At the risk of hijacking a perfectly good religious-war label, this morning finds me inclined to attack established practice in secular areas as well.
My "Danger, Will Robinson!" alarm went off when I read this sprightly citation of comments by Lyn Robison (yeah, I noticed, "Robison" does look a lot like "Robinson") of the Burton Group:
Mr. Robison also said some businesses just aren’t suited to SaaS. The accounting sector, for example, has well-established business practices, so there’s enough commercial off-the-shelf software to meet any accounting office’s needs and budget. Meanwhile, certain companies’ business processes are what really differentiate them in the market, and so they need to buy or create highly customized applications. Wal-Mart’s supply chain, for example, ensures low prices on a wide variety of merchandise, so it needs special supply chain management software.
Then there are the problems associated with Internet access—or rather, the lack of it. A company has little influence on when an SaaS vendor takes its system offline for maintenance, while disrupted Internet service could prevent a business from getting to its data. “These things could happen at inconvenient times for the enterprise,” Mr. Robison said.
Your honor, I object. On multiple grounds.
What bothers me are the multiple assumptions embedded in these remarks: assumptions that may impede any number of well-considered recommendations to adopt an On Demand model for an application, whether using the salesforce.com platform or another.
- "…well-established business practices…" — that doesn’t make them best practices
- "…enough commercial software to meet any…needs and budget" — also known as "enough shelfware out there to sink a ship"
- "…business processes are what really differentiate…" — so why waste 80 to 98 per cent (if you don’t believe me, ask Gartner) of the IT budget on non-differentiating infrastructure?
- "…need to buy or create highly customized applications…" — which in no way precludes using On Demand platforms and technologies
- "…little influence on…offline maintenance…" — which is the bigger problem, a scheduled offline maintenance window (undoubtedly planned for an off-peak time) or an unscheduled problem with an on-premise stack? How many on-premise installations are backed by the kind of multi-site redundancy and failover provisions that are offered, for example, by the salesforce.com network?
- "…disrupted Internet service could prevent a business from getting to its data" — sorry, but I consider that a 1990s argument:
- Using an On Demand model does not mean giving up on-premise data access
- If your Internet access is down, you’re not worried nearly as much about access to your past as about the destruction of your future
I believe I’ve made my point. Take the time to deconstruct the blithe generalizations that sometimes pass for technical discussions of application development options.