All the techniques used in the JMark Services Information Environment Advanced Analysis (IEAA) course apply to supply chain networks. I’ve selected three and added a new fourth type based on Project Socrates (see endnote xi).  Our first step is to characterize the IE to understand the battlespace. We begin with baselines used in IEAA practical exercises.[ix]

Baseline Analysis

The IE is ever-changing, so we need to identify its various states relevant to our problem set at any time. Three enduring questions about the international system can help us decide what to include in a baseline.[x] Who are the relevant actors? What do they want to do to each other? What can they do to each other? These questions are not all-inclusive. They serve as a starting point for advanced analysis (includes synthesis).

Baselines are informational, “normal” starting contexts filled with dynamic relationships among actors and environmental forces. Four types of baselines are deeply relevant to supply chain networks:

• Cultural
• Technical
• Functional
• Sources of Competitive Advantage

Cultural

This baseline provides vital information on patterns and behaviors deriving from traditions, beliefs, attitudes, identities, and social norms. Culture tends to be described and interpreted as it is. Most people don’t explain culture as actionable causes because it presumably takes a long time to change, though the IE may be changing that for socially immersed populations. Consequently, “thick descriptions”[xi] are being replaced by a mishmash of historical and psychological factors characterized as mindsets, interpretations, personalities, perceptions, approaches, and worldviews. Adherents often claim their culture as unique or exceptional to find or reinforce their identities. Such peculiarities are critical to understanding behavior that otherwise may seem unfathomable or unexpected. Cultural differences affect supply chain labor and management issues in multiple ways–avoiding and accepting uncertainty, performing a job in collectivist and individualist ways, establishing trust, and implementing best practices.

Technical

This baseline is more tangible than culture. I’ll describe technical factors from Ervin Ackman’s perspective of creating advantage in three key attributes and related purposes.[xii] First, the structure of interconnected technologies provides insight into a state, society, or firm’s capability to compete, survive, and advance. Critical information includes the data, sensors, tools, equipment, material, skills, and knowledge to satisfy market wants and needs. A prime example today is the race to dominate the 5G technology that connects other technologies. Second, technology flow across individuals and groups reveals advances and gaps. Being ahead or behind affects performance and can change rapidly. A venue technology like 5G can shape advances in other technologies such as robotics and remotely controlled sensors and shooters. Third, the duration and timing of technology structure and flows characterize past and current conditions so we might anticipate future conditions. From differences in technology’s pace and rate change, actors develop advantageous upstream and downstream pathways. Again, 5G technology is a critical example, particularly if an authoritarian regime controls the global telecommunications network of devices.

Functional

Functional baselines reflect the IE’s diversity and potential for integrated control. Controlling AI technology relies on objective functions (tasks) that can vary, such as past performance versus the ability to learn. Here are four types of functional baselines subject to change. Attitudes and opinions among the public or select decision-makers function as inputs to a product or service’s demand, affecting supply. Economic performance such as revenue from energy sales, current account deficit or surplus, government debt, and income levels can influence sales and investment instantly. Government structure and informal power networks shape communication flows and incentives for suppliers across regions and locales. Finally, national and private media may be controlled or decentralized, serving as a censor or facilitator of marketing, shaping production and distribution options. The value that humans or a future artificial general intelligence place on a function is information that determines optimal solutions across different states and space.

Sources of Competitive Advantage

Sources of Competitive Advantage. All the previous baselines harbor sources of competitive advantage. Cultural advantages such as collectivist or individualist tendencies are not objective facts; they are contextual subjects that vary according to conditions. So, superior awareness of cultural practices and beliefs is an “information intelligence” (see ICSL Paper #39) advantage over competitors about what and whom to trust. Technological advantages drive competitive supply chain networks and much more, whether domestically developed or acquired abroad. Theft of intellectual property rights can create a technological advantage if domestic innovation does not suffer.  Functional advantages derive from a technology of some sort but not necessarily from a technological advantage. For instance, the strategic application of old technology can outperform “superior” technology ill-suited to the social context.

Developing baselines that characterize the competitive space is an artful science. All the baselines above impact a supply chain’s sources, manufacturing, and delivery. Baselines are critical reference points that need to be updated with other advanced techniques, starting with linkages.

Linkage Analysis

Mapping the nodes and relationships in the IE can simplify the complexity of the information environment. At first, it may not look like simplification until we recognize the structure of the complexity. When it’s unclear which relationships matter to our problem, collecting more data and assigning meaning to convert it to information can generate just another complex morass. AI-assisted human-directed collection, filtering and sorting can help us understand structures and relationships.

