Читать книгу Mind+Machine - Vollenweider Marc - Страница 4

Thank you for buying this book.

In 2015, after 15 years of operations in the field of research and analytics, we decided to adopt the notion of mind+machine at Evalueserve. We believe this marriage of the perceptive power of the human brain with the benefits of automation is essential because neither mind nor machine alone will be able to handle the complexities of analytics in the future.

The editorial team at John Wiley & Sons approached me in November 2015 to ask if I would like to write a book on how our mind+machine approach could help with the management of information-heavy processes – a topic that is of increasing interest to companies worldwide. We got very positive feedback from clients, friends, and colleagues on the idea, and decided to go ahead.

Mind+Machine is for generalist mainstream middle and top managers in business functions such as sales, marketing, procurement, R&D, supply chain, and corporate support functions, particularly in business-to-business (B2B) and B2C industries. We're writing for the hopeful beneficiaries and end users of analytics, and for people who might need to make decisions about analytics, now or in the future. The book is not a technical text primarily addressed to data scientists – although I firmly believe that even those specialists have something to learn about the primary problem in generating return on investment (ROI) from analytics.

We won't be looking at super-advanced but rare analytics use cases – there are specialized textbooks for those. Instead, we're looking at the efficient frontier, offering practical help on dealing with the logistics of managing and improving decision-making support and getting positive ROI at the same time.

After reading this book, you should know about key issues in the value chain of mind+machine in analytics, and be in a position to ask your data scientists, IT specialists, and vendors the right questions. You should understand the options and approaches available to you before you spend millions of dollars on a new proposal. You'll learn some useful things to demystify the world of analytics.

We're also proposing a novel approach, the Use Case Methodology (UCM), to give you a set of tangible and tested tools to make your life easier.

We've included 39 detailed case studies and plenty of real-life anecdotes to illustrate the applications of mind+machine. I'm sure you'll recognize some of your own experiences. And you'll see that you're far from alone in your quest to understand analytics.

What makes me want to put these ideas about the problems and solutions to analytics issues out in the world is conversations like these two.

The first words to me from a very senior line manager in a B2B corporation:

“Marc, is this meeting going to be about big data? If so, I'll stop it right here. Vendors are telling me that I need to install a data lake and hire lots of increasingly rare and expensive statisticians and data scientists. My board is telling me that I need to do ‘something' in big data. It all sounds unjustifiably expensive and complex. I just want to make sure that my frontline people are going to get what they need in time. I keep hearing from other companies that after an initial burst of analytics activity, real life caught up with them, the line guys are still complaining about delays, and the CFO is asking a lot of questions about the spend on big data.”

During a meeting with the COO of an asset manager to define the scope of a project:

“We do thousands of pitches to pension funds and other institutional investors every year. We have over 25 different data sources with quantitative data and qualitative information, with lots of regional flavors. However, we still put the pitches together manually and get the sign-offs from the legal department by e-mail. There must be a smarter way of doing this.”

Why is analytics becoming such a controversial and challenging world? Why are managers either daunted by overhyped new initiatives and processes that they don't understand or frustrated by the feeling that there should be a better way to do something, given all this talk about better, bigger, brighter analytics?

Typical line managers want to get the right decision-making support to the right people at the right time in the right format. The proliferating number of analytics use cases and available data sets is not matched by an expansion in individuals' and companies' capacities to mentally and logistically absorb the information. Additionally, existing and new compliance requirements are piling up at a remarkable speed, especially in industries with a high regulatory focus, such as financial services and health care.

Analytics itself is not truly the issue. In most cases, the problem is the logistics of getting things done in organizations: defining the workflow and getting it executed efficiently; making decisions on internal alignment, the complexities of getting IT projects done, and other organizational hurdles that hamper the progress. These complexities slow things down or make projects diverge from their original objectives, so that the actual beneficiaries of the analytics (e.g., the key account manager or the procurement manager in the field) don't get what they need in time.

Many other issues plague the world of analytics: the proliferation of unintuitive jargon about data lakes and neural networks, the often-overlooked psychology of data analytics that drives companies to hold too dearly to the idea of the power of data and makes the implementation more complex than required, and the marketing hype engines making promises that no technology can fulfill.

Based on hundreds of client interactions at Evalueserve and with my former colleagues in the strategy consulting world, it became increasingly clear that there is a strong unmet need in the general managerial population for a simplified framework to enable efficient and effective navigation of information-heavy decision-support processes. Simplicity should always win over complex and nontransparent processes – the analytics space is no exception.

I want