There is a truth held dear in traditional IT management that with scale come economies of scale.
In the old world of on-premises technology in server rooms that were run exclusively for the company consuming the technology, this might have been the case. Costs nestled in the price of running the infrastructure plus the costs of software licensing for application server software, database software and server operating systems.
For the consuming organisation, these costs were significant, and consolidating (for example) a CRM system into a single product running in one set of infrastructure reduced the high costs of technology infrastructure management. The downsides of everyone having to use the same product (and therefore it being variances of not quite right for everyone) would be significantly offset by the costs saved in server rooms.
Moreover, the software vendors could be negotiated down on per-licence pricing if you were buying in bulk. A traditional software licence has little or no incremental cost for the supplier, so you can negotiate hard with them if you are going to buy 50,000 licences instead of 5,000 or 500. Ultimately the costs of selling 50,000 licences were pretty much the same to the software company as selling 50.
I’m not convinced that any of this now applies in the world of Software as a Service.
For starters, at least on the face of it, in our API world the ability to inter-operate between systems from different vendors should be getting easier. The integration costs that used to be removed by consolidation should be lower these days (I say should advisedly).
But the broader picture changes quite dramatically.
Because in the world of cloud you are no longer managing server infrastructure, much of the traditional cost base moves from big capital investment followed by a lot of tending and watering, into subscription costs.
Subscription costs for cloud vendors have a very different profile in comparison to old software licences. Whereas the incremental costs of additional licences was pretty much zero for licence vendors, in the world of cloud every additional user subscription means additional cost for the Software as a Service vendor. This is important – whilst there’s no doubt that there is opportunity for getting better prices with more users on SaaS, it’s never going to scale as well as it could for licencing models.
Given those two conditions, I’m wondering whether single platforms might be exactly the wrong strategy these days. A number of CRM systems from different vendors, interconnected at a data and reporting level where necessary, would keep the providers on their toes in a way that having one provider simply wouldn’t. The more you consolidate on a single vendor in cloud or traditional hosting, the less likely it is you’ll move products because the costs of change scale up. Swapping out smaller implementations is much more feasible. Having multiple suppliers promotes innovation and cost effectiveness.
The world of Cloud computing isn’t one of merely swapping out one technology layer for another. It forces fundamental rethinking of business models for providers and customers alike. Monolithic, enterprise-wide software might be one of those models for IT management that needs to now change.
mmitII 