fix) for this? Single-tenancy SaaS
All that's needed is a proper excuse, well thought-out, but I'm sure SAP will manage. Single-tenancy SaaS will keep SAP very, very close to its core business, and not only uphold the support fee model, but even some of its level. I think SAP can get away with single-tenancy SaaS on 10-15% in support fees - it's a bargain compared to the current world!
Of course, that still is only 45-70% of the current fees, and there's still money needed to turn on-premise into off-premise, but if SAP can combine this with private Cloud, they can almost just take the current stack and move it to their own datacentre!
Yes, it will be an outrageous lie and the Clouderati will despise it, but the customers will save 15-35% a year in cost (I just sugguestimated the private Cloud cost to be 15% of support fee) and simply love it
There will be a very slow convergence from on-premise to off-premise, and this is one of its most likely scenarios to get that convergence started. Don't like it? Tough. Disagree? Please say so