Lean Six Sigma in IT – Article
How apply Lean Six Sigma in IT
Article written by Tomas Tiefenbach
In this series of articles, I want to introduce to you the 8 biggest “money wasters” in the company. What is more, I will also show you various ways to get rid of them so your company can increase its profitability. There are actually two ways to increase profitability of your business:
One way is to increase sales, turnovers, create new products, and search for new clients. However, we both know how difficult it is to find new clients or convince all clients to buy more of your products. You also know how difficult it is to keep your sales team motivated in situations where sales targets go up almost every second month.
The second (I belive more easy) way to increase profit is to focus on money that is already in your business by just stopping money leaks. Is it effective? Definitely! The simplest way to stop money leaks is to fix inefficient processes. If you improve your processes using ROI (Return on Investment) criteria, you will definitely decrease operation expenses and even if your turnover remains constant, your profit goes up.
In fact, reducing money wastage generates you extra money that you can use to develop your business or spend on higher bonuses; it’s just up to you. If you choose this second easier way, you can continue reading because I will show you where to find this “extra money.”
Some of the service providers falsely believe that they have no over production because they provide services ad-hoc = only when they are needed.
Example: a manufacturer who produces more goods than he can sell VS a call center that provides services only when customers call. In this article, I will show you how the belief is far from the truth.
Overproduction occurs when the company buys more resources than it actually needs – for example, IT department builds an overzised solution. This solution can be oversized in many different ways – hiring more operators at the call center, buying more server resources, paying for thicker Internet line than is not needed…
Buying more resources than is actually needed is the most common argument supporting overproduction. In fact, it’s also the most EXPENSIVE way to cover peaks. In most organizations, utilization of server resources (HDD, CPU, RAM, connectivity..) is very poor: “Only about 25 percent of the available processing power of virtualized servers is utilized by many companies that adopt virtualization”, says Gartner analyst David Cappuccio. More than half of the clients we talk with have this situation” he says.
It’s not difficult to figure out why companies waste so much money: “This efficiency gap occurs because businesses typically add new virtual servers rather than place more workload on existing virtual servers. And that is a big waste of money” says Cappuccio in the interview.
It’s also quite easy to count how big this waste of money really is – let’s assume your company bought servers worth of 500k EUR and you utilize only about 25% = 75% waste = 375k EUR wasted + operation expenses to support these extra servers.
What are cheaper ways to cover peaks? The first way is proper demand management. Demand management is nothing else other than motivation of clients (those who produce demand) by positive or negative incentives in order to influence their behavior. Real life example of demand management:
Telco companies offer cheaper call rates for off peak hours. Such a discount motivates clients to make long non-urgent calls during off peak hours. As a result, Telco infrastructure is less busy during peak hours (company don’t have to buy additional capacity) and is better utilized during off peak hours (Telco provider has to pay for infrastructure even when it’s not used so it is better to earn something at night than nothing).
Another example – hotels increase prices during high season when they are fully booked (when demand goes up, prices also go up in order to maintain equilibrium). It’s far more effective to increase prices when there is high demand rather than increase capacities of rooms (that would be otherwise empty for the rest of a year).
One friend of mine (CIO of a mid-size Internet retail provider) told me “We need more servers to cover peaks at the end-of business cycle. Our accounting department has to prepare a lot of invoices and they consume most of our server resources at this time. There is no way to motivate clients and manage demand so we can produce less invoices.”
I told him “Yes, you are right. Maybe you cannot decrease the number of invoices but you don’t even have to.”
He looked at me quite surprised. So I asked him “Why does your company bill all clients in the first 2 days after the end of the cycle if your legal obligation is to bill clients within 15 days after the end of the cycle?”
He thought about it and replied “hmm, I don’t know. I need to ask our CFO but if it’s really possible to schedule billing over longer period, we would have lower peaks and won’t need so much resources.”
It is possible to apply demand management to ANY kind of business and I’m serious about this. However, there is even a stronger killer of overproduction.
Most CIOs think they need everything triple redundant just to make sure business operations can run smoothly. They invest in high-end storage systems, double backup Internet lines to every branch… I’m not saying it’s wrong. I’m just saying it’s EXPENSIVE!
A friend of mine, a successful IT consultant, told me a very interesting story about one of his business continuity project. One of the business locations he was auditing with other colleagues was the backside car entrance to the Data Center. There was one old woman sitting at the reception. She was writing car drivers ID details into a computer while opennig the gate. Basicaly, she was manually feeding the system with data about who and what time entered the Data Center. A security analyst who was also doing this audit immediately found out a single point of failure – only one computer, no external power supply and connectivity is also without backup. If this computer breaks up or Internet connection fails, the company would have no log about who entered or left the Data Center. This would create a possible threat to security.
The security analyst proposed a replacement of the computer with a laptop (in order to prevent short power failures), backup connectivity and monitor the Internet line: “We have to do this because it’s a possible threat to the company’s security. I’m just wondering why nobody did this before.”
My friend thought about it and asked the old lady “What do you do when your computer breaks up?”
The old lady thought for a while and responded “No problem, I just write those ID details on a piece of paper and when somebody fixes my computer, I put all the data from the paper to the system.”
Waw! Solution for €0.10 that fix the same security issue than super-secure solution for €1200/4 years.
I read from somewhere that there are more planes in the World than parking places on all airports. How is it possible? The probability that all airplanes stop flying at same time is almost zero. I don’t know what is the risk mitigation plan for such an extreme scenario but what I know is that even Telco operators don‘t have enough capacity to provide call services to all users. How do I know this? Try to call between 00:01 to 0:05 at New Year’s eve and you would be lucky to make a call (and I assume not even 70% of all connected users try to call at this time).
The reason why Telco operators don’t have enough capacity to cover peaks at New Year’s Eve is simply because it’s not worth it. Why buy and maintain massive infrastructure when they actually need it only 0.001% of the year? People simply got used to it and almost nobody quits the service just because he couldn’t make a call during the first 5 minutes of New Year’s eve…
In hotel industry, it‘s very common practice during high season to allow booking of more rooms than are available. I was (un)lucky few times and found out that the hotel I booked online was full. The first time I was a bit upset “Why did I book a room if you don’t have room for me now?”
The receptionist always calmed me down by saying: “Sir, I’m really sorry we don’t have room for you but we can offer you room in a hotel just cross the street. It’s actually a 5 star hotel (I’ve booked 4 stars). You can stay in the hotel with no extra charge.”
I tought…Hmm, sounds good. Better hotel for same money, why not… so I told him “OK, I’ll try it.”
What is the business case behind this practice? Let’s say a hotel has 50 rooms and they allow booking of 51 rooms just in case a customer doesn’t show up.A room in a 4 start hotel costs €50 and a room in that 5 star hotel costs €70. However, when a hotel is booking room in another hotel (e.g. same chain of hotels or partners), they don’t pay full price – for 5 star hotel, they pay for example just €40 and not €70
Calculating possible loss
Loss of profit: €25 (turnover €50 minus expenses €25)
Probability client doesn’t show up: 80% (historical data)
€25 profit * 80% probability = €20 possible loss of profit
Calculating costs of risk mitigation
Cost of solution: €40 (paying for another hotel room)
Probability that a client shows up: 20% (historical data)
€40 cost * 20% probability = €8 cost of solution
Is it worth to risk €8 if you can earn €20? Definitely yes!
What is different in your IT department compared to hotel when talking about monetization of the risk? All you need to know is just
- Probability of the risk (mostly it’s not difficult to count it based on hard data you already have).
- Count loss in profit.
- Price of a solution that helps you to mitigate risk.
Maybe you find out that paying fines for SLA breach is much cheaper compared to large investment that helps you prevent SLA breach…