% fortune -ae paul murphy


One of the things people like Ben Rockwood complain about a lot is the complexity and general pain of navigating Sun's website - and, in particular, the paid access portions of the documentation and support sites.

Here's one last quotation from the comments, this one by Fazal Majid, to Jonathan's blog on branding:

Well, Sun's brand promise is very similar to what DEC's used to be - excellent software and hardware, but completely mismanaged company. I am trying to buy an Ultra 40M2 for home (I'd like to use the 8-drive capacity to set up a ZFS RAID server for my home storage needs). For my company orders, I just don't want to deal with Sun and hand over my business to a VAR who knows how to navigate through Sun's abysmally dysfunctional business processes, but in this case it's not really an option.

First of all, your site wouldn't even allow me to log on yesterday all day. How can you maintain any credibility when you talk about RAS, scalability and availability when you can't even keep your own storefront open on a Monday? The NOT in not-com indeed.

Now, after managing to add the item to my cart (after ignoring the multiple SSL/non-SSL mismatch warnings, how amateurish can you get?), I try to click on the checkout button. I get an empty page. Well, not quite empty - it does have the Sun banner, and the feedback footer, but nothing in between. How am I supposed to make my purchase?

This is a recurring pattern with Sun, and it's amazing your company is still alive despite such epic levels of incompetence at so many levels.

Since quite a lot of mid range Sun people suffered their formative professional years at DEC this is less than a co-incidence, but there's a lot more to the story.

Take a good look at the complaints, by Majid and many others, about Sun's web sites and you should see three recurring themes:

  1. the sites are slow;

  2. the sites are difficult to navigate; and,

  3. the sites often seem not to work well.

These results ultimately reflect organisational design issues - most importantly the fact that these sites are still fundamentally run as cost centers.

There are two main reasons costs center operations generally become divorced from their markets:

  1. the budget process desynchronises the operation from external change; and,

  2. the budget management process permits short term tactical change while preventing long term adaptive change via capital investment or strategic personnel change.

As one result cost center operations tend to go from bad to worse until senior management takes dramatic action to restart the cycle - often, at least in IT, by alternating out-sourcing and in-sourcing.

More locally, one result is that cost center management often makes tactical changes for no better reason than that doing so is within their span of control - and then chooses those changes which build their own, or staff, resumes. You often see, for example, such groups adopting whizzy new tools without making the training and server upgrades needed to really use them - and without considering their impact on service users.

In Sun's case, for example, there's no structural impediment to the adoption of advanced page features - meaning that people who try to access the sites, as I suspect Mr. Majid did, with a browser never intended to cope with 1,800 lines of include files for every HTML page may find their access silently denied by their own technology.

So what can the President of Sun do? Not as much as you'd think, but I do have some suggestions:

  1. institute a strategic review aimed at determining which elements of Sun's web services efforts can be turned into profit centers and, for each, how that can be done.

    For example, turning the web store into a direct profit center sounds easy, but requires that the store's managers be given more autonomy and be rewarded on what they sell - thus bringing the entire sales and customer engineering compensation issue to the fore.

    That's not "low hanging fruit" - but spinning bigadmin back out as a separate business is: there's a dedicated core group, significant applicable expertise - and an obvious way for Sun to support them without limiting them: a long term advertising contract for their site.

  2. institute a review of the entire documentation operation to see how it can be turned into an asset for Sun marketing. It might be possible, for example, to conceptualise Sunsolve and its relatives as a public benchmark site for Sun - providing anonymous customer access to the shoemaker's latest shoes - and, in that process, set up some sensible rules for both customer access and software use. For example:

  3. talk to the people currently responsible for these sites to determine first whether change is needed, and, if so, what changes make sense - it may be, for example, that personnel change is indicated, that they need something they don't have, or that an experienced web editor needs to be brought in from outside to reconcile the drive to standardisation with the differing needs of each site.

The big challenge, of course, is to turn the documentation and support sites into marketing assets instead of goodwill sinks and ultimately that's going to come down to control structure and compensation - but there are things that can be done quickly, so let me join Majid et al in urging Sun to get something positive done - soonest, please.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.