Recent news of airline seat allocation and the bumping of paid passengers has brought increased awareness to the complexity of airline seat yield management and capacity optimization. In the 1970s, passenger load factors averaged under 60%, and the airline industry was losing billions. Today, the industry’s load factor is approaching 90%, and in 2016 international carriers’ estimated profits reached $36 billion. This is a direct result from increased efficiency in the airline industry’s yield management and sophisticated capacity optimization.
If an airline flies a 747 400-passenger jet between Orlando and Miami, traditional optimization would try to find the maximum revenue for the flight. For many years, airline optimization maximized the profits on an assigned capacity between cities. For some aircraft on some routes, it was profitable. However, for the airline as a whole, optimization meant finding ways to maximize the profitability across the entire fleet. It would often have some aircraft fly in unprofitable routes as a feeder into significantly more profitable routes. Capacity and yield management are constantly analyzed and change by the minute. As we have recently watched in the news, overbooking and paying to bump passengers may not appear to make financial sense but, as a whole, it results in the airline realizing a greater profit than not selling at the capacity level and allowing seats to go empty.
The development of airline yield management (optimization) comes as a result of understanding the true sophistication of applied mathematics. It is significantly more complicated than simple inventory management.
As children, when we learned arithmetic and the ability to add and subtract, the world suddenly became bigger. When we later became introduced to multiplication, then algebra and geometry, we became confident that we had just been taught advanced mathematics. We may even do well with Calculus, Differential Algebra and Probability until someone introduces Krylov Subspace Methods and the Algebraic Eigenvalue Problem. It is then we realize how complex the world of Mathematics is.
Learning Software Asset Management is similar to the progression of mathematics and airline yield management. If you talk to 100 different SAM “experts” you will get 100 different definitions of Software Asset Management. However, the common thread throughout all definitions is knowing what is licensed and what is deployed. That isn’t mature Software Asset Management. That’s inventory tracking. That’s basic arithmetic.
Some more knowledgeable SAM consultants will offer an additional layer of Software Asset Management by rationalizing the licensing metric for the software purchased. As valuable as that it, it is analogous to an airline trying to justify the best seat pricing on a 747 from Orlando to Miami. It may be a form of moving from basic arithmetic to algebra, but it isn’t mature Software Asset Management either.
Mature Software Asset Management begins before business decisions are made. Traditional growth in a datacenter begins with a line of business deciding to move forward with a need. They communicate the need (new reservation system, claims system, customer service system, etc.) to the hardware capacity planners who decide how much hardware capacity is required to support the business requirement. A hardware decision is made and then passed on to the software department to deploy the necessary software on the hardware. Software engineers then pass their actions on to procurement to purchase the licenses “at the best price”, and SAM managers monitor the deployment against existing entitlements.
Mature Software Asset Management isn’t managing a sequence of decisions post-mortem. It’s getting in front of the hardware decision to make sure the right hardware was purchased, deployed in the right environment, configured in the right virtual pool, with the right software, licensed for the right use, and purchased at the right price. Mature Software Asset Management is knowing what to buy, when to buy it, where to deploy it, and when to terminate it.
Having personally seen the inside of hundreds of data centers and collected millions of data points from over 1000 data centers over 20 years, there is a reason the majority of data centers spend significantly more than industry best in class in software. Most data centers confuse Mature Software Asset Management with inventory management.
For a data center to do Mature Software Asset Management effectively they need to follow a few simple steps:
In addition to traditional Procurement and SAM managers, an overall knowledge and coordination of optimization projects is required:
- Hardware Architecture – The Hardware team needs a solid knowledge of how the underlying hardware decisions impact software licensing and costs. Understanding how to plan, design and grow a data center’s overall capacity at the lowest possible TCO is crucial. This is the airline director of planning and scheduling who knows all the similarities and differences between aircraft manufacturers, size, configuration, and performance capabilities. They understand the market demands and can plan the appropriate capacity to meet the demands accordingly.
- Virtualization Architecture – Individuals who take into consideration software licensing while virtualizing an environment to avoid unnecessary increasing the risk or costs of non-compliance. This is the airline crew scheduling expert who understands how to avoid putting crew trained on a Boeing aircraft from routes only flying Airbus. They understand how to regulate and maximize the mix of aircraft capacity necessary to meet fluctuations in the market demands.
- Software Licensing– An individual who understands every possible licensing metric and coordinates and communicates with the Architecture teams. They know the differences in licensing options, even those which the vendor ignores or will tell you, often incorrectly, are not available. This would be the airline scheduling manager who knows that a 747 can’t fly in airports with a runway that is too short, and can manage the number of flights landing in an airport to ensure they don’t utilize more landing slots than allocated.
- Finance expert – Someone who can run every possible financial scenario before the hardware and software decision is made. They perform analysis on the total cost of ownership across all decisions, hardware and software. Analysis will include deciding if 1000 licenses at 40% discount is better than an optimized deployment of 500 at full list price. This is the critical piece to ensure that the required software technology is acquired at the best possible cost structure.
There are 101 SAM processes published and utilized, but if the process doesn’t get your costs to industry best in class, it’s not an optimal process. Don’t build a process for the sake of claiming to follow a process. Then, the process becomes the end goal and not the means to the end. Build a process that, if followed closely, will enable your data center to grow its hardware capacity without spending additional money on software. A U.S. federal agency has seen its hardware capacity grow by over 40% with its software costs growing less than 3%. Their cost structure is lower than any large data center I’ve seen. It didn’t use to be that way, but over time they’ve built processes to get in front of the necessary decisions so that they can grow capacity or virtualize servers, without increasing their software costs or becoming non-compliant.
This is the holy grail in software asset management. Unfortunately, for most data centers, entitlements and deployments are all they collect. For many, building an “Effective License Position” becomes the final outcome. Software discovery tool vendors claim that having their tool will get you to SAM Nirvana. The problem isn’t having data; the problem is having the right data and using the data properly and effectively. Having an ELP is essential across all software vendors. That is like airlines knowing precise seat capacity and passenger count on every flight in the system. However, knowing whether the proper capacity is in the right market at the right time at the right price is optimization.
Mature Software Asset Management is getting data centers to think beyond basic mathematics to advanced mathematics. It involves a team effort, hard work, and the ability of a SAM team to model the lowest possible cost structure before business decisions are made. It involves knowing industry best in class cost structure and having the right people, processes, and data to get there. Mature Software Asset Management requires delivering the highest level of technology at the greatest value by maintaining industry best in class cost structure, providing flexible growth of data center capacity needs while remaining compliant with software licensing requirements.