Agility

Yesterday our new prime minister Malcolm Turnbull talked about staying agile, trying to figure out what’s working and what not and to see where we have to adjust and where we can improve.

It’s true, our manufacturing policies of the 1980ies and 1990ies can’t work anymore, because that was before we had that economic power house China buying all our commodities. Now we have to adapt to the changes imposed by the emerging and faster and faster developing digital age.

What holds in politics holds as well in all other aspects of the economy.

The problem people have with agile development was nicely expressed by Waleed Ali during a radio interview yesterday, when he asked on how we could anchor this, that is how can we hold the prime minister responsible when he has the right to change his mind and to something else than promised before the election. I think in politics we will have to change our mindset and look at what the goals are and what was achieved in respect to those goals, not how. And if we figure that the goals were wrong, then we have to keep an open mind and feel free to change them as well.

In business the situation is not dissimilar, and the questions asked are similar too. “How do you measure the success of an agile project?” was one question asked recently on LinkedIn. The problem is that agile and project do not mix. You become agile because of the realisation that you cannot fix the requirements even for the time the project is developed. What use is it if you spend weeks and months to get the requirements and even in that process figured that requirements were shifting, just to nail them at some arbitrary point in time and start the project on those fixed requirements? You know that requirements will continue to change, and if you keep them fixed then all you have after two years is a measurable project delivered but out-dated already.

Even in the early days of computing in manufacturing, when computers cost millions and were not much smarter than a simple pocket calculator, I saw centuries old companies with top reputation go out of business because they were not able to handle that transition properly. The situation today has not changed. If you don’t adapt, you perish. So how do you measure success? At the bottom line! Yes right, that’s the overall success of the business, and a business is not only software you might say. Well, in the mass customisation projects we do, software is everywhere and it has become the critical success factor, although not the major cost factor.

But still- how do you measure the success of an agile software development? At every iteration of course! So don’t be scared, you don’t lose control, you get a lot more of it! However, you have to measure at every iteration, and to be able to do so you have to understand what’s going on. But nobody ever promised that better control means less work.

And don’t forget: Agile might have become big now in the software community, but you have to be agile everywhere. In your marketing, your product development, your production technology, your software, your distribution. A lot is changing now in all those fields and many more, and at an ever increasing pace. Scary? As Malcolm Turnbull put it, we live in exciting times. It’s not only threads, it’s a lot of opportunities as well.

Where to Start with Mass Customization

Just let’s say you’ve figured out that you could go into mass customization and provide a much wanted added value when delivering your products mass customized. So where do you best make a start?

Well this depends a lot of where you come from. If you currently supply a bespoke product, handmade at a high price, then it sometimes might help to get some machines that can increase your productivity. However, in that case you will have to make sure your new machines will fit into the big picture, that is you will later be able to load production data directly onto that machine, rather that entering it all manually.

But in most cases, what you need first is all necessary production data on your server. This means that someone has to specify that bespoke product on the computer. At best, this is your customer. However, your customers are only prepared to do that if there is a benefit to them as well.

For example your locksmith is happy to enter all the data for your master key system, because this helps him with the specification with you, he can manage all the key and cylinder types and with their functionality, and apart from ordering electronically with the manufacturer it helps him as well to generate quotes for their end customers and later he can easily manage all the changes and extensions the end customer wants.

If you need a new kitchen and can’t design that yourself, then the installer is happy to do that for you. If his manufacturer can automatically quote and he can generate his quote to the end customer from that simply by adding his margin, labor and transport, then that is a big help to him.

However, if you try to sell to the end customer directly, then you face quite a different challenge again. In that case you are communicating with a user that will use your system very infrequently. That means he is not familiar with that user interface at all. And that in turn means you have to provide an easy to use interface which resolves all the complexities in the background.

From a programming point of view, this is challenging enough. However, in most cases it is not good enough, because people want to be able to verify what they get, not only in a technical sense by verifying the measurements and/or all the other data they entered, but they would like to have a realistic image of what they get as well.

For example if produce custom sofas, would it not be great if the user in the end could have a 3D colors image of it? If you can provide a few so your customers can show them around to their friends and ask for comments before ordering, then there is a significant bit more value in it for your customers and that will attract them to your site.

As programmers we know that paradox that the easier a user interface is to use, the more effort there was to make it all work properly. On the other hand, if a user interface looks simple then our customers usually think that it must be simple to program as well. However, just let your software provider explain what all has to happen and you will understand. More often than not, this design tool will be the most complex and expensive piece of software in a mass customization project.

However, as far as cost is concerned, the machinery will be the dominating cost factor anyway. But I will discuss the production side of mass customization in a later blog.

Mass Customization

On this site I would like to talk about mass customization and later on maybe a bit more about other forms of digital manufacturing. So in my first blog post I would like to discuss a bit in depth what mass customisation is and in later posts give and discuss examples.

Mass customisation is the automated manufacturing of bespoke products. That does not mean that every process and everything exceptional and a bit too far out of the ordinary is handled completely automated. But to me, mass customisation means that the big mass of similar products is produced automatically.

And there is another thing. Not only production has to be automated, order processing has to be automated too. This can go over the whole supply line, so if a window manufacturer can mass customise, then the glass manufacturer that delivers the insulation glass units (IGUs) has to mass customise too. In fact this is a good example for mass customisation. This industry can mass customise for a long time already, and we have been lucky to provide the software for many such projects over the last twelve years.

In the window case the process can start with the architect exporting the window specifications from his drawings and send them electronically to the window manufacturer. The window manufacturer will then add his data, for example profile types and IGU specifications and order with the glass manufacturer. IGU sizes are calculated automatically by the window manufacturer’s software. Because the window manufacturer knows his supplier’s pricing model, he can quote automatically. The same happens at the glass manufacturer and if they have to order special glass in, then that goes automatically to another glass manufacturer too.

When released for production all production data goes to the machines and production is managed automatically. Not everything is automated though, for example nonrectangular shapes are cut automatically on the glass cutter, but broken out manually. Because this is a small part of the overall production and a process that’s difficult to automate, this is a reasonable way to approach this.

So this is my take on mass customisation. To make it work and to be competitive with standard size mass produced goods, one can charge a premium for customisation, but not a fortune. Thus all products within a regular size and shape range must be produced with a degree of automation as high as possible. More often than not this manufacturing infrastructure then allows with little extra effort to produce goods a bit further from standard with only little additional manual intervention. In the glass case almost everything except breaking out a nonrectangular piece is still automated, including quoting and order processing.