Updated 20-Aug-2009, 6:50pm.  Corrected the cost of the F1 and provided some numbers for comparison.

Is Total Specialisation the Future?

In case you hadn’t noticed, there’s a trend occurring which affects you and your future, namely the increasing degree of specialisation in the IT industry (and world in general). 

It used to be that you’d get asked what you wanted to be when you grew up: a lawyer, teacher or fireman; although that question is still relevant today, the number of career specialisations available (and the knowledge they require) is placing increasing pressure on us to make good choices earlier that have greater impact.

There are a number of factors and influencers behind this, and I think there is also some basic laws that we can identify and which may help us (and our children) make good decisions: whether it be career choice, architectural decisions or business strategy.

I was reading a piece called “Protect and Preserve: Mobile Art-Conservation Van Helps Save Treasures”; It’s basically about a small bunch of specialists (with specialist gear, no doubt worth a specialist cost) who go around helping to preserve works of art.  It seems that (like most things) art preservation is increasingly complex and as that complexity rises so do the costs; and there comes a point when it’s no longer via to do it in-house.

See a pattern here?  Out-sourcing, software services (cloud or local) – it’s all the same thing; eventually you get to a point where the result you need is beyond your capacity (or the amount of time and money you’re willing to throw at it).

Let’s put the idea into a picture, and we’ll use the art preservation context.  In the beginning (the left hand side) an art gallery could perform a certain amount of preservation itself at relatively little expense – and gets good value for money.

As the time went by preservation became more demanding, requiring more resources (skills and technology) until eventually, at a certain point in time the costs and benefits were balanced, and then as things continued it wasn’t cost effective or possible to sustain.

Of course it is possible for the trends to go the other way.  There was a time when developing websites was extremely specialist – not many people knew how and the necessary tools were not as available or capable as they are today, and so on.  But, with the passing of time, building websites has become much easier; there are far more resources available to help and the bar for entry is lower; on the other hand the limits of what is possible has dramatically increased.

The idea of specialisation is closely linked with the concept of continuums: where specialisation occurs over time, gradually pushing out the extreme of what is possible and what is expected.

So, what are the laws of specialisation and what factors do we need to keep in mind?

The Laws of Specialisation

1. The amount of effort required to achieve a degree of specialisation increases every time specialisation occurs.
2. As specialisation increases so do its dependencies – and therefore its exposure to risk.
3. The value gained is proportionate to when you get there, and when your competitors do.
4. The more that specialisation occurs, the less specialised previous specialisations are (relatively).

So what drives specialisation?  The underlying driver is need: what is the result you require and what is it going to take to get it?  This boils down to improvement – continuous improvement:

Factors / Drivers:

• Effort
• Cost (investment in tools, time)
• Skill
• Size and complexity of the problem domain

Limits:

• Resources
• Us, how much a person can deal with / understand.

Driven / Influenced By:

• Competition (personal / national security, business): bigger, faster, better
• As available resources dwindle the need to specialise increases - in order to increase efficiencies or extend the scope of available resources.
• In areas of competition, the more specialised things become smaller improvements and refinements become more important.
• Evolution constantly raises the bar.
• As more is understood (within a problem space / domain) people expectations rise.

Benefits of Specialisation:

• Access to resources or capabilities that others don’t have.
• Greater efficiency.
• Better results for the same cost, or
• Same result for lower cost, and eventually
• Better results for lower cost.

Disadvantages of Specialisation:

• Become more dependants on a limited set of skills.
• The context of use / range of possible uses decrease.

The Laws in Detail

The amount of effort required to achieve a degree of specialisation increases every time specialisation occurs.

Anyone who has worked in software development for a reasonable length of time will be aware of (even if not by name) the 80/20 rule or Pareto Principle; when 80 percent of functionality is achieved by 20 percent of the work or system.  This translates into the amount of effort put in, and as any seasoned software developer knows, its finishing off the last features and bugs which usually take up the most time – haggling over details with the client; in other words the details which are increasingly specialised.

Another example that illustrates this specialisation principle is advancements in the automotive field.  The lineage of the modern automobile is long, but the Model T Ford is a good place for us to start:

“As the auto industry grew and competition got stiffer, everyone in the company – from Ford’s employees to his family – pushed him to update the design”

http://stevelutz.wordpress.com/2008/04/23/henry-ford-the-cost-of-not-innovating/

Let’s jump forward to the Ford Escort.

Obviously evolution has pushed the capabilities of the car a long way, and we can see that essentially the same car can be used for different purposes.  I’m sure an off-the-shelf Escort is capable of racing around a sealed racing track or a dirt road, but if we invest a bit of effort into improving the car for a specific use (i.e: we specialise it) we can get much better results; and most importantly, the effort required to do this is arguably quite reasonable.

How much effort do we need to go to improve yet again?

