Related topics
  • ,
  • ,
  • ,

Blog: Housebuilding as a metaphor for software development

Software is a thing called Comfort

I've long believed that when you start a project you worry about the technology and whether it works. but in the post-mortem afterwards, you find that most of the real issues were to do with project management. Even technology failure can be mitigated, if the project is managed properly. Sometimes the right course is to cancel something based on promises your vendor can't deliver against - before inflicting it on the business.

Of course, you can't always criticise the project management, because project managers sometimes have political influence and power – so it's then safer to blame the technology. Sometimes, even the project review - post-mortem - is discouraged, in case it gets too negative...

Which is why I've been following Mr Screwpole's lessons in project management failure here with some interest and posted a comment here, concerning how well his metaphor worked in the real world. Screwpole sent his lowly amanuensis, Phil Rice, to put me right.

Phil: The last extension I was involved with took two to three builders six months, and cost nearly £50K. The actual day-to-day project management followed the “writing a newspaper” strategy. The project manager (not me…a hired one) spent 20 per cent of his time visiting the site, reviewing progress, scheduling the arrival of raw materials, and allocating jobs of work. There were a few hiccups (that window in the wrong place…that window is the wrong size), but overall the investment in the project management (actually about £7k) was money well spent.

Yes, that makes sense – until the project gets too small to justify a separate project manager and the builders have other jobs on at the same time. I sometimes wonder if proper IT project management is a luxury for big enterprises...

Phil: In my extension, the timescales weren’t really important, the cost and quality of build were! I didn’t need to request specific timescales, as the professional project manager kept me informed. The costs were controlled by having a fixed price contract, but variations were discussed, and usually paid for.

Ah, so it wasn't really a “fixed price” contract (like many IT contracts with big Systems Integrators) in practice and you weren’t living in the house at the time...

Phil: I think it’s worth noting that the world of building contractors is nearly as bad as the world of IT. Every estimate I was given could comfortably be doubled. I’ll be discussing in next month’s email how to deal with this kind of estimation, and how to work out when things are available, and what they will be.

Yes, that email has appeared now (see here). You make some good points - estimating is often thought of as a “black art” but, in fact, it’s a well-understood process – if you know how to do it...

Phil: In the second part of your comment, it’s interesting that you equate comfort with knowing future forward plans. It seems to me that in the world of bespoke construction for private house owners, comfort often equates to the quality of the deliverable, and timescales are secondary.

For rich people, perhaps. For people trying to juggle a job and home improvement projects in a house that's barely big enough the first place, I think timescales (and knowing when the disruption will finish) are very important to comfort. Which, in the IT world, is perhaps why “timeboxing” is such a feature of the DSDM agile method – functionality can be negotiated but delivery deadlines don't slip.

Phil: In a prefabricated world (such as housing estates built to prescription by property developers) comfort often equates to accurate costing. The constructors will be working to low margins, often well into single figures. This world needs tight project management, mapping of dependencies and optimisation around the usage of resources.

Ah, in that world, comfort for the developers may not equate to comfort for the eventual occupiers. Project managers need to know exactly who they have to keep happy - and, for some IT projects in somewhat dysfunctional companies, that isn't the end-user...

Phil: I ’m tempted to pontificate on how to provide customers' comfort, but I suspect there’s a categorisation model different to that on project types. I’d be interested in other people’s opinions on this…

Perhaps our readers will comment here...

Sponsored: Driving business with continuous operational intelligence