--- title: "Why do we fetishize \"best practices\" but not \"best infrastructures\"?" date: 2006-09-25T21:27:31Z modified: 2006-09-25T21:27:31Z permalink: "https://redmonk.com/jgovernor/why-do-we-fetishize-best-practices-but-not-best-infrastructures/" type: post status: publish excerpt: "" wpid: 728 categories: - Uncategorized timestamp: 2006-09-25T21:27:31Z tags: - Uncategorized --- [Reinventing Local Government ｫ Manage By Walking Around](http://alignment.wordpress.com/2006/09/23/reinventing-local-government/ "Manage By Walking Around") [Jonathan Becher](http://alignment.wordpress.com/about/), CEO at Pilot Software, is a contrarian thinker in corporate performance management, and he asks a great question in his latest blog entry, jumping off a speech by Bob Hanson, CIO of Sarasota County Florida. > “Sarasota County spends more than $100K every year to operate their email system. > > Since Sarasota County is probably larger than most, assume the average yearly cost is $25K > > Therefore, something like $2.5B is spent every year on municipal email systems. > > What if we replaced the 10K individual systems with one shared email infrastructure modeled on Yahoo! Mail or Microsoft Hotmail? It’s certainly technically feasible and we’d end up with a system that would cost a fraction of the $2.5B to operate but that would have greater redundancy, reliability, and security than most of the organizations enjoy today.” Becher talks to the goodness of best practices but it seems to me the language of best practices is often utterly bankrupt. A best practice, to my mind, must be based on direct practitioner experience. Unfortunately in our field it can mean almost anything. In the web services community, for example, “best practice” is often used as shorthand for implementation of a standard. The disease of calling arbitrary processes or implementations best practices holds the industry back, again and again. In this [age of shared services](http://www.redmonk.com/jgovernor/archives/001706.html), shouldnt we be looking for “best infrastructure” though, rather than best practices? After all, the best infrastructure has already proven itself in production, while best practices on the other hand have often not proven themselves anywhere other than on a technical document. I don’t like to fetishize the services provided by web companies either- but hotmail et al obviously do scale for the task in hand. What companies would benefit most from “best infrastructure” thinking? Best in class service providers. EBay, Amazon, Google, Fedex and so on. And Sun – CEO Jonathan Schwartz is puttiing [most of his eggs into the shared service provider basket](http://blogs.sun.com/jonathan/entry/explaining_our_share_gains). Jonathan wants to focus on customers that potentially provide “best infrastructures” or “best services” , and is prepared to focus less on those that are obsessed with building to “industry best practices”, long after the market has established them. Sun is prepared to bet on best infrastructure. Are you? Who would suffer in a best infrastructure world? Likely IBM for one. IBM Global Services and Business Consulting make tens of billions of dollars selling best practices. Even SAP, which sells best practices instantiated in code, is still selling best practices, rather than the best solution to the job. IBM needs to become a switch in the sky if it is to play more effectively in a best infrastructure world. Currently its customers have too many opportunities to poorly implement its software (the same is obviously true of Microsoft .NET) I am talking about the difference between specification and implementation. Of course political issues make the adoption of best infrastructure difficult, in any sector, from local public sector to the biggest enterprise. Over the last thirty years organisations have concentrated efforts on managing their own assets, and “IT competitive advantage”. But what has doing so bought? In many cases large management overheads. But everyone fights for budget, whether they need it or not. Ever spent money in order to get a budget for next year, rather than because you need to spend the money for new infrastructure now? Exactly. If we buy into ITIL say, a best practice framework for IT Service Management, that means a significant investment in training and cultural change. We have to a create a culture of excellence, and best practice. In some cases it would surely make sense to just subscribe to existing successful management services rather than building our own. That’s one of the problems with trying to sell ITIL to SMBs. They might be better off with a simple managed service. Best infrastructure can also be fetishised in its own right. The UK government is currently spending upwards of ten billion dollars on a project that takes purchasing power out of the hands of local medical practices, and hands it too five unelected “regional service providers.” Of course, none of these service providers has ever actually provided the service in question, before so to my mind they are neither best infrastructure nor best practice. Needless to say the players are missing deadlines and pointing fingers. As a taxpayer I would like to see the government recommend choices from service providers that can show a working system for the service they plan to provide. I am not saying the organisation has to already be in that exact sector, but at least be solving the problem in question. Thus for example, local government organisations should be encouraged to try salesforce.com – because it offers a working system, rather than some best practice ideas and piece parts to implement them. I haven’t used the SOA word yet in this piece, but that is a likely context for a discussion of best infrastructure going forward. Of course the world isn’t going to move wholesale to a try before you buy, acquire service rather than products world, whatever salesforce.com would like. But when you see CIOs from public sector organisations start questioning whether they have too much budget and too much freedom to implement stuff, you have to wonder whether the shared service bug isn’t catching. So lets fast-forward to a world of bureau-based computing… everything old is new again. disclaimers: Microsoft and IBM are clients. Pilot and salesforce.com aren’t. tags: [ITSM](http://technorati.com/tag/ITSM), [ITIL](http://technorati.com/tag/ITIL), [Best Practices](http://technorati.com/tag/Best+Practices), [SOA](http://technorati.com/tag/SOA), [Sun](http://technorati.com/tag/Sun), [IBM](http://technorati.com/tag/IBM), [Microsoft](http://technorati.com/tag/Microsoft), [IGS](http://technorati.com/tag/IGS), [Pilot Software](http://technorati.com/tag/Pilot+software), [management by walking around](http://technorati.com/tag/management+by+walking+around), [Sarasota County](http://technorati.com/tag/Sarasota+county), [NHS](http://technorati.com/tag/NHS), [salesforce.com](http://technorati.com/tag/salesforce.com)