It’s efficiency but not as we know it because we always get out of a system less than we put. Or do we?
In efficiency terms most things are inefficient and some things are less efficient than others.
Take electricity production using coal:
Now I don't have anything against coal. I once spent a very happy year delivering it and when asked what I did would reply that I "Worked in the weights and measures department of precious minerals".
Coal
But this much coal
Produces this much electricity
Electricity
that when used by an electrical appliance
does this much useful work
Work
How did we get from a huge pile of coal to a small amount of work?
The law of diminishing returns could have something to do with it but it's really that it takes a huge amount of energy to boil the water that turns the turbines, that generate the electricity.
Each link in the generation train takes energy out of the system mostly in the form of heat
Which is why there can never be a device that creates as much energy as it consumes, running for ever on it’s initial charge.
Efficiency returns are relative though. Coal to electricity to work done is a good example of an inefficient system but one that’s easy to understand.
So how do software and business systems stack up?
Well, it’s pretty well understood that without computers we would need many more people working in business, government and communications.
But how do we measure efficiency in a computer system?
I mean, it could be just a displacement activity. If there were no computing industry those that work in it would have to be employed elsewhere, typing memos, scratching out accounts, making tea or any of the hundreds of other jobs that having computers has empowered us to do.
In a way the computer age has made typists of us all. Instead of hundreds of people in typing pools tapping away, we now do it ourselves.
Computing cannot be the infinity engine we looked at earlier, can it?
Does computing break the second law of thermodynamics and become a law of increasing returns!
Let’s try to quantify the energy cost of a typical business computer system.
Let’s be honest. Coal is a whole lot simpler, we don’t have to worry about the effort it took to get out of the ground. As an entity that contains energy in proportion to it’s weight the sums are easy.
1 Ton of coal contains 21 to 22 gigajoules of energy. Around 7.2 GJ of electricity is generated, delivered to the grid and used by appliances.
So 30% or thereabouts of the original stack of coal is converted into work.
In other words if we dug out the coal with and electric digger that used the electrical energy produced from the coal that it dug, it would eventually grind to a halt.
Because each ton of coal lifted would take 70% more energy to shift than it produced in electrical power.
I’m struggling to quantify the software that makes up a business system in the same terms.
We have to discount as we did with the coal the act of creation (or digging), so the only useful benchmark is cost.
We can say that cost = energy, as every cost is paid for by the work we do and the energy we expend doing it.
For sure there are losses in computer systems and equations that describe work done over time by processors and while these are very interesting to computer scientists they don’t answer the fundamental question that we are posing here.
Which is ‘Does a modern computer system offer value in terms of efficiency and how do we define it’.
In simplified terms we can look at the cost of a typical system and compare having it against not and we can do this in perceived terms.
A ten user ABSsoftware system using a subscription model with licenses for a server could be £1500 p.a. or £150 per user per year. (2017)
Add in bespoke development at £5000 and assuming that a developed system will be in service for between 5 - 10 years we can average that to 7.5 yrs divided by ten with a cost of £67 per user per year.
So we have £150 plus £67 making a total of £217 p.a.
Assuming that each user interacts with the software for 5 hours per day through a business year of 240 days (excluding holidays) That’s a total of 5 x 240 x 10 = 12000 hrs of work (1200 hrs per user per year)
£217 ÷ 1200 hrs = 18p per hr
So how do we work out the efficiency of the system?
We could say that without it more time would be taken filling spreadsheets and paper forms, retrieving information on these and that lost time is a function of the efficiency of a software system.
So let’s take an example.
Sally receives a call from Bob. He asks if she has widget123. She puts him on hold and calls stores, stores as ever is busy with a customer trying to locate a part that he wants. So he doesn’t answer.
Sally says “Sorry, can’t get stores at the moment can I call you back?” The caller says OK.
Sally writes down Bob’s number and the part he wants and adds a reminder to call him back later.
This whole episode takes over five minutes, Sally still has more work to do on it and the potential customer hasn’t got what he wants.
Contrast this with a software system that knows what stock there is, where it is and allows Sally to see it.
Bob calls up and Sally answers.
Bob asks for Widget123.
Sally says “I’ll just have a look for you”.
She opens the parts app and types in widg, the app short lists all parts with widg in their name, there are eight items. Sally looks down the list and clicks on Widget123 it shows there are ten in stock with the price.
“Yes we have those, would you like one?”
“Yes please” says Bob
In the same time it took Sally to say ‘I don’t know’ she has said ‘Yes’
Sally is happy and so is Bob. The job is done with no loose ends.
There is a high chance the Bob will be a returning customer and Sally can get on with her next job.
Software isn’t as easy to quantify as Coal but it’s benefits and low cost per user provide significant contributions to the way a company can grow and make a modern joined up system a must have for any size of organisation.