The modern Formula One car is a very specialised beast indeed, and it’ll be no surprise to anyone that achieving that level of ability requires some pretty focused specialisation and a lot of effort (which we can roughly translate into cash).  If I had a lot of spare time I could probably provide a detailed break down of costs (effort) against metrics for performance (specialisation) and show quite clearly how the degree of effort increases, but unfortuately I don't.  I can provide some data to consider though:

In terms of cost:

• The standard 4-seat open tourer of 1909 cost $850 (equivalent to $20,091 today).
• A Formula One car is in the region of $2.4 Million (and Not $6.8 as I mistaken had earlier. Doh.)

Now I’m not sure if that’s a high-end car or not, and I don’t know what vintage, but it’s certainly more than I can afford on a car.

In terms of engine size and performance:

• A modern Formual One engine typically produce 224 kilowatts (300 Break Horse Power (bhp)) per litre of displacement, far higher than most internal combustion engines.
• For comparison, the naturally-aspirated piston engine production car with the most specific power is the Honda S2000 engine with 92 kW (123 bhp) per litre.
• The Model T Ford had a 2.9 Litre engine which produced 15 kilowatts (20.2 hp).

This gives us some rough figures for direct comparison:

In other words, spending 24 times as much will get you 6 times more Kw, but spending as much again won't get you twice as many Kw. Yes I know this is all very approximate, hardly scientific, and not really a truly accurate set of figures - but I think its not so far off as to be completely meaningless.

It's also worth noting that not only is the F1 car very specialised, but so are the rules with which its expected to comply.

As specialisation increases so do its dependencies – and therefore its exposure to risk.

Perhaps the most classic example of specialisation is bio-diversity; this is basically specialisation as evolution.  New Zealand isn’t the most bio-diverse land mass on the planet, but it does have some examples that amply demonstrate how vulnerability and specialisation are linked, such as the Kakapo.

The Kakapo is a native to New Zealand, is flightless and nocturnal, it has specific dietary needs, it’s also a heavy bird (not very agile) and not capable of defending itself from predators.  The Kakapo isn’t capable of doing anything that other birds can’t, so in its case it specialisations are more like limitations, and yet for a considerable period of time it existed in plentiful numbers and filled its place in the environment.

The vulnerability of its specialisations was fatally exploited by introduced predators, against which the birds had no defence, as well as destruction of its natural habitat (a common issue for many species).  The Kakapo is now an endangered species and its extinction has been avoided only by coordinated and sustained human intervention.

The value gained is proportionate to how when you get there, and when your competitors do.

The amount of value gained from a degree of specialisation is entirely subjective, but it’s safe to assume that if your there first you’ll be a winner.  But then, assumption is maternal; being first doesn’t always guarantee success – execution is important.  The use of the atomic bomb in World War Two marks an achievement in specialisation, namely in aeroplane deliverable destructive devices.  The degree of effectiveness – coupled with the Allies ability to use it first (by a wide margin) brought significant success.

Use of the first Tiger Tanks wasn’t so successful. Although the Tank itself was nothing new, the capabilities of the Tiger were expected to produce favourable results; to quote Wikipedia:

“Its predecessors balanced mobility, protection, and firepower, and were sometimes out gunned by their opponents.  The Tiger I represented a new approach that emphasised firepower and armour at the expense of mobility."

Unfortunately the expectations which commonly accompany the introduction of new technology were to be frustrated:

“His staff remonstrated that the terrain he had chosen made tactical deployment of the tanks impossible because of the marshy subsurface on both sides of the road… The Russians calmly let the tanks roll past an antitank gun position, then fired direct hits at the first and last Tiger.  The remaining four thereupon could move neither forward nor backward, nor could they take evasive action to the side because of the swamps, and soon they were finished off.”

Albert Speer, “Inside the Third Reich”.

The particular context here is the tank arms race against the backdrop of modern mechanised warfare; the main objects of the Tiger were to mount a main armament of a certain size, and protection sufficient to protect against a similar weapon.  The outcome was a specialised vehicle of 75 tonnes; and although it could effectively engage targets at ranges of 2000 meters plus, it was limited in where it could be used and the full advantage of its specialisation was only realised in certain situations.  All in all it’s a classic example of the laws and forces of specialisation at work.

The more that specialisation occurs, the less specialised previous specialisations are (relatively).

This is typical in the IT field; as mentioned earlier, building websites is no longer seen as being as specialise as it once was – demand and supply, coupled with a competitive environment have driven evolution (and specialisation) in IT, and the availability of matching skills has followed.

Conclusions

So what does all this mean?  What’s the impact on us?

It’s fairly clear that you need to be aware of the laws of specialisation and that you’ll get benefits from applying them correctly.

If your architecture, design needs to meet a specific goal that is out of the ordinary you’ll likely need to specialise – beware how this potentially exposes you in terms of required skills and resources.  If your area of interest is fiercely competitive you’ll already be aware of the need to improve your performance – at this point you’ll need to look at areas where you can safely specialise without putting your venture at risk, or more accurately model those risks.  Finally, if expectations on your performance are rising – as well as the costs of meeting those expectations – perhaps it’s a good time to look further into the future and bring in specialists early; perhaps there aren’t any to use – in which case you might have stumbled on a whole new area of specialisation, and it’s yours for the taking (for now).