The shovel or the hole?

This is a lesson that took me long and painful time to learn.

It also explains a lot why superior technologies tend to fall in favour of their lessers, whether it be Beta vs VHS, Ethernet vs Token-Ring, ia32 vs PowerPC, Windows vs (just about any other OS) or HD-DVD vs Blu-ray.

It's all about the shovel and the hole.

Say you want a hole dug in your backyard and you want to contract someone else to dig the hole for various reasons. You either don't have the time, don't like manual labour or you just figure someone else can dig a bigger hole than you. In any case, you narrow the list down to two contractors.

The first says he can dig the hole to a certain depth, with the dimensions you require on time and on budget. No problem, this guys clearly in the running.

The second contractor starts talking about the shovel he's going to use. It's a superior shovel - made from a titanium alloy. Lightweight, faster, better than the competition, it can shift more dirt per shovel than any other and more digs per hour. In the hands of a skilled operator, much more efficient than the competitors shovel.

What's your reaction? What would you say?

You'd most likely say you don't care about the shovel. You only care about the hole.

Your concern is the actual deliverables. If his shovel is really that good, then that should show up in the bid for the project. Oh, but he's a little more expensive, well, because his great shovel costs a lot more, and that's reflected in the overall price.

No points for guessing who wins at this point.

You don't have to read far into The Road Ahead to realise that Bill Gates understood this lesson very early on in the game whilst those around him had no idea.

Before the IBM PC was released, Digital Research's CP/M (and also MP/M) was the dominant Operating System for 8-bit Z80/8080 computers. It was fairly ubiquitous and for many it was an aspiration to own a CP/M computer. Microsoft (then a fairly small software company with only two products) was commissioned by IBM (along with Digital Research and UCSD) to write an OS for the new IBM PC. IBM would sell the new IBM PC and allow people to purchase whichever OS they wanted. The three products were PC-DOS, CP/M 86 and the UCSD P-system. Both Digital Research and UCSD structured per-unit licencing for their OS. Microsoft, however, pitched a flat-fee to IBM at a low price. IBM could then charge whatever they wanted and keep every cent. This incentivised IBM to sell MS-DOS at whatever price they could. MS-DOS (marketed by IBM as PC-DOS) sold for $40, CP/M-86 sold for $240 and the P-System cost a whopping $499. As a result, 96% of IBM PCs shipped with MS-DOS. It didn't take long for MS-DOS to be the only OS available for the IBM PC.

Microsoft didn't make very much from the initial deal with IBM, but once the clones starting appearing, 'MS-DOS' had to be purchased from Microsoft - unless the manufacturers also cut a deal with Microsoft.

The rest - as they say - is history. The clearly superior (for the time) CP/M-86 was left to be lamented by its cadre of faithful users.

Another example I can think of is a microwave I saw once for sale that proudly advertised "With inverter technology!" Now I doubt more than 1% of people would know what an inverter was, let alone pay an extra $50 for it. If instead they had said "Heats food more evenly", that would have been different.

IBM used to have a TV advertisement that illustrated this effectively. An IT Manager is called into his boss' office and told "I need you explain our web strategy to the board." He quickly replies "Sure, no problem." To which his boss adds "In terms they can understand." A look of horror crosses the IT Managers face for a few seconds before he blurts out "Every buck we spend this year will make us two bucks next year!"

I experienced something similar to this when I worked for AV Jennings. I had given what I thought was a short presentation in lay terms to the IT Steering Committee explaining our DR plan. After I finished I asked if there were any questions; there were none so I left the room and that was it. A few months later my boss came out of a Steering Committee meeting and said "I found out how to use you in those meetings". In response to my quizzical reply he said "I went into the meeting prepared to answer questions on eight topics, but they peppered me with questions about a topic I wasn't prepared for. They started grilling me on why I didn't know the answers and I said 'Look, I'm just not across all the technical details, if you want I can get Wayne in to answer them.' They stopped me and said in unison 'No! No! No! That's fine! Don't bother.'"

As an IT professional, you may have your favourite OS, hardware platform, dev tool, app environment, programming language - whatever. You may have excellent reasons why it is the ants pants. However, your client doesn't care. Chances are they are so far abstracted from your reality it is difficult for you to be objective about it. However it's crucial that you try to view the world from the user perspective. What you deal with on a daily basis may as well be a black box that runs on magic to them. Importantly, perspective is the key.

Novell had an email suite called Groupwise that was vastly superior to Microsoft Exchange (arguably it still is). Groupwise 7 had five separate btree databases including a guardian database to check and rebuild any of the other databases if necessary. The message database had 128 separate data stores. It supported live migration of users between post offices, live in place upgrades and dynamic load balancing. This at a time when Exchange 2003 stored everything in one large, slow ISAM database. The TCO of Groupwise was one tenth of Exchange 2003.

Unfortunately, the Groupwise client application sucked. The client also would only work with Groupwise.

In contrast, Outlook could talk to pretty much anything (including Groupwise). The end user didn't know anything about btree databases, dynamic load balancing, separate data-stores or the fact the sysadmin could upgrade the email system in 40 minutes with zero downtime instead of it taking three weekends to do so. That, after all, was the sysadmins JOB.

But they did that they could get nicely formatted email messages with graphics in-line: something the Groupwise client couldn't do. And for those that deployed the Outlook client with Groupwise, there where lots of nice 'features' of the client that would only work with Exchange. Couple that with the fact that when your boss' boss plays golf with a Microsoft exec, then your opinion as an engineer is not going to count for very much.

The whole Groupwise/Exchange war could easily have been won by Novell if they'd had spent a bit more time and effort on the client application.

And what applies to large companies applies equally well to you as an IT Professional. Take a close look at the deliverables your client needs. Think of the user experience and what could enhance it. Anything that that you regard as important but the client either doesn't or wouldn't understand, list these as a secondary benefits under one grouping or make it an optional extra if you can. The client has what they what in mind and you can bet lowest price is high on that list.

Remember: It's all about the hole - not the shovel.