Microsoft CRM Online: not so sharp

At least Microsoft has finally admitted that software has to be completely rearchitected before it can be offered as a service. But even if CRM Online, released today, has the right architecture, Microsoft's on-demand business model remains seriously flawed.

Let me start off by offering grudging congratulations to Microsoft, which took the wraps off CRM Online today, for finally acknowledging that it has to totally re-architect its software to be able to offer it on demand. Tim O'Brien, senior director of platform strategy for Microsoft, spelt it out in a media briefing today:

Dynamics CRM is the first product that Microsoft has rearchitected for multitenancy, and to do so, the company had to fully rewrite the program, he said.

"We're not as far down the path of rearchitecting on the Exchange and SharePoint side, but that is directionally the way we're headed," he said. "We have a huge portfolio of applications that we'll over time take in this direction."

Rearchitecting for multitenancy includes the following key actions, he explained:

  • Making sure it can support millions of users rather than stopping at the tens of thousands typically expected at the high end of enterprise software.
  • Partitioning user content so customers can't access each other's data.
  • Supporting high levels of client-side configuration so that individual users can design user interfaces or presentation skins to suit them.
  • Ensuring that one customer running high usage or a bad query can't take down the system for everyone else.

This much is good, and of course Microsoft has the market presence to make some headway with CRM Online. But I can't agree with Josh Greenbaum's assertion last week that the product is going to be a big challenge for Salesforce.com this year [disclosure: Salesforce.com is a client].

For one thing, the assertion that CRM Online is a winner because customers can choose between hosted and on-premise doesn't strike true to me. According to Tim O'Brien, quoted above, the software is optimized for multi-tenancy — so installing it on-premise is presumably sub-optimal. That doesn't sound like a level choice to me.

But what really lets the product down I think is this notion that the integration to Office and other on-premise Microsoft products will make it a winner. This really exposes everything that's wrong with the software-plus-services strategy, in which software delivered as a service remains subservient to old-fashioned on-premise server products. Microsoft's supposedly low-cost pricing for CRM Online doesn't really undercut Salesforce.com when you factor in all the integration servers you have to buy, install and maintain to get your desktop and on-premise apps working with the online services.

I think customers are pretty wise to this, and if they're already committed to the Microsoft infrastructure, then maybe CRM Online won't look too expensive. But Microsoft isn't going to steal customers from Salesforce.com with this strategy. It's a bit like the old razor and blades business model, except that it's the wrong way round. Microsoft is selling the razor blades cheap so that it can make money on the razors. But why would anyone want to buy those razors unless they were locked in to doing so?