Technology pervades everything just about everything we do and touch. Technology is many things to many people. For some it is a catalyst for change. For others it is the vessel to channel innovations. Some people view technology as the pivot point to outdo their peers in a fiercely competitive environment. It is the bane of existence for those who have yet to tame technology's potency and intricacies.

Like it or not, technology and the world that it brings with it is here to stay and endure.

The one constant about technology is that it continually evolves; so it should not surprise you that our company name is Evolving Technologies Corporation. To put it plainly, this web site is all about our passion, technology, and how to put it to use.

Enough with the platitudes. It's the specifics that really matters.

What do you do if your data has gaps or missing pieces of information? or how do you smooth out the conflicting information in your data without altering your original data? In our article on constructing a "data overpass" we show you how to deal with such situations.

You may have noticed that some technologies are easy to grasp, easy to start with, but don't scale particularly well. It can be a simple process to write formulas in a spreadsheet, and replicate them. It's definitely not fun managing a large and complex spreadsheet that suffers from too many loose ends, a morass of formulas, and provides no easy way to verify correctness of formulas or results of computations. No doubt for many, scalability in spreadsheets is the challenge de jour. The problems with scalability doesn't stop there. It has an evil sister called Total Cost of Ownership. Large and complex spreadsheets are notoriously expensive to maintain, are are costlier still, when they are inadequately prepared to support good decision analysis. Fortunately, there are effective techniques and practices, and augmenting technologies to deal with these challenges. On our site, you'll lean about the Layering Approach design pattern which decouples the source data, the computations, and the presentation layers. Techniques like this helps to keep the process of maintaining your data and analytical tools, sane and manageable.

Sometimes the key in assessing situations, especially when the data you are working with is complex or voluminous, is to visualize the data, be it by sub-setting or intelligently partitioning the data, summarizing it according to various categorizations (we like to think of this as "digital origami"). Other times, tremendous insight can be derived if you could easily tweak a few parameters by turning a knob or moving a slider in your spreadsheet or analytical models. Such visualizations are often called dashboards. We have a lot to say about this. You may have already seen our web site www.XcelsiusJournal.com. If you haven't do check it out. You'll quickly discover, tools like this are a lot more than fancy visualizations. Behind the scenes of the point and click dials, sliders, and dynamic menus, are full fledged interactive mathematical models. They retain their full interactivity (see Figure 1) when are converted to a PowerPoint slide, PDF/Acrobat, or Flash file that you can email to a colleague or post on a web site.

Figure 1: This sample dashboard is not a screenshot. It is fully interactive (try clicking the various controls).

Taking What-If to the Next Level

Without realizing it, almost everything you do with numbers and data is a mathematical or quantitative model. Although a spreadsheet with data and a few formulas can easily be used as a mathematical model, its power becomes greatly enhanced when you superimpose some software enhancements and apply effective spreadsheet design principles.

Here's an example worth thinking about.

Imagine that your company is a pharmaceutical firm who is developing a portfolio of new drugs and drug therapies. Each of these products are at various stages in the development pipeline. Some hold tremendous promise but may be risky. Committing to some of these products may impact your manufacturing capabilities for other products. The timing of how soon you can bring a product to market, could have tremendous impact on your company's business and financial situation. There can be other forces at work, such as those of your competitors, who could introduce a competing drug six months before yours gets approved. Their product may be less or more efficacious (that is, effective) than yours. It may even be discovered that your drug gives rise to adverse effects for a specific group of patients.

There is a fundamental challenge with preparing a forecast or analyzing your "portfolio" with scenarios of this kind. The scenarios can be extremely complex. If you were to know in advance, some of the information you could figure out precisely where you stand. All other things being equal, if you get to the market before you competitor does, it enhances your positioning. The problem is that not all other things are equal. If you want to make an informed decision, you really have no choice but to take into account the full range of possible scenarios.

The problem is that not all other things are equal. If you want to make an informed decision, you really have no choice but to take into account the full range of possible scenarios.

Let's just focus a moment on timing to market. If you edge out your competitor by six months you might grab the market share. You come out around the same time, it's a horse race. Your competitor beats you by six months, you loose the advantage. Notice that in this isolated instance, there's a predictable relationship here. It is almost as if you could move a slider control (much like dashboard) to adjust the timing, and see the impact on profitability. The idea of the slider is nice, but what if you reliably believe you can get to market earlier than you competitor? How is that going to affect your financial assessment? Obviously, it improves it, but the concept of varying your timing using a simple slider no longer fits the situation. Think of having a slider (on timing to market) that you move back and forth by hand, watch what happens in your simulation, but most of the time, the values to examine place you in advance of your competitors, that is, the probabilities favor you outpacing your competitors. If you aggregated your outcomes based on one or more probability distributions, you could develop a risk or cost benefit profile.

The reality is still more complex. Your timing to market is dependent on a variety of other factors. Throwing more dollars in an R&D budget to accelerate the development of product X may cut out dollars for product Y. Unbeknownst to you, product Y could be a breakthrough miracle drug.

If you could set up a mathematical model in a spreadsheet, and vary all the different possible parameters or assumptions according to a variety of constraints and probability distributions, it would in principle be possible a usable composite picture.

Building a mathematical model in the form of a spreadsheet, it turns out, is not the overriding hurdle. Instead, it turns out to be overlaying the probabilities and constraints. Fortunately, there's a technology for doing just this. It's called a Monte Carlo simulation (named after a celebrated casino in Monaco). Better yet, Monte Carlo simulations can be performed completely within a spreadsheet. There are tools like Crystal Ball (a product of Decisioneering Inc., a Hyperion/Oracle company) that run as an Excel Add-In (see Figure 2). Better still, this kind of Add-In makes it possible to play out tens or hundreds of thousands individual scenarios for a complex spreadsheet model, varying hundreds or even thousands of variables, from a desktop computer, with relatively inexpensive software. The skill threshold is little beyond what you would need to create a spreadsheet without the Monte Carlo component.

Figure 2: Crystal Ball running inside of Excel

The SAT is not what you think it is

Simulation analysis and the technologies that can enable this kind of analysis is obviously important. For this reason, Evolving Technologies Corporation is launching the Simulation Analysis Technologies Journal, to focus on effective techniques and uses in simulation and modeling technologies.

While we don't shy away from being product specific in our journal articles, our focus is on the ideas, concepts, and things you can do with these technologies. Where appropriate, models and files will be made available, though you have to be registered to logon and download the files (registration is a one time process and it's FREE).

As you read through the articles, you'll discover that our goal is to get you to think about things differently. We welcome the dialog that results from this journal. Feel free to contact us. Click the Contact Us link appearing that the top of any of the pages on our site.

©2007 Evolving Technologies Corporation - all rights reserved.