INFOGRAPHIC: A BEGINNER’S GUIDE TO MACHINE LEARNING ALGORITHMS

We hear the term “machine learning” a lot these days (usually in the context of predictive analysis and artificial intelligence), but machine learning has actually been a field of its own for several decades. Only recently have we been able to really take advantage of machine learning on a broad scale thanks to modern advancements in computing power. But how does machine learning actually work? The answer is simple: algorithms.   

Machine learning is a type of artificial intelligence (AI) where computers can essentially learn concepts on their own without being programmed. These are computer programmes that alter their “thinking” (or output) once exposed to new data. In order for machine learning to take place, algorithms are needed. Algorithms are put into the computer and give it rules to follow when dissecting data.

Machine learning algorithms are often used in predictive analysis. In business, predictive analysis can be used to tell the business what is most likely to happen in the future. For example, with predictive algorithms, an online T-shirt retailer can use present-day data to predict how many T-shirts they will sell next month.  

REGRESSION OR CLASSIFICATION

While machine learning algorithms can be used for other purposes, we are going to focus on prediction in this guide. Prediction is a process where output variables can be estimated based on input variables. For example, if we input characteristics of a certain house, we can predict the sale price.

Prediction problems are divided into two main categories:

• Regression Problems: The variable we are trying to predict is numerical (e.g., the price of a house)
• Classification Problems: The variable we are trying to predict is a “Yes/No” answer (e.g., whether a certain piece of equipment will experience a mechanical failure)

Now that we’ve covered what machine learning can do in terms of predictions, we can discuss the machine learning algorithms, which come in three groups: linear models, tree-based models, and neural networks.

WHAT ARE LINEAR MODEL ALGORITHMS

A linear model uses a simple formula to find a “best fit” line through a set of data points. You find the variable you want to predict (for example, how long it will take to bake a cake) through an equation of variables you know (for example, the ingredients). In order to find the prediction, we input the variables we know to get our answer. In other words, to find how long it will take for the cake to bake, we simply input the ingredients.

For example, to bake our cake, the analysis gives us this equation: t = 0.5x + 0.25y, where t = the time it takes the bake the cake, x = the weight of the cake batter, and y = 1 for chocolate cake and 0 for non-chocolate cake. So let’s say we have 1kg of cake batter and we want a chocolate cake, we input our numbers to form this equation: t = 0.5(1) + (0.25)(1) = 0.75 or 45 minutes.

There are different forms of linear model algorithms, and we’re going to discuss linear regression and logistic regression.

LINEAR REGRESSION

Linear regression, also known as “least squares regression,” is the most standard form of linear model. For regression problems (the variable we are trying to predict is numerical), linear regression is the simplest linear model.

LOGISTIC REGRESSION

Logistic regression is simply the adaptation of linear regression to classification problems (the variable we are trying to predict is a “Yes/No” answer). Logistic regression is very good for classification problems because of its shape.

DRAWBACKS OF LINEAR REGRESSION AND LOGISTIC REGRESSION

Both linear regression and logistic regression have the same drawbacks. Both have the tendency to “overfit,” which means the model adapts too exactly to the data at the expense of the ability to generalise to previously unseen data. Because of that, both models are often “regularised,” which means they have certain penalties to prevent overfit. Another drawback of linear models is that, since they’re so simple, they tend to have trouble predicting more complex behaviours.

WHAT ARE TREE-BASED MODELS

Tree-based models help explore a data set and visualise decision rules for prediction. When you hear about tree-based models, visualise decision trees or a sequence of branching operations. Tree-based models are highly accurate, stable, and are easier to interpret. As opposed to linear models, they can map non-linear relationships to problem solve.

DECISION TREE

A decision tree is a graph that uses the branching method to show each possible outcome of a decision. For example, if you want to order a salad that includes lettuce, toppings, and dressing, a decision tree can map all the possible outcomes (or varieties of salads you could end up with).

To create or train a decision tree, we take the data that we used to train the model and find which attributes best split the train set with regards to the target.

For example, a decision tree can be used in credit card fraud detection. We would find the attribute that best predicts the risk of fraud is the purchase amount (for example that someone with the credit card has made a very large purchase). This could be the first split (or branching off) – those cards that have unusually high purchases and those that do not. Then we use the second best attribute (for example, that the credit card is often used) to create the next split. We can then continue on until we have enough attributes to satisfy our needs.

RANDOM FOREST

A random forest is the average of many decision trees, each of which is trained with a random sample of the data. Each single tree in the forest is weaker than a full decision tree, but by putting them all together, we get better overall performance thanks to diversity.

Random forest is a very popular algorithm in machine learning today. It is very easy to train (or create), and it tends to perform well. Its downside is that it can be slow to output predictions relative to other algorithms, so you might not use it when you need lightning-fast predictions.

GRADIENT BOOSTING

Gradient boosting, like random forest, is also made from “weak” decision trees. The big difference is that in gradient boosting, the trees are trained one after another. Each subsequent tree is trained primarily with data that had been incorrectly identified by previous trees. This allows gradient boost to focus less on the easy-to-predict cases and more on difficult cases.

Gradient boosting is also pretty fast to train and performs very well. However, small changes in the training data set can create radical changes in the model, so it may not produce the most explainable results.

WHAT ARE NEURAL NETWORKS

Neural networks in biology are interconnected neurons that exchange messages with each other. This idea has now been adapted to the world of machine learning and is called artificial neural networks (ANN). The concept of deep learning, which is a word that pops up often, is just several layers of artificial neural networks put one after the other.

ANNs are a family of models that are taught to adopt cognitive skills to function like the human brain. No other algorithms can handle extremely complex tasks, such as image recognition, as well as neural networks can. However, just like the human brain, it takes a very long time to train the model, and it requires a lot of power (just think about how much we eat to keep our brains working).

Companies That Learn Together, Earn Together

Roch Parayre

You and your customers may work in different sectors of an industry, operate within different organizational structures, or be located in different geographies, but you have one very important thing in common: you both need to learn in order to adapt and succeed. So why not learn alongside your customers and strengthen your relationships in the process? That’s the idea behind a trend that can best be described as customer-centric co-learning, where organizations and their customers come together to learn side by side in an executive education–type setting.

General Electric is already famous for the leadership development programs for company managers that it offers at its leadership center in Crotonville, New York. It is also one of the best examples of a company that has embraced—and benefitted from—a customer-centric co-learning model. Over the past few years, I’ve had the privilege of teaching at more than half a dozen of GE’s co-learning events around the globe, and I can tell you firsthand that the inherent value of co-learning with customers is immense. Not unlike the benefits of co-creation, which show that developing beta products with active users works much better than traditional iterative improvement techniques, co-learning offers the opportunity to work collaboratively with key influencers to strengthen business relationships in ways that can’t necessarily be achieved otherwise.

In GE’s case, here’s how it works:

The company identifies its top customers and partners and invites them to send a small cohort of executives to participate in the learning experience. Most events I’ve been involved with are capped at about 120 participants (10 senior executives from 12 customer companies), which is a small enough group to remain intimate, while also offering diverse perspectives and viewpoints.

The course, which typically runs for one week, is designed to focus on one area or theme that is particularly relevant for the specific audience. The company assembles a team of presenters—educators and thought leaders who specialize in that area—while tapping GE representatives to moderate and facilitate group work. During the course, teams learn how to apply relevant concepts within their own organizations. At face value, this setup may very well look like any other executive education event—but it’s not.

Here’s why it works. Customer-centric co-learning enables organizations to do the following:

• Get closer to their customers. The opportunity to spend a week, or even a few days, interacting with customers outside of the normal course of business doesn’t come around often. But when it does, it enables both parties to learn things about the other they may not otherwise know—things that may not be directly connected to their business dealings at the moment, but that lead to a deeper understanding and enable companies to better serve their customers in the short and long term. It is also worth noting that greater trust and collegiality are by-products of this sort of transparency.

• Break free from sales tensions.These engagements are about learning and relationship building—not making another sale (although that may happen as a longer-term result). Sales conversations are prohibited during these programs, unless initiated by the customer, freeing participants from their everyday vendor–customer mode and encouraging customers to share freely without needing to fend off sales pitches.

• Encourage outside-in thinking.When customers and partners from various industries come together, everyone benefits from different viewpoints. Too often leaders are limited by their own perspectives, but a diverse group of peers can help one another see the bigger picture and adopt a more macro lens that can help them on a strategic level. 

• Demonstrate their investment in customer success.Including customers in such a high-caliber learning event clearly demonstrates that an organization values its partnerships. It sends the message that the organization is vested, and invested, in its customers’ success. It’s not unusual for companies to commit substantial resources to make this all happen, and while there is no guaranteed hard return on investment, it pays dividends in building long-term business relationships.

For all of these reasons, I’m convinced that companies that learn together, earn together. While GE is not alone in its co-learning endeavors, most organizations have yet to recognize its value, let alone take the plunge. But those willing to try will reap the rewards.

Using ORG-MASTER for knowledge based organizational change

D. Kudryavtsev, L. Grigoriev, V. Kislova, A. Zablotsky

Abstract: Enterprises in growing markets with transitional economy nowadays encounter extreme necessity to change their structures and improve business processes. In order to support knowledge processes within organizational change initiative  enterprises can use business modeling tools. On one hand software vendors suggest many tools of this kind, but on the other hand growing markets with transitional economy determine quite special requirements for such tools. This article reveals these requirements, assess existing business modeling tools using these requirements and describes ORG-Master as a tool specially created  for support of process improvement initiatives in the growing markets with transitional economy. 

Keywords: Business information modeling, business modeling, knowledge process, organizational change, business process improvement, growing markets, transitional economy.

Introduction

ORG-Master is a business modeling software, which was initially created as a response to growing need for computer aid to consulting projects in the field of  organizational change and business transformation. In spite of the diversity of products for business modeling ORG-Master has certain advantages that can be revealed in solving certain tasks in certain environment.

Certain tasks include such organizational change components as business process improvement, business restructuring, quality management implementation and holistic improvement of management system. In the current article organizational development will be described by the example of business process improvement (BPI) initiative. 

Certain environment include growing markets with transitional economy (GMwTE) which determine specialties in organizational change initiatives. GMwTE include post-soviet countries (Russia, Ukraine, Belarus, Kazakhstan) and in the current article will be described by the example of Russia. In order to reveal these specialties Section 1 describes features of GMwTE from management point of view. Section 2 focuses on the flow of knowledge within BPI initiative and gives an ability to define requirements for business modeling tool at the GMwTE (section 3). Section 4 reveal imperfections of existing business modeling tools with respect to above-mentioned requirements and show the niche for ORG-Master. Section 5 explain the main concepts and consequent advantages of ORG-Master. Section 6 describes practical application of ORG-Master.

1. Business process improvement initiatives in the growing markets with transitional economy

The most important features of GMwTE from management point of view are:
1. Extremely high pace of change in market conditions and business environment
2. Low level of managerial culture
3. Predominance of informal methods of management

Quick changes and competition growth make companies to change in the same pace and the main objectives in the organizational change is to fit company structure with business needs and to implement client-oriented business processes that allow to achieve company goals. This results in the necessity to launch restructuring or BPI initiatives.

The main prerequisite for BPI initiative is transparent management at every level of organization. In this context transparency implies holistic knowledge describing What functions and processes are realized in the company, Who performs the functions, How the functions are performed, What for are the functions performed. While low level of managerial culture results in absence of clear knowledge in this field. As a results BPI initiative in the GMwTE usually involve a wide range of preliminary stages directed towards understanding of company “big picture” in order to make conceptual changes and define the processes for improvement or re-engineering.

The third feature of GMwTE — predominance of informal methods of management results in small amount of documents and business rules. Such a situation has its roots either in skeptical attitude to archaic and out-of-date formal documents at post-soviet enterprises or in quick growth of small start-ups. In some situations informal intuitive method of management brings fruits, but it is terminated by the scale of business and is one of the barriers in development of managerial culture. As a result BPI initiative in the GMwTE has an important objective – to  switch company from informal methods of management to formal procedures and business rules.

2. Knowledge process in the business process improvement initiative

BPI or restructuring initiatives deal with business organization knowledge.  Under business organization knowledge in the current article we will understand knowledge domain covering organizational goals, structure, processes, functions, rules, rights, authorities and relationships between this objects. Thus in order  to raise effectiveness of BPI initiative project team should support knowledge process in the domain of business organization knowledge. As described in [Strohmaier, 03a] knowledge infrastructure   is determined by the nature of knowledge process, which in turn can be understood through analysis of business processes covered by improvement initiative (figure 1).  

Figure 1: Business process and knowledge infrastructure relationship

The main constituents of improvement initiatives are organizational change processes - a subset of the whole system of business processes: 

• business process analysis
• business process improvement
• organizational structure control
• performance management

Business analysts (either internal, or external consultants) together with domain experts (head of departments and other managers) generate business organization knowledge, store and transfer it throughout the organization in these processes.

Application of business organization knowledge is distributed between all the other business processes — operating, management and support processes. Organizational roles of performers vary from workers to executives (top managers).
According to [Strohmaier, 03b] business organization knowledge can be visualized (figure 2).

Figure 2: Knowledge process in the business improvement initiative

The most important and influential feature of this process consists in different organizational roles involved in it and especially in the knowledge transfer process. During transfer process business analysts deliver their knowledge through the mediation of domain experts to personnel from different domains and organizational levels. As it was mentioned in [Section 1], one of the goals of BPI initiative in the GMwTE is to shift the priorities of management from informal methods to formal business rules. Thus the basis of knowledge transfer is formalized knowledge and the main factor of its successful internalization [see Nonaka, 03] by personnel is type of knowledge representation.

Type of knowledge representation depends on two specific knowledge processes generation on one hand and application on the other. While the way these processes are performed is determined by the involved organizational roles.
Business analysts have developed competencies in organizational management and system analysis. They have detailed understanding of business from different points of view and can operate with different objects and their relations (organizational units, functions, processes, goals etc).

Personnel from different domains and organizational levels have low competencies in organizational management see [Section 1]. They usually have only dim understanding of business from “organizational unit” point of view (answer for the question “who do what?”).
As a result business analysts primary use diagrams of different types and notations (IDEF0, UML, EPC) as a mean of knowledge representation. The other personnel use job descriptions, documents describing the functions of business units / departments and other regulating documents. In addition these regulating document for application usually prepared according to national, industrial or corporate standard. Namely these regulating documents solve one of the objectives of BPI initiative in the GMwTE - shift the priorities of management from informal methods to formal business rules see [Section 1].

Thus the necessity to facilitate communication between people speaking “different languages” predetermines important requirement for knowledge infrastructure.

3. Requirements to business modeling tool for GMwTE

For the current analysis we assume business modeling tool primary as a support system for knowledge generation and storage during BPI initiative.
Previous section described the necessity to have different types of knowledge representation during BPI initiative in the GMwTE. Assumption that the process of knowledge transfer do not change the type of knowledge representation imply the necessity to generate knowledge both in type of diagrams for analysis and regulating documents for application. This requirement for knowledge generation process determine the first requirement for business modeling tool:

1. Ability to represent knowledge in different types and formats.
Section 1 highlight the necessity of preliminary stages within BPI initiative in the GMwTE. For example, companies should define goals, composition of functions, change organizational structure, reassign responsibilities for function realization, reveal a list of business processes. This tasks can be done both in series and in parallel and include several analysts concentrating either on different tasks or on different levels of detail. Such a nature of BPI initiative determine the next requirement:

2. Ability to work both with a complex model (e.g. business process model) and with separate parts of this model (relate functions with organization roles, roles with infrastructure etc) using different views of enterprise.
Fast dynamic of the enterprise development is especially relevant for GMwTE and require constant improvements in business processes thus a model once created should be constantly up-dated. Model is a system of constituent objects and their relationships, but both objects and their relations changes constantly. This situation generate the third requirement:

3. Ability to reflect changes in objects and in their relationships throughout the whole model after changing any part of the model. In order to reveal a tool which satisfy all the abovementioned requirements an analysis of  the tools existent in the Russian market was carried out.

4. Analysis of existing business modeling tools in the Russian market

Although in some BPI initiatives knowledge is created and stored using typical office applications like MS Word or Excel or simple graphical packages like MS Visio this tools obviously do not satisfy requirements see [Section 3].
The main business modeling tools existent in the Russian market that are usually used for organizational development and BPI purposes are:
ARIS http://www.ids-scheer.com/
BPWin (AllFusion Modeling Suite) http://ca.com/

There are also some Russian products that contain either limited functionality or slight modifications of foregoing tools. Differences of these products are immaterial from chosen requirements point of view and as a result they appeared beyond the scope of our analysis.

There are also a broad range of CASE tools (e.g. Rational Rose) for corporate systems development. These tools include business process modeling, but their primary function is information architecture development and it determine their whole viewpoint for enterprise modeling. As a result they are nor convenient for organizational management and business process modeling, nor efficient. Thus they appeared beyond the scope of our analysis.

Here is generalized result of the analysis:

Requirement 1: Ability to represent knowledge in different types and formats.
ARIS: It include a broad library of object types and corresponding diagrams, but it has a very complicated mechanism for generating regulating documents. It is hard to customize necessary templates and consequently requires unique and expensive specialists
BPWin: It allow to generate IDEF0 diagrams, but it is also very hard to generate corresponding regulating documents in customary standards.

Requirement 2: Ability to work both with a complex model (e.g. business process model) and with separate parts of this model
ARIS: Satisfy. There are both a whole process model and separate constituent models.
BPWin: Dissatisfy. User works either with one object type (functions, roles) or with a whole model of business process (one type of composite diagram).

Requirement 3: Ability to reflect changes in objects and in their relationships throughout the whole model after changing any part of the model.
ARIS: Partially. Centralized library of modeling objects guarantee the reflection of changes in the particular object throughout the model (e.g. changing function name in one diagram cause changing this name in every diagram in the model), but changes in relationships between objects of different type do not appear automatically throughout all diagrams.
BPWin: Satisfy. All the objects stored in centralized library and are used in one type of diagram.

Thus presented tools do not completely satisfy suggested requirements. Besides this tools are quite expensive and requires extremely professional analytics to support business model.
There is a necessity for more effective business modeling tool for organizational development.

5. Main concepts and advantages of ORG-Master

Concepts and methodology

The main idea of ORG-Master consists in division of business modeling interface from model representation one. As a result each interface and type of knowledge representation is optimized for the solution of own tasks. This idea is contrary to an approach of ARIS and BPWin. In the foregoing products user input, editing and represent business model in the same knowledge representation type and format.

Division of interfaces in ORG-Master allow to represent knowledge both in different types (diagrams in different notations, reports, tables) and from different point of views.
On the other hand business model editing interface has its own type of knowledge representation based on two instruments: classifier (ontological models, see [Gavrilova, 00]) and matrix (table).

Classifier – hierarchical tree of particular objects (e.g. organizational roles, functions, material resources, documents etc), which can have different attributes: type, meaning, comments etc. In the process of building classifier objects become structured into a hierarchy/ tree – they receives relationships of AKO (“A Kind Of” [Gavrilova, 00]) type (figure 3).

Figure 3: Structuring informational objects

Matrix (table) – models that define relationships between objects of different classifiers in any combination of the later (figure 4). Relationships can also have different attributes (directions, type, name, index, meaning). 

А) Matrix as a relationship of objects from 2 classifiers	В) Tabular representation of a matrix of 2 classifiers	С) «Triple matrix»
Figure 4: Conceptual framework of matrix (table)

As any material object of any complexity (e.g. building) can be described using definite number of  2-dimensional (flat) schemes (e.g. design drawings) so and several matrix allow to receive multidimensional description of complex business system and make it both holistic and visible (see figure 5).  

Figure 5: Business process model as a system of classifiers and matrixes

ORG-Master advantages for end user

Foregoing concepts of ORG-Master provide the following features of this tool:

• ability to generate multidimensional reports based on matrixes from business model with different level of detail, which allow to analyze company from many viewpoints for people at different levels of organizational hierarchy
• ability to fine-tune knowledge representation reports, that allow to generate regulating reporting on the basis of business model for particular needs and customary standards
• ability to generate visual diagrams of business processes that support business analysis
• ability to decompose the whole business model into constituent separate submodels, which allow to divide complex BPI initiative into manageable tasks and solve problems in separate domains with adequate (pared-down) tool
• all the objects and relationships from different submodels are integrated into centralized holistic model that allow to reflect changes in objects and in their relationships throughout the whole model after changing any part of the model

These features of ORG-Master satisfy requirements to business modeling tool for GMwTE and together with relatively low price and low training complexity characterize it as effective and efficient tool.

Among the relative disadvantages of this tool the most obvious is absence of quantitative analysis of business processes, but this feature is of low importance with respect to foregoing requirements.

6. Application of ORG-Master and practical results

ORG-Master has 6-year history of  application in the organizational change and BPI initiatives. There is a broad range of ORG-Master clients located in Russia, Ukraine. Size of ORG-Master clients vary from small companies to large holding structures (up to 10000 people).

The results of typical ORG-Master application in BPI initiatives include:

• Business model which describe functions, organizational roles, goals and measures,  functions distribution among organizational roles and description of necessary business processes.
• Regulating documents based on business model (job descriptions, procedures etc)
• Diagrams of the necessary processes based on business model

Conclusions

Since organizational change and BPI initiatives become a life-style of every company it is useful to support such an activity with adequate tools for business modeling. But choice of  the tool is determined by the objectives of tool application and business environment. Current paper revealed specialties of BPI initiatives in the GMwTE and analyzed existing business modeling tools from that perspective. As a result of this analysis ORG-Master can be considered as an effective and efficient for knowledge process support during organizational change initiatives in the GMwTE.

Acknowledgements

The author of this paper is thankful to the advisor Dr. Prof. Tatiana Gavrilova (St. Petersburg State Polytechnic University) for her useful suggestions about content and destiny of this paper.

Bibliography

[Strohmaier, 03a] M. Strohmaier Designing Business Process Oriented Knowledge Infrastructures Proceedings der GI Workshopwoche, Workshop der Fachgruppe Wissensmanagement, Karlsruhe (2003)
[Strohmaier, 03b] M. Strohmaier A Business Process oriented Approach for the Identification and Support of organizational Knowledge Processes Proceedings der 4. Oldenburger Fachtagung Wissensmanagement, Oldenburg (2003)
[BIG, 96] BIG&Expert “Seven notes of management”, Moscow (1996)
[Nonaka, 95] Nonaka I., Takeuchi H.: “The knowledge creating company”; Oxford University Press (1995)
[Bukowitz, 99] Bukowitz W., Williams R.: “The knowledge management fieldbook”; Prentice hall, Pearson Education Limited (1999)
[APQC, 96] APQC’s International Benchmarking Clearinghouse Process Classification Framework www.apqc.org, (1996)
[Gorelik, 01] Gorelik S., “Business-engineering and management of organizational change”; (2001) http://www.bigс.ru/publications/bigspb/metodology/
[Gavrilova, 00] Gavrilova T., Horoschevsky V. “Knowledge bases of intellectual systems”; Piter / Saint-Petersburg (2000)
[Рубцов, 99] Рубцов С., Сравнительный анализ и выбор средств инструментальной поддержки организационного проектирования и реинжиниринга бизнес процессов http://or-rsv.narod.ru/Articles/Aris-IDEF.htm
[Репин, 01] Репин В. Сравнительный анализ нотаций. http://www.interface.ru/fset.asp?Url=/ca/an/danaris1.htm

Dimensions of Knowledge Management

One of our current efforts at a-connect is to improve our knowledge management capabilities and processes. The framework below from PA Consulting highlights some of the key elements:

In each dimension, the capabilities will vary along a spectrum, from very basic tasks to quite sophisticated skills. On the technology side, for example, one company may have the basics in place through a shared drive with key documents and basic search features. Another, more sophisticated company may have a full-fledged knowledge management system, a portal with internal and external access, e-learning capabilities, etc.

A diagnostic of the six dimensions will help companies understand where they are and what actions are critical to take knowledge management to the next level.

It struck me that the framework is not only applicable to knowledge management systems, but also to a variety of other topics that require the implementation of an IT solution: Anybody who has ever wrestled with the implementation of a CRM system will recognize the key issues!

Staying in the Know

Davide Nicolini, Maja Korica and Keith Ruddle

In an era of information overload, getting the right information remains a challenge for time-pressed executives. Is it time to overhaul your personal knowledge infrastructure?

A common thread runs through many recent corporate setbacks and scandals. In crises ranging from BP’s Deepwater Horizon oil spill debacle to the Libor rate-fixing scandal in the City of London, the troubles simmered below the CEO’s radar. By the time the problems were revealed, most of the damage had arguably already been done. Despite indications that large companies are becoming increasingly complicated to manage,1 executives are still responsible for staying abreast of what’s going in their organization. But how do you keep tabs on what your competitors and employees are doing? How do you spot the next big idea and make the best judgments? How do you distinguish usable information from distracting noise? And how do you maintain focus on what’s critical?

Many management experts have assumed that better information systems and more data would solve the problem. Some have pushed for faster and more powerful information technologies. Others have put their faith in better dashboards, big data and social networking. But is better technology or more tools really the most promising way forward? We think not. In this article, we maintain that the capacity of senior executives to remain appropriately and effectively knowledgeable in order to perform their jobs is based on a personal and organizational capability to continually “stay in the know” by assembling and maintaining what we call a “personal knowledge infrastructure.” And while information technologies may be part of this personal knowledge infrastructure, they are really just one of the components.

We are not the first researchers to make this claim. More than 40 years ago, organizational theorist Henry Mintzberg suggested that information was central to managerial work and that the most important managerial roles revolved around information (monitoring, disseminating and acting as a spokesperson). Mintzberg described managers as the nerve centers of organizations and said informational activities “tie all managerial work together.”2 Other researchers suggested that management itself could be considered a form of information gathering and that we are quickly moving from an information society to an attention economy, where competitive advantage comes not from acquiring more information but from knowing what to pay attention to.3 Later research confirmed that dealing with information is critical and found that managers’ communication abilities are directly related to their performance.4

While the importance of informational roles and activities is well established, we take the idea a step further, arguing that managers — and especially senior executives — are only as good at acquiring and interpreting critical information as their personal knowledge infrastructures are. Managers rely on specific learned modes to manage and allocate their attention.5 However, how we pay attention is not simply a matter of internal mental processes that we can do little about. Rather, attentiveness (in other words, the capacity to stay on top, and the ability to distinguish between what matters and what doesn’t) mostly stems from what managers do or don’t do, whom they talk to and when, and what tools and tricks of the trade they use. In short, attentiveness relies on and is facilitated by things we can observe — and things we can do something about.

Technologies and new tools are not and cannot be “silver bullet” solutions. At times, simpler things such as talking to customers or networking with board members may be more important, provided they are done methodically and with some purpose. Selecting when particular elements are appropriate depends on the circumstances. As a result, understanding and, when needed, overhauling one’s personal knowledge infrastructure should be routine. In this article, we explain how this can be done, drawing on insights obtained by shadowing individual CEOs as they went about their daily jobs.6

About the Research

To uncover how top executives deal with information and knowledge, we conducted an observation-based exploratory study using a rigorous ethnographic protocol successfully employed in the past. We followed seven chief executives through their working days for several weeks. We went where they went, watched what they did, listened to what they and others said and asked what was going on when we did not understand. We also discussed our findings with them and with invited colleagues as part of structured feedback sessions.

Our sample comprised seven CEOs of acute and mental health organizations that are part of the National Health Service in England. In England, health care is provided by public sector bodies called trusts. Our sample included organizations that run multiple hospitals, have an annual budget of more than 500 million pounds and have up to 10,000 employees. The CEOs have both legal and financial responsibility. The sample included both men and women (3:4). The CEOs had diverse professional backgrounds (NHS management, private sector, nursing and medical) and were at different points in their careers, both in terms of tenure in their present post and overall experience at the CEO level. The sample also included organizations with different performance levels according to indicators by which their performance was monitored by national regulators (for example, financially sound vs. struggling).

CEOs were observed for five or more weeks, apart from one subject, where observations lasted 3½ weeks. The researchers had good access to the CEOs and were able to document nearly all aspects of their work, with exceptions such as one-to-one supervisory meetings with junior colleagues, HR-related meetings concerning individuals and private meetings with patients. When CEOs worked from home, data was collected in interviews afterwards. We conducted semi-structured interviews with five of the CEOs and a number of informal interviews with the other two CEOs. In addition, we conducted two formal interviews with two different personal assistants, which were recorded and lasted approximately half an hour each. Additional data came from meeting papers, articles referenced by the CEOs and copies of publications consulted, and they were supplemented by externally available information such as annual trust reports and regulatory documents. Following the study, our results were shared with two groups of CEOs at dedicated sector events. Those CEOs helped us to refine our findings and elaborate on the notion of the personal knowledge infrastructure.

Our research is based on a two-year study of the day-to-day work of seven CEOs of some of the largest and most challenging hospital- and mental health-based organizations in England. (See “About the Research.”) We chose to study health-care executives because they sit at the crossroads between the private and public sectors and therefore are expected to meet multiple, often competing, demands. To say that the informational landscapes in these organizations are complex is an understatement. Yet the organizations are increasingly subject to pressures to become more transparent, even as they compete with each other. Therefore they seemed to be a good choice as settings for studying the challenges of using information and knowledge to stay on top and ahead of the curve. Throughout our research we sought to answer a simple question: How did the CEOs know what they needed to know in order to be effective at their jobs?

“Nothing but Talking”

One of the first things that struck us was that, in contrast to the popular image of CEOs as lonely, heroic decision makers, the individuals we studied did not seek information or utilize discrete pieces of evidence for the purpose of making decisions. Rather, they often sought something much more ordinary: to make themselves knowledgeable in order to be ready for any eventuality, so that they could understand what to do next. Indeed, one of their main preoccupations appeared to be staying on top of what was happening within and around their organizations. As one put it: “The worst thing for a CEO is to find yourself asking after the fact: How could this happen without me knowing?”

Notably, staying on top was not a separate activity in addition to what the CEOs already did, but rather something they mostly did without thinking and without noticing — or something that they achieved while doing something else. Calling a former colleague who was in the news for the wrong reasons (an accident, a bad report from inspectors, protests about the closure of a loss-making hospital) can produce multiple outcomes: reinforcing a relationship and demonstrating solidarity, but also finding out what is going on. Indeed, many of the CEOs had difficulty acknowledging that checking in with people was an integral part of their job — hence the often-heard comment, “I do not know what has happened to my workday … It seems I have done nothing but talking.”

The Personal Knowledge Infrastructure

The CEOs we studied didn’t leave the process of staying informed to chance. Rather, they relied on a habitual and recurrent set of practices, relationships and tools of the trade, which constituted a personal knowledge infrastructure that supported them in their daily tasks of understanding, foreseeing and managing. This tacit and rarely discussed infrastructure, which was very different from their IT system, helped them to know what needed to be done and to get a sense of the right way forward. What made some CEOs more effective than others was not merely the characteristics of the individual components of their personal knowledge infrastructure but also the quality of the whole and its fit with the specific needs of the job. This personal knowledge infrastructure comprised three main elements: routine practices, relationships, and tools and technologies.

First, every CEO had a set of routine practices she or he relied on — things such as checking the morning news, running periodic review meetings, dropping by immediate collaborators’ offices to ask what was often “just a quick question,” walking around and occasionally even going to the cafeteria “to check how things are going.” These practices were not just internal. The CEOs also met with board members and managers of other organizations, attended conferences and staff events and participated in ceremonial functions such as charity events. Some of the gatherings were framed as leisure opportunities (having a drink, playing golf), but they weren’t entirely social. CEOs returned from such events with information and news they shared with various associates. Similarly, sitting on boards of other organizations was often seen as a necessary evil that helped the CEOs get a broader overview of what was going on beyond their organization.

Second, the personal knowledge infrastructure contained a number of social relationships. Like prior researchers, we found that most CEOs’ work was conducted verbally and was accomplished with and through others. Our CEOs used their relationships both to gather information and to make sense of it.7 For example, every CEO we studied carefully cultivated strategic relationships within and outside the organization. These relationships were usually engineered to produce a combination of breadth and depth of intelligence. Their network of contacts constituted a form of social capital that had been accumulated over time. On many occasions, we observed CEOs interacting with long-term colleagues, previous members of their staff and people with whom they had done business. They used their contacts to gather weak signals (on the principle that today’s gossip can become tomorrow’s news), triangulate information and confirm or contradict their evolving insights. Some of the CEOs were extremely strategic and nurtured relationships with various stakeholders (for example, management consultants, politicians or local leaders), whom they saw regularly for dinner or a drink. Some CEOs also relied on small groups of peers, whom they met with on a regular basis. These groups of peers, who were often also “comrades in adversity” facing similar challenges, operated both as a support group and as a precious setting where sensitive information was exchanged on the basis of reciprocity.

However, not all of the contacts were treated the same way.8 The CEOs appeared to have an informal hierarchy: those who were more distant, who could be used as a source of signals and needed to be taken with a grain of salt; those who were trusted and tended to provide reliable intelligence (for example, board members or colleagues, as well as their assistants); and finally, those with whom the various streams of information could be discussed and processed — the inner circle. All of the CEOs in our sample relied heavily on such an inner circle, usually composed of selected executive team members with whom they had the most intense interactions. The CEOs used these individuals not only to obtain information, often informally (given the open-door policy that was in operation for most CEOs), but also as sounding boards to test emerging understandings and reconcile possibly competing insights. These interactions allowed the CEOs not only to connect the dots but, more importantly, to figure out which information qualified as a dot that had to be connected further.

Finally, the CEOs’ personal knowledge infrastructures included a variety of tools of the trade. These included traditional tools such as phone, email, reports and journal articles from industry magazines, as well as less traditional sources such as Twitter, blogs and other social media. Most CEOs utilized some form of electronic reporting system or audit-based dashboard that helped them track critical performance indicators, and most consulted these tools regularly, but the sophistication of the tools varied substantially. (See “Components of a Personal Knowledge Infrastructure.”)

Components of a Personal Knowledge Infrastructure

To stay on top of things, CEOs need a combination of three elements: people who feed them information and act as sounding boards; routine activities; and technologies that allow them to track performance and pick up signals. These elements constitute a CEO’s personal knowledge infrastructure. What comes to the CEO’s attention and what stays under the radar depends on the makeup of this infrastructure.

Although most CEOs had a small pile of “will read” books in their office, they rarely had time for books or magazines during the course of a working day. Personal preferences played an important role here, more so than in the two previous categories. For example, while some CEOs relied heavily on mobile phones for calls and texts, others used email almost exclusively. Contrary to our expectations, most of the CEOs dealt personally with a range of emails — this was how their work was done. Some CEOs made very little use of written documents and required short summaries. Others wanted to have complete documentation “just in case.” Some CEOs still found comfort in printed paper; few were happy to go paperless.

One critical aspect of personal knowledge infrastructures was the extent to which individual elements were often intended to support each other. For example, CEOs who liked to run large formal meetings also invested significant time in social relationships, consulting collaborators on a one-to-one basis. The CEO of an organization that was struggling with issues of trust and hidden or misplaced information was working to triangulate soft data with hard knowledge. This often required him to follow up with people individually (for example, phoning staff members directly to corroborate information or requesting documentation from outside sources) while also working to set up formal structures currently lacking.

How the Personal Knowledge Infrastructure Evolves

Although all of the CEOs relied on a personal knowledge infrastructure, how these were composed varied greatly. We saw differences across all seven CEOs in terms of whether personal knowledge infrastructures had particular elements and also the amount of emphasis each element received. Two examples illustrate how different personal knowledge infrastructures contrasted with specific leadership styles in different situations.

CEO 1: Knowing the Details in a Struggling Organization

A newly appointed CEO was running a struggling hospital-based organization that was receiving increased regulatory attention for financial reasons. His personal knowledge infrastructure was designed to help him closely monitor his organization. He ran large, often long, weekly management meetings that provided an opportunity for all team members to examine operations and share and obtain a wealth of information. After the formal meetings, conversations continued in the executive offices. The CEO spent the bulk of his time in regular meetings with local managers of health-care organizations and funding agencies, picking up signals and providing insight into the work and progress of his organization. He also spent time working on wards and visiting and talking to staff. Moreover, he cultivated a wide network of colleagues whom he often consulted in rather informal ways and maintained external links to support his key strategic tasks. He had an open-door policy, both as a symbol of change and as a permanent invitation. He felt comfortable digging through reports and documents, and he set aside time on trains to go through what he called “the train pile of documents.” Though the CEO used the phone and the Internet, he liked to attend conferences and networking events to develop a broad view of the business environment.

CEO 2: Managing Via a Mix of the Formal and Informal

The second CEO, who had been in his position for more than five years, used a very different set of practices and tools. He worked with an established team to run an organization that prided itself on its ability to achieve operational excellence and strategic growth. Throughout the day, this CEO had a series of chats with executives, which often expanded into conversations between him and several people. Indeed, much of his working day was spent in what appeared as free-form interaction: sharing information informally, only sporadically framed by discussion about an immediate problem concerning a patient or a medical concern. This CEO rarely attended local meetings, so the other executives were an information gateway to local strategic issues for him. However, the headquarters-based, executive-team orientation was reinforced by several other structures, relationships and tools, purposefully arranged by the CEO so he could remain in the know. First, there was an executive who the CEO felt had a very different approach from the others, which gave him another voice and view to consider. The CEO also had an IT performance system, which he consulted every morning and which allowed him to identify any serious performance issues in the organization without needing to rely on reports from executives. The CEO supplemented such insights with visits to wards and other areas of the hospitals late in the evening and on weekends, which allowed him to gain informal insights from veteran staff. The internal systems were supported by national-level policy work and involvement via leadership positions in sector organizations and initiatives and networking. These allowed the CEO to both formally and informally stay in the know regarding strategic issues of potential relevance, and also to influence their direction to the benefit of his organization.

What Makes It Personal

As we have seen, different CEOs use different knowledge infrastructures that reflect both what they personally need and where their organizations are at a particular point in time. In each case, the context was particularly relevant. We saw different types of knowledge infrastructures (in other words, combinations of tools, practices and relationships) in relation to seven factors:

The CEO’s Experience

More experienced CEOs often had a more defined personal style that they carried with them when they changed jobs. Some had specific practices that they tried to reactivate in the new workplace and a network of contacts that constituted the social capital they had accrued over the years; we saw this in the case of a CEO facing an operational challenge in his new organization, when he called a former colleague for advice.

Job Tenure

The more time CEOs spent in the same organization, the more they learned, often the hard way, about sources they could trust and how they could make these sources work, given their existing infrastructures and approaches to work.

Makeup of the Executive Team and Board

The composition of the top management team, how competent its members are perceived to be and how well they work as a team affected the makeup of the inner circle. CEOs often included in their inner conversational circles directors who were easy to talk to or were particularly good at collecting and relaying intelligence. However, many CEOs, like the second CEO discussed above, also saw value in having friendly “devil’s advocates” on staff who were able to present different views and act as meaningful counterweights.

Organizational Conditions and Pressures

Organizations with different financial, efficiency, quality and safety environments posed different issues for the CEOs. When the conditions changed, they had to retune their antennae accordingly. A particular challenge involved the tools and technologies available. Systems can be powerful but are costly and difficult to change. Most CEOs worked to modify and develop existing systems but often they didn’t have a lot of room to make immediate changes. They worked with what they had while instigating long-term interventions so that the system would suit them rather than the other way around.

Strategic Vision

Entering new markets or introducing new products or services required CEOs to adapt their personal knowledge infrastructure accordingly. For instance, a CEO facing a possible merger began adding M&A events to his calendar.

Economic, Competitive and Regulatory Environment

The macro environment determined whether a CEO’s personal knowledge infrastructure was appropriate. Changes in the environment forced CEOs to adapt their existing personal knowledge infrastructure and reactivate old relationships.

The Kind of Manager the CEO Wants to Be

Ultimately, the above factors were filtered through the prism of “what kind of manager I would like to be.” For instance, a CEO who valued transparency and closeness to his organization’s users established a strong presence on social media and utilized this channel to garner insights into the experience of patients and their families. This sometimes allowed him to identify problems (such as low quality of service in a particular location or low staff morale) before his managers reported them.

Taken together, the factors above suggest that effective personal knowledge infrastructures tend to be unique and personal and conform to the preferences of the manager. They need to be continually adapted, tweaked and refined in keeping with the shifting nature of the CEO’s job, the environment and new opportunities.

Although most of our CEOs were reasonably successful, everyone saw room for improvement. Indeed, a CEO’s effectiveness was a reflection of his or her situation and person-specific alignment. For example, one CEO had spent years building a sophisticated IT performance monitoring system. Another CEO didn’t see having such a system as a priority; in his view, being an effective manager entailed moving away from operational considerations and focusing more on strategic and systemwide issues. The challenge of changing as the organization changes was highlighted by several CEOs: The personal knowledge infrastructure that serves you well during a period of crisis and turmoil may get in the way in calmer waters. The lesson is that there is no single best personal knowledge infrastructure. Through personal reflection, managers and CEOs need to learn how to ask themselves difficult questions regarding the quality and fit of the practices, tools and relationships that they rely on to become knowledgeable. They also need to develop structured ways of asking such questions consistently and over time — rather than waiting for something to go terribly wrong.

Four Traps

The quality and fit of the CEO’s personal knowledge infrastructure is critical because it determines how he or she sees the world and defines himself or herself as a manager and CEO. It is the prism through which managers understand what is going on, and it provides the horizon of information sources through which this understanding will be probed and evaluated. However, a poorly designed personal knowledge infrastructure can lock the manager inside an information bubble and create information biases and blind spots.9 Managers may only realize this when something happens that was not on their radar or when an incident exposes the misalignment between the current demands and needs of their job and their own role. By closely examining the work practices of our CEOs over time, we identified four potential traps:

1. Not Obtaining the Information You Need

Although conventional wisdom suggests that the main problem for today’s executives is too much information, the real problem is not enough relevant information. Due to insufficient monitoring, an inappropriate mix of monitoring practices, inadequate or insufficient social relationships, and information overload, managers can find themselves without the information they need.

Evaluating Your Personal Knowledge Infrastructure

Executives can use the six questions below to create personal knowledge infrastructures to match their own particular circumstances.

2. Developing a Personal Knowledge Infrastructure That Points You in the Wrong Direction

A typical problem with personal knowledge infrastructures is that they can be poorly aligned with the demands of the job. For example, if a CEO wants to foster innovation but the infrastructure informs her about operational issues only, the CEO is likely to focus on things that aren’t of primary importance. A personal knowledge infrastructure not only reflects the rules of attention but also shapes those rules. Researchers have highlighted lessons from major spectacular failures from the past, from the Challenger space shuttle disaster to the global financial crisis.10 Many of the managers in question were completely current on the wrong information — or information about the wrong things.

3. Setting Up a Personal Knowledge Infrastructure That Is Not “You”

A manager’s personal knowledge infrastructure can clash with his management style, both in terms of what he does, the tools he uses and the type of manager he would like to be. In our study, we observed a CEO who wanted to be a manager who delegated. However, his personal knowledge infrastructure systematically drove him to focus on details, which led him to take a hands-on approach — against his best intentions. The most effective managers we observed were those who reshaped their personal knowledge infrastructure to fit their work, their management style and what they considered important.

4. Starting With Technology Rather Than Personal Need

Last, some managers make the mistake of addressing the issue from the wrong end — considering technology first rather than later. Rather than being technology-centered, personal knowledge infrastructures need to be geared toward personal development, not toward buying new technologies. Rather than asking, “Is this technology good?,” CEOs should ask, “Will it do any good for me?”

Improving Your Personal Knowledge Infrastructure

So how do managers improve their personal knowledge infrastructures? We found that although CEOs easily recognize the importance of their personal knowledge infrastructure, they very rarely pause to reflect on its effectiveness and fit. More often than not, they discover its inadequacies through comparison with others’ practices or, more commonly, following breakdowns and failures. Developing, refining and testing the effectiveness or present fit of your personal knowledge infrastructure should be routine. (See “How to Improve Your Personal Knowledge Infrastructure.”)

How to Improve Your Personal Knowledge Infrastructure

For CEOs or other executives concerned about improving their personal knowledge infrastructure, we have developed six steps designed to initiate learning and reflection. Examine the following:

1. Take stock of what you do now — what tools, practices and relationships do you use in day-to-day work?
2. Perform an initial personal assessment of whether your personal knowledge infrastructure is currently fit for the purpose intended, using the questions we provide. (See “Evaluating Your Personal Knowledge Infrastructure.”)
3. Ask a peer, mentor or coach to observe you and help you observe yourself and reflect.
4. With a trusted peer group, discuss how each person in the group operates and what differences you deploy within your different personal knowledge infrastructures.
5. Develop a personal plan to rebuild and enhance your personal knowledge infrastructure over time.
6. Put the plan into place and set a goal of continuous learning, committing to specific dates (for example, bimonthly peer group meetings).
7. There is a great deal that individuals can do for themselves. The starting point is being aware of the composition and functioning of your personal knowledge infrastructure and also being candid about its internal contradictions, potential misfits and misalignments. In our study, we found that this was best done through discussions with others: a mentor, a coach, colleagues or a trusted counselor. After all, your personal knowledge infrastructure is very much a part of you. Having a personal knowledge infrastructure in place is one thing; being honest about how well-suited it is to your particular circumstances is very different.
   To this end, in addition to studying CEOs in action, we developed the outlines of a reflection and developmental process one can apply to one’s own circumstances. This is a framework to guide individual and peer reflection, built around a set of questions. (See “Evaluating Your Personal Knowledge Infrastructure”)
   Being a manager in today’s complex world requires becoming information-savvy in ways that are manageable and work for you in your specific context. What we learned from the CEOs we studied also applies more broadly to executives in general. Becoming and remaining practically knowledgeable is a critical task. It is a capability that managers need to learn, develop and continually refine, and it becomes increasingly important as the manager moves through his or her career and up the corporate ladder, when the risk of information overload significantly increases.