Savant X Seeker is an example of a hyperdimensional relationship analysis platform that finds interconnections among keywords and concepts in a database. I provided the human direction, framing “information” as described in Paper #48: operating system information (input-change-output) in physical and psychological dimensions.

Figure 4 is the initial visual rendering of complex relationships (the node and linkage output) after providing keywords (the input highlighted on top). In this case, the problem was to characterize the breadth of Iranian influence strategy. Seeker’s latent semantic analysis program calculated 4-6 dimensions of insight (the change). I provided the database in the form of credible articles and reports. Each relationship in the figure pointed back to passages in the database.

Figure 4

I adjusted the scale and scope of the search using the Node Control, Settings, and Nodes features to the left of the picture. This search led me to discover regional and global connections that otherwise would have required more research time at best. I might not have discovered them at all. Seeker also has an “Insights” feature that I successively activated for specific terms in the visual depiction. This sequence enabled me to examine fourth, fifth, and sixth orders of relationships beyond the initial 3-D picture (3-D insights yield 4th-dimensional relationships; insights from those reveal 5th, etc.)

This type of analysis could include keywords for time-sensitive chains in any industry. A database, desired performance effects, and multiple risks would be appropriate inputs, ingested by Seeker for subsequent human-directed search.

In this Iran example, what initially appeared as an intractably complex sphere yielded insights into proxy warfare, political interference, cyberattacks, narrative warfare, and financial flows.[xiii] Mapping linkages is an initial characterization of the IE. We also want to know how the linkages change over time. To do that, we turn to the next analytical technique, pattern and trend analysis.

Pattern and Trend Analysis

One way to capture the dynamic variety of relationships is to characterize patterns and trends as linear or curvilinear. I use these terms instead of “linear” and “non-linear” because they are more accurate. “Nonlinear” or “disproportionate” examples still use lines to describe relationships; they’re just curved lines as in an exponential increase or decrease.

A pattern is a consistent or repeating set of traits, acts, or any observable information. A linear pattern continues for a period in fixed ratios, perhaps changing to other fixed ratios. The pattern is usually represented in 2-D or 3-D.

A trend is a longer-term pattern, long enough to reasonably expect it to continue. We can calculate trends as derivatives of patterns. Just as the derivative of position is velocity (position over time) and the derivative of velocity is acceleration (velocity over time), the derivative of a pattern is a trend. A linear trend is a direction or vector—up or down, more or less, west or east, and so forth.

The US stock market is an excellent open-source example of linear patterns and trends.

Figure 5

A curvilinear pattern or trend is a changing linear relationship. Relationships consist of inputs that result in outputs with variable ratios:

Figure 6

The figure above is from a study that explored the relationship between a message’s rate of speed and the message’s persuasiveness on an audience. The pattern revealed the optimum rate of 181 words per minute–not too fast and not too slow for most people.[xiv]

The study concluded that the optimum speech rate was a conversational baseline rate. However, the researchers stopped short of predicting the pattern would become a trend. They might have calculated that derivative. Instead, they recommended investigating source perceptions and comprehension levels, factors that arguably influence the optimum speech rate for a particular person.

Supply chain planning often involves linear charts, patterns, and forecasted trends, but no plan survives contact with the competition or the environment. Curvilinear disruptions intervene in the way of shortages, declining capacity, attacks on infrastructure, domestic crises, and natural disasters. They may be intentional, accidental, or systemic. Indicators that precede unplanned-for events or appear as outliers in a pattern may be anomalies.

Anomaly Analysis

An anomaly is an unanticipated or unexpected value or phenomenon, a departure from a baseline, pattern, conditions, or trend. For instance, the figure below shows a sudden decrease in the level of violence:

Figure 7

Other general questions can probe the anomaly and lead to root causes:

Figure 8

Anomaly detection in supply chains is often designed to find incidents for managers to resolve:

“Anomaly detection can assist these managers in analyzing the data quickly to accurately spot unexpected behaviors such as identifying issues in orders, fulfillment, inventory, shipping, and more.”[xv]

AI-assisted decisions such as those provided by Azure Metrics Advisor (noted above) rely on collecting data from transactions. Managers may get pulled into the details of setting parameters to filter the data–what is an anomaly, and what is normal baseline activity for the conditions at hand?

When considering any solution, decision-makers should step back and ask, why do we think that what we observe is an anomaly?

Anomalies may be deliberately hidden activities or bait to lure our attention to them. They could also result from insufficient analysis and serve as signposts for further problem-solving. Here are three examples.

First, analyzing the scarcity of semiconductors can provide insights.[xvi] Increasing demand for microchips in more products has caused more opportunities for criminals to sell counterfeit chips. Would it be an anomaly for your competitor to risk purchasing semiconductors from an unvetted vendor? Seemingly irrational behavior may be an anomaly for a risk-averse competitor and expected behavior for a competitor who diversifies supply by any means. Can your competitor increase inventory or invest in joint ventures? If the answer is no, risky behavior is not an anomaly.

Second, a sudden spike in maritime shipping costs and ghost bookings may not be an anomaly if we understand why purchasers make multiple bookings. This method is one way to diversify supply if you cannot commit to long-term contracts. There are better ways to increase the reliability of supply, but many small businesses cannot afford them.

Third, AI-assisted surveillance of pharmaceutical cold chains can help resolve anomalies. Robert Boehm, former head of Bristol-Myers Squibb global supply chain, described an anomaly in the temperature parameters of a product. The company isolated the problem to a particular port, where 97% of arriving shipments were found frozen below acceptable limits. AI analysis of shipment histories discovered the root cause—flawed insulation gel.[xvii]

Knowing your own business and your competitor is key to revealing some anomalies as patterns. Asking why is this an anomaly and why is it not? is critical to understanding a dynamic information environment. Answering that question will not explain or eliminate all anomalies because uncertainty is pervasive. The point is to be competitive. Three approaches to anomalies are passive, active, and aggressive.

• Passive: a passive approach is reactive: collect more data for a later decision. As previously mentioned, anomalous risk-taking or price increases may become an explicable pattern after processing data into meaningful information.
• Active: an active approach anticipates anomalies and reports them. Like incident reporting, an anomaly incident management system can develop procedures to guide an appropriate response beyond collecting more data. For instance, Tive’s Open Visibility Network claims to provide logistic service providers a complete solution to track and optimize shipments in real-time. Suitable solution architectures can help managers convert passive approaches into action plans to build resilient supply chains.
• Aggressive: an aggressive approach tests or shapes the supply chain by inducing an anomaly. A simulation might test the implications of a labor strike, change in local policy, or weather event as a risk mitigation measure. An actual induced anomaly is what we associate with various kinds of warfare–a cyberattack, seizure of assets, hostile takeover, or other predator-prey scenarios.

We need baselines, patterns and trends, and anticipatory analysis to deal with anomalies aggressively. Without anticipatory analysis, one wins battles and loses wars.

Anticipatory Analysis

Anticipatory analysis is about forecasting changes in the inputs of information, which create new outputs. Another way to describe this thinking is the dialectic—a clash of contrary arguments.

Figure 9

This figure integrates dialectical thinking with our operating system definition of information. We need to synthesize what we’ve analyzed so far to anticipate or forecast a future state based on the current state. Synthesis is the result of dialectical thinking. The thesis is our initial input or understanding, which we compare to an opposed or changed argument—the antithesis. As we reconcile those two opposing views, we try to achieve a new output or understanding—the synthesis. That synthesis will generate theses to test, so the cycle of thinking goes on, or should.

Our advanced analysis techniques are a decomposition process—breaking down the whole of our IE into those parts. Linkage, pattern and trend, and anomaly analyses understand how those parts of the whole relate to one another and interact. In that sense, decomposition helps us determine the current state of the IE.

In anticipatory analysis, we imagine how the future could happen by rearranging the linkages, patterns, trends, and anomalies. We recompose the IE with those parts. There’s a lot of uncertainty involved—anticipating what might happen is not risk-free. Next, we reassemble the parts differently than what we decomposed. That synthesis of ideas produces two conclusions–what the future state will likely look like if we make no changes and what conditions we must change to shape the future we desire.

Now let’s apply that general theory to a recent example of anticipatory analysis: locating and eliminating the latest leader of ISIS in northern Syria.[xviii] The example is based on incomplete open sources but demonstrates how to create and use information in an operating cycle of dialectical thinking.

• Thesis: the baseline activity of ISIS brutality extends across northern Syria. Because land routes are limited and ISIS has no airpower, we can identify baseline movements and look for patterns. There are areas through which ISIS leader Abu Ibrahim al-Hashemi al-Qurayshi (“Abdullah”) would have to traverse. He’s likely to set up residence to direct operations, creating anomalies in the local population’s behavior that we can detect.
• Antithesis: our persistent surveillance of those areas yields nodes and linkages of activities and people, but no pattern on the leader himself. Rather than create an anomaly, he may blend in with normal local movements. At the same time, we observe that several possible locations do not exhibit normal local behavior. These may be diversions instead of a headquarters.
• Synthesis: the anomalies at one location match the pace of activity we expect to find at an ISIS leader’s residence—couriers, heavy security, and occasional sightings of an individual on the roof of the three-story building. Local informants provide tips while airborne and ground cameras confirm Abdullah’s identity. An innocent family lives on the first floor. Our continual observation of the location may create the opportunity to kill the ISIS leader if he leaves the building into an open area away from non-combatants. Should we raid the location or observe to collect more information on the ISIS network and risk his escape?

Over several months, allied forces tracked the pattern of life of visitors and residents. This information helped anticipate the best time, weather, and method to kill the terrorist with minimal non-combatant casualties. Based on this information, the helicopter commando raid launched on 2 Feb 22. During the attack, Abdullah detonated explosives that killed himself, his family, and other residents whom the special operators could not evacuate in time.

The mission was a success, including its information effects. While President Biden’s announcement of the raid emphasized no US casualties and another ISIS leader’s death, he included a critical contrast. US forces risked the air assault raid rather than an airstrike to save civilians versus Abdullah’s last act of cowardice and murder. First responders immediately reported civilian casualties, which some news sources attributed to US forces.

The US won another kinetic battle, but the fight to gain and maintain the initiative in the information environment does not end. ISIS and other terrorist organizations use suicide as a recruitment tool, labeling Abdullah a martyr per their extremist ideology.

The operation was a supply chain of sensors (sensor-shooters), infrastructure, and wireless communication. The key ingredients were the highly trained experts who understood the entire mission and trusted one another. Sensors included remotely operated aerial and space platforms and specialized ground forces for detection, tracking, communication, and delivery. Robust transportation, intelligence, and command and control infrastructure was in place. The sensors and infrastructure provided real-time visibility of system conditions, threats, and vulnerabilities that could cause deviations. Every participant was prepared to execute a precise yet flexible plan to achieve the mission’s output—the capture or killing of the ISIS leader. That accomplishment, however, was not the end of mission fulfillment. The operation’s information effects had just begun to be contested.

Conclusions

This two-part series, Paper #48 and Paper #49, proposed and applied a new joint military doctrinal definition of information for operating to win significant advantages in the information environment.

• “Information” is an operational process in an environment filled with physical and psychological matter, energy, data, and interactions.
• We use baselines and linkage, pattern and trend, anomaly, and anticipatory analysis to influence the information environment with cycles of inputs, changes, and outputs.

The definition above allows us to disprove cause and effect propositions, not simply confirming what we believe to be accurate. We need to falsify what we think and know because innovative or ruthless competitors, including inexplicable AI, will do that at our expense.

The sample of advanced analysis techniques above applies the competitive, dynamic concept of information.

Leaders must transform legacy practices and identities to integrate operations and information in ways that win information wars not just kinetic battles.

Applying this approach to supply chains helps clarify how to win advantages in the IE, from development to marketing to procurement, fulfillment, and customer service.

The critical baselines are cultural, technical, and functional, and from those, sources of competitive advantage.

• Cultural: vital information on patterns and behaviors deriving from traditions, beliefs, attitudes, identities, and social norms
• Technical: the structure of interconnected technologies to compete, survive, and advance, technology flow to reveal advances and gaps, and the duration and timing of technology
• Functional: examples include demand for a product or service, economic performance, government structure and informal power, and all levels of media
• Sources of competitive advantage: awareness of cultural practices and beliefs, acquisition and application of technology, and socially contextual functions

Informed by the baselines, advanced analysis characterizes the information environment where supply chains compete for sources, manufacturing, and delivery:

• Linkage analysis: mapping the nodes and relationships in the IE can simplify the complexity by structuring it, revealing otherwise hidden relationships
• Pattern and Trend analysis: discerning consistent or repeating traits, acts, or any observable information (pattern) over the long-term (trend); linear or curvilinear
• Anomaly analysis: detecting unanticipated or unexpected values or phenomena, departures from a baseline, pattern, conditions, or trend
• Anticipatory analysis: informed by the preceding analyses, this forecasting process is a dialectic of inputs, changes, and outputs to out-think and out-execute one’s competitor

In the information environment, competitive supply chains are critical to securing economic advantages that sustain technological advances and drive global political change over the long term. Winning advantage requires proactive analysis for exploiting opportunities. To survive, operators need to respond to change faster and better than competitors and predators. To thrive, leaders need to shape the conditions for desired change better than the competition. We need to recognize information as an operational process that must be mastered.

Automated analytics and AI-assisted decision-making can help us optimize efficiency and value. However, AI already develops inexplicable solutions we do not understand, but accept. Advanced information environment analysis is key to maintaining a holistic perspective and asking the right strategic questions to compete and stay in control.

[i] Overtraining an artificial intelligence platform occurs when the training data is so narrow that the AI cannot make sense of new data. Neural networks such as Alpha Zero train themselves by self-play alone. Stuart Russell and Peter Norvig, Artificial Intelligence: A Modern Approach, 4^th ed, p. 164.

[ii] John H. Miller, A Crude Look at the Whole: The Science of Complex Systems in Business, Life, and Society, Basic Books, 2015, p. 194.

[iii] Joint Publication 3-13, Information Operations, 27 Nov 12, Change 1 Nov 14, p. GL-3. https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_13.pdf.

[iv] Joint Concept for Operating in the Information Environment, Joint Chiefs of Staff, 25 Jul 18, p. 42, https://www.jcs.mil/Portals/36/Documents/Doctrine/concepts/joint_concepts_jcoie.pdf.

[v] Ibid, p. 2.

[vi] There is a philosophical difference between human-centered and data-centered that influences how we characterize the information environment. Anti-realist holism assumes that the human mind must act for reality to exist, so humans are the exclusive subject of analysis. Realist holism considers individuals as just one type of actor in a larger whole. These differences influence the propositions we decide to test for causes and effects. For an example of realist holism applied to complex adaptive systems, see John H. Miller, A Crude Look at the Whole: The Science of Complex Systems in Business, Life, and Society, Basic Books, 2015, p. 194.

[vii] JP 3-13, p. GL-3.

[viii] Ibid, p. I-6.

[ix] Thanks to the work of Larry Bruns, the author of JMark’s practical exercises that run the duration of the course.

[x] As related by a former student of Professor Hoffman–Lt Gen ret. Ervin J. Rokke. These questions are a distillation of the structure of the world in terms of basic actors and relative power among them, forces of change operating on the actors, relationships between domestic and foreign policy, and interactions among the actors. See Stanley H. Hoffman, “International Relations: The Long Road to Theory,” World Politics, April 1959 (346-377), pp 371-374.

[xi] This term is taken from Clifford Geertz’s classic work, Thick Description, Toward an Interpretative Theory of Culture, Basic Books, 1973.

[xii] For an extensive explanation, see the four dimensions of “technology space” in Project Socrates as described in Ervin Ackman, President Reagan’s Program to Secure U.S. Leadership Indefinitely: How All Americans Can Participate and Reap the Benefits, Ervin Ackman, loc 1346-1411.

[xiii] Seeker has an “Insights” feature that I successively activated for specific terms in the visual depiction. This sequence enables us to examine fourth, fifth and sixth orders of relationships beyond the initial 3-D picture.

[xiv] Sang-Yeon Kim, Mike Allen, & Raymond Preiss, “Meta-analysis of the Curvilinear Relationships between Rate of Delivery and Message Persuasiveness,” Communication, Society and Media, Vol. 2 No. 9, 2019, p. 9, file:///Users/thomasdrohan/Downloads/Meta-Analysis_of_the_Curvilinear_Relationship_betw.pdf.

[xv] Tony Xing and Neta Haiby, Supply Chain Anomaly Detection and Root Cause Analysis with Azure Metric Advisor, Microsoft Azure AI Blog, https://techcommunity.microsoft.com/t5/azure-ai-blog/supply-chain-anomaly-detection-and-root-cause-analysis-with/ba-p/2871920.

[xvi] Colin Campbell, “Buyer Beware: Avoid these 3 semiconductor procurement pitfalls, Supply Chain Dive, 10 Feb 22, https://www.supplychaindive.com/news/semiconductor-supply-chain-strategies-fraud/618624/?utm_content=ad-EDIT_NOTE&utm_term=39442&utm_source=Sailthru&utm_medium=email&utm_campaign=Issue:%202022-02-10%20Supply%20Chain%20Dive:%20Procurement%20%5Bissue:39724%5D.

[xvii] studioID for Tive, “Critical & Time-Sensitive Cold Chain Shipments Best Practices,” Supply Chain Drive, https://web-us11.mxradon.com/l/DownloadFile.aspx?oid=27153&&eid=9a52c28f-4082-11ec-b63c-0e9d901a4035&&etype=lp.

[xviii] Jacoby Warrick, Dan Lamothe, Matt Viser and Karoun Demirjian, “With watchers on the ground and spy drones overhead, U.S. zeroed in on Islamic State leader’s hideout,” The Washington Post, 10 Feb 22.