The commercial Butterfly Effect, and how to control it

The commercial Butterfly Effect, and how to control it

If you’ve heard of the ‘Butterfly Effect’ you might think it’s an unmemorable movie made in 2004 starring Ashton Kutcher, or you might confuse the term as associated with ‘chaos theory’, which states that the number of variables in certain systems (for example meteorology) are so vast that it is impossible for humans to predict future outcomes with any measure of accuracy.

There is a common misapprehension that the Butterfly Effect is so called because a famous Massachusetts Institute of Technology (MIT) professor, Edward Lorenz, was studying weather patterns in the 1960s; he was supposed to have said that meteorological systems are so difficult to predict, that an event as tiny as a butterfly flapping its wings in India could cause a hurricane in the USA some weeks later.

Lorenz never actually said this, but compared the potential accuracy of forecasting to the likelihood of the Butterfly Effect itself. In short, almost impossible.

However, in the business world,  there’s no denying that in some systems, especially within the manufacture of complex products, a tiny change in specification at any part in the process can have extremely far-reaching consequences for the final outcome of the manufactured product as a whole.

Clearly, in commercial terms, if a company manufacturing aircraft parts were to quote a price to an aerospace company like Boeing, then to discover they had not accounted for a product specification change, the final price would be inaccurate and indeed their company reputation would be in jeopardy.

To prevent this industrial butterfly effect from occurring, many larger companies are turning to CPQ software (Configure, Price, Quote) – which uses artificial intelligence (AI) to account for and control the vast number of interrelationships between components in a manufacturing project.

In computing terms, this is not known as a butterfly effect, but, more accurately, as the concept of ‘combinatorial explosion’.

Combinatorial explosion refers, in commercial and manufacturing terms, to a concept where the number of possible combinations or permutations within any given project increases exponentially as the number of factors to consider increases.

Such is the complexity of combinatorial explosion that before AI was able to be applied within practical computing solutions, problems were so vast that they couldn’t be predicted or solved with accuracy.

Projects often failed for unforeseen reasons. This could cost manufacturing and service companies huge sums of money, or, in the worst case, could even put them out of business.

But CPQ software uses AI to deal with combinatorial explosions in a decision tree, rules-based architecture.

The platform compartmentalizes and assesses the interrelationships between relevant factors in any industrial project, then even suggests workarounds and solutions that might take a team of highly qualified engineers days or even months to conceive.

Let’s make an oversimplification for an example. Imagine that an aircraft spares manufacturer, Acme, is asked by Jetairlines.com to quote a price for the overhead reading lamps that sit above every passenger’s head on the aircraft.

A typical large airliner can seat 475 passengers, and the company is refitting all the 500 aircraft they’ve ever manufactured and still in service. That’s about 238,000 lamps to quote Jetairlines.com a price before manufacture commences.

The initial (fictional) problem is that the wiring on any aircraft built before 2010 is of a different specification to those manufactured after that date.

So now it’s not just 238,000 lamps to quote for, it’s the 200,000 pre 2010 models and the 38,000 post 2010 units.

To add another complication into the mix, some airlines want to have daylight balanced bulbs to help passengers stay awake on flights, other airlines fit dimmer, lower wattage bulbs to their lamps on long haul flights.

This is another example of technology being used for human wellbeing, because passengers’ body clocks expect darkness at a given time.

So, Acme now must source and quote for the lamps separately to the bulbs for two different types of fitments and also various wattage powers and differing color temperatures.

And guess what, the higher wattage bulbs become too hot to pass new safety regulations, so Acme is obliged to fit capacitors inline before the bulb holders to keep the voltage down. Already the complications are becoming intractable!

But if Acme were using CPQ software, each and every component would be assigned with a series of compliance and specification rules by drop down fields being filled in by engineers on a dashboard at the outset of any project.

When a lamp holder was input into the CPQ platform as a component, it might contain a sequence of rules in computer code, in the back end of the system; for example:

[Lampholder model X] = true = compatible = [bulbs a,b,d,f] incompatible {not true} [bulbs c & e]. [Heat Resistance] = true date compatible [post 2010] {not true} [pre 2010] [if {dateincompatible} required capacitor model Z retrofit 12volt]… 

The number of factors can be almost endless. But the AI within the CPQ can determine at lightning speed how many lamp holders, capacitors and bulbs are required to determine a unit price, along with adding minimum profit margins, international shipping costs and numbers of spare bulbs required per lamp.

If that seems complex, imagine the same scenario with an aircraft’s fuel pump or hydraulic wing-flap systems! 

Using CPQ, the AI can not only provide accurate transparent quotes, but cope with tiny changes causing butterfly-effect long term consequences.

Even within the context of contemporary world po

litical tensions, technology is making our lives more comfortable and often easier, but the complexities involved in getting there can be mind blowing. Fortunately, the AI in a configure, price quote (CPQ) platform takes the headaches away for businesses competing in an ever-tightening marketplace.

Read more business articles from our guest authors at SugerMint.