Searching for the Moon

Shannon Clark's rambles and conversations on food, geeks, San Francisco and occasionally economics

Network analysis of a corporation

Posted by shannonclark on January 10, 2005

I am seeking a corporation (or multiple) to perform the following analysis. As this is a trial run, my rate would be competitive, all data would be cleansed and shared prior to any publication, other details negotiable (no rights to my research or work, but a right to have access, perhaps to use for internal purposes, external purposes negotiable).

The research would do the following analysis:

Look at the corporation’s business and finances from a “pure” network perspective.

That is all parties and counterparties (vendors, partners, employees, governments, shareholders, etc) as nodes within a multi-dimensional, temporally bound network. The modeling would seek to model ALL transactions against time in terms of network transactions – mapping duration of link, value of link, and ALL parties impacted via the transaction(s).

From this initial set of data – which likely requires historic data from accounting systems as well as very likely other systems – I would then also extrapolate data and trends in terms of logical time units – based on the data, not preconcieved breakdowns. Additional analysis on the basis of more traditional timeframes such as quarters and the fiscal year would also be analyzed.

The main aspects of this would be to try to isolate and highlight the short and long term trends around – for example – specific lines of business, particular units/sales channels etc.

My thesis going into the research is that this type of analysis will very likely highlight opportunities, as well as anomolies. It may also help in mapping how complex organizations and partners are tied together.

My past research, into contracts for example, indicates that it is not uncommon at all for the actual transactions a corporation engages in to differ from how they are “supposed” to happen – payment/recievables do not occur in a timely manner, penalties are incurred (or conversely penalties that should be enforced whether on suppliers for mistakes/delays or on customers are not enforced etc.). All of these and more have a very real impact on the bottom line.

Further, many corporations – airlines perhaps being a prime example – while claiming to have a firm understanding of pricing – clearly do not at present manage to sell their services for more than they have costs (i.e. they are not currently profitable – or at least claim not to be). A better understanding of the timeline as well as inputs/outputs of the network around, say, a particular flight – over time – could be highly valuable and informative.

My thesis is that this study could help isolate and in a generalizeable manner explain the value a company such as Dell manages to achieve, but many other firms do not manage.

On one level it may appear an exercise in cash accounting – but I think there are key and vital differences in the approach, the process and the goals. My goal is not to provide an auditable set of accounts. Rather, my goal is to map out and highlight the network process via which the company opperates – placing it inside of a greater network of other firms, individuals, organizations and governments.

To start I would like to work with a small to midsized firm, or a very isolatable division of a large firm. As the details are worked out, especially around how the data will be manipulated and especially visualized, I would hope to work up to being capable of providing a rich analysis of a large firm, organization, or even government agency.

The process would be somewhat untraditional – this is not a “simple” case of spreadsheet modeling, yet it is also not a case of pure custom software development. rather it is a research and analysis process, as well as a test of the process of large scale data warehousing. I am envisioning a graphics process that would involve scaled data and which likely would involve animation to depict the passage of time – i.e. to show slices of data over time. Ideally the tools might allow for the tracing of a specific transaction and its effects over time across the nodes of the network being shown. (limited by the data and effects from transactions within the bounds of the dataset – effects from other external sources would initially not be modeled.)

To explain this – consider an employee’s salary from the corporation. On the 15th of the month she receives a direct deposit of her salary. From a network perspective a few days prior to the 15th money from the corporation will have been transferred from one source of funds, probably a bank to another bank from which payroll would happen. From this account funds would then be transferred to the employee’s bank, as well as into some holding accounts (perhaps) earmarked for various insurance providers and related taxes, as well as perhaps deposits into flexible spending accounts, HSA accounts, 401k accounts etc.

Each of these accounts while earmarked on the employee’s paystub as occurring on the 15th might, from the perspective of the corporation be settled on a very different schedule – end of the month, quarterly, even yearly.

So, just one employee’s salary could represent a web of connections and linkages being exercised on behalf of the corporation.

The employee, however, from the perspective of the corporation could also very easily incur a wealth of other costs (and hopefully in many cases generate income as well). These could include monthly licensing costs for tools and software used, expenses that are specifically expensed, obligations that are acrued (vacation, training dollars, education reimbursement, stock options, stock grants, bonuses etc.). Most employees occupy space in a facility somewhere owned by the corporation and at that facility use equipment and supplies, etc. In short understanding what a new employee represents to the corporation can be a complex task – I posit that by analyzing the corporation from a network perspective, this web can be made sense of and new relationships, anonomlies and interdependancies can be highlighted (and thus potentially acted upon).

Clearly this could be a big project.

If you are interested in discussing it further with me and/or have a firm whom you think would be interested in this type of analysis and consulting, I welcome your comments and/or emails.


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

%d bloggers like this: