Platt Perspective on Business and Technology

Quantifying business intelligence valuation in terms of systems-indeterminacy 5: deterministic and stochastic valuation models and methodologies and their alternatives 1

Posted in macroeconomics by Timothy Platt on December 30, 2013

This is my fifth installment to a series in which I discuss and analyze the valuation of information in a business context and from a due diligence and risk remediation perspective, and in terms of what in a physical systems context would be called quantum indeterminacy (see Macroeconomics and Business, postings 137 and following for Parts 1-4.)

I have been discussing the valuation of information and particularly of business intelligence under circumstances where:

• Valuation might appear to be more deterministically set, with agreed-to standards for establishing a market-wide agreed-to fair market price for specific types of information products (see Part 3) , and
• Circumstances where like-for-like comparisons and market-standardized processes cannot be expected, and where ranges of prices offered and accepted would be expected to fall along a statistical distribution curve if some same-information package could be repeatedly sampled for realized valuations through independent bids and in such a way that separate biddings did not influence each other. For the buyer’s side of that, and as what might be considered a close possible approximation to it, consider sealed auctions where potential buyers make their bids without knowing what any other potential buyer might be offering or even if there are other bidders – and only that there might be.
• In the real world, I have to add, the type of multiple-independent bidding process that I suggest above would still create influence between bids because all bidders would know that other bids were possible, influencing what they would consider offering.
• And it is of the nature of business intelligence that even if bidders know something about what would be included in an offering, they cannot know in advance exactly what that offering consists of and in all details that they would need for a full due diligence evaluation of its true worth to them – and certainly for novel and one-off information products as discussed in Part 4.

And this brings me to explicit consideration of deterministic and stochastic valuation models and the gray are in-between them of semi-deterministic models. I have been invoking them all in this series, even if I have not always explicitly been identifying them as such. My goal here is to more directly bring these approaches into this narrative. And I begin that by considering valuation process subsystems and the relationships between small numbers of causally connected variable factors in them.

• If such a subsystem is deterministically connected, then if you know the precise numerical value of any all-but-one subset of the variables there, and you know the mathematical relationship between those variables and your still unknown, then you can predict a precise numerical value for that last deterministically related variable. For a simple physical systems example of this, consider Charles’ law from thermodynamics. If you know the volume of a fixed amount of a substance that is in a gaseous state and you know its temperature, and you make a precise known change in the temperature of that gas, then you correspondingly change its volume and to a specific new exactly predictable value, at least if the substance in question behaves as an ideal gas in the temperature range that you are testing this system at.
• If such a subsystem is stochastically connected, then if you change the value of one of its variables by a precise amount and in a precise direction, one or more of the other variables in this subsystem can assume any of a range of values. But their values actually arrived at do not follow a random pattern. If you repeatedly perform the experiment of making the precise same change in the one variable and with the second variable always starting with the precise same starting value, you will see its new value following a predictably reliable probability distribution curve and the overall subsystem to show statistically predictable behavior as to values reached.

In principle, there is always going to be some statistically representable variability in essentially any real-world process of determining the valuation of business intelligence. But in practice, there are also going to be at least some system subsets that for all intent and purpose behave as if deterministically interrelated, and with a set value for one causally connected variable in them predictably associating with a single value for another, or at least with a valuation for it that stays so close to one fixed value that any deviation from it can be considered insignificant.

As a tricky example of that, consider in-house determination of the realized value of a well-used and relied upon trade secret formula for a beverage. The business holding and using this trade secret business intelligence has as a matter of record, precise information on what it costs to produce, package and distribute its product and they know precisely what their gross profits are for it, and their profits net of those and related costs (marketing expenses, etc.)

• And if these values have held steadily within narrow limits and are predictably consistent,
• Then this facet of overall marketplace valuation determination for this trade secret can be said to behave as if highly deterministic,
• For its capacity for predicting the value for any variable in it with acceptable levels of range accuracy,
• Given knowledge of the values of the other variables in play here.

But I identified that as a tricky example. Let’s consider the tricky part of this, and with a specific working example company and beverage: Coca-Cola, and more specifically coke classic – the standard Coca-Cola beverage that people have been buying and enjoying for generations now.

There are a seeming multitude of competing alternative-formula cola beverages on the market ranging from major competitors to small local and niche market offerings. But Coca-Cola, with its securely protected trade secret recipe for its flagship beverage has consistently been a leader in its industry. What would actually happen if a competitor was able to acquire and use their precise original formula and in a way that would allow them to explicitly market their product as being a direct knock-off and a precise duplication of Coca-Cola? If this were a new product with little if any realized name recognition value yet, then this loss of proprietary control over their trade secret might make a tremendous difference for the profitability it would bring to the Coca-Cola Company. But with its deeply entrenched name recognition and its loyal customer base in the face of competing cola products, one more new competitor would probably not make so much of a difference. And many if not the vast-majority most of current Coke customers would still want and buy “the real thing.”

So if you add into this in-house valuation subsystem, measures of valuation from holding the manufacturing process for this product as a closely held, exclusively proprietary secret and the value of that variable of how closely it is held were to change – you might not actually know in advance what the impact of that change would be or even if there would be one. This I add, brings a non-deterministic element into this valuation subsystem and one that I note would not be expected to follow any predictable statistically modeled stochastic approach either. I have just added in a wild card variable here, and a none-of-the-above for the modeling systems I have just been discussing.

I am going to continue this discussion in a next series installment where I will look at these issues from the marketplace and buyer’s perspective, and as a full-context business information valuation problem. Meanwhile, you can find this and related postings at Macroeconomics and Business.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: