Last Modified:                                                                                                



Homepage / Publications & Opinion / Silicon.com

Peter Cochrane's Uncommon Sense: Cochrane's Law
Too many cooks... do what exactly?
Three people rarely do three times the work of one person doing the same job. Will the wired economy change that? Peter Cochrane has some anecdotes and formulae to share...

In those labour intensive days before the PC and office automation, I had the experience of working with a number of secretaries and assistants in large offices. On each occasion when I started a new operation it would involve me, a secretary and perhaps an assistant. We would build the business, gain orders, recruit people and create an organisation to satisfy a given demand.

At some point I would have to engage additional office staff and then I would begin to notice a very definite and clear effect in terms of the real work output. This I called 'Cochrane's Law of Secretaries', and it went roughly like this:

1 secretary = 1 secretary's output

2 secretaries = 1.8 secretaries' output

3 secretaries = 2.6 secretaries' output

4 secretaries = 3.3 secretaries' output

And so on!

It always seemed to be the case that putting extra people into the same work area merely leveled off the work output to some asymptote with the total number employed. This is sometimes referred to as the 'painters in a room' problem. You can employ more painters to paint a room and get the job done faster but eventually there are so many painters they can't all get to the wall and start to apply paint to each other.

Most definitely 10 secretaries don't do 10 secretaries' work. It seems a basic human need and desire to talk, communicate and interact and this becomes amplified by the added complexity of an organisation growing.

So in an office environment we quickly see an increasing percentage of time dissipated or distracted by other activities best defined as socialising work and processes. As work becomes more complex, more networked and faster, the communication overhead required in a strictly human-to-human network grows rapidly and takes individuals energy from the primary purpose.

I am not picking on secretaries in particular you understand. This seems to be a general condition that has parallels with the behaviour of other large groups. Over the years I have noticed this general condition (or 'law') to be true of administrations in general or indeed almost any group employed in a highly interactive environment. More, ultimately, becomes less for the old modes of working.

On another plane I have also observed a more damaging aspect. If 10 people start a new company with a single administrator, it is clear that admin is there to support the nine in the creation of the company, finding customers, realising orders and achieving profitability. This set up applies for some time as the company grows to 100, 500, 1000 and so on.

Only at some point the value and belief system get inverted and the administration becomes the means to an end. Divorced from customers and markets, finance and personnel become the core of the company and the rest are there to support them. This condition is not only wrong, it is dangerous and often fatal, with death engendered by growing lack of customer awareness and increasing insulation from reality.

I'm not sure there is a 100 per cent modern equivalent to all this. We now see companies and virtualised, distributed and networked, with huge numbers of people chit-chatting on email, instant messaging, sending text messages and talking on the telephone as opposed to involved in discussions across the desk or via pieces of paper.

The best managers no longer broadcast and dictate but communicate, create and support. Everyone can see the company condition via electronic access to constantly updated briefings and databases.

Another general observation on the ebusiness world: our work output tends to increase about 10-fold every 10 years but without a complete dissipation of individuals' energy in socialising work.

However, it is clear that similar laws apply to the overall intelligence of organisations in general. I tend to think of this in terms of corporate IQ. In short the IQ or smartness of an organisation does not appear to go hand in hand with the number of people employed. Often it is quite the reverse. It is not unusual to find very smart people employed in apparently very dumb organisations or alternatively very smart people creating very dumb organisations.

In engineering and science there is a very well known and common phenomenon often referred to as coherence. In brief, if two coherent signals or forces are added together they do so arithmetically, so: 3 + 4 = 7. But when they are incoherent they tend to add on a power basis giving SQRT(3^2 + 4^2 ) = SQRT(9 + 16) = SQRT(25) = 5.

This also seems to be a fair approximation of what happens with our mental capability. It would appear the best you can hope for is that two people are jointly SQRT(1^2 + 1^2) = SQRT(2) = 1.4x smarter. If the number of people increases, then the effective IQ seems to rapidly level off well before the total equals 100 people. The exception - if there really is one - is when we restrict team sizes and network all the participants in a company or on a project.

The next big step is going to be the inclusion of smart machines, and the big question is whether they will follow the same limiting laws? I suspect not.

This column was typed on my laptop on BA496 flying Lisbon to London and emailed to silicon.com from The Seal Pub in Woodbridge UK using my GSM phone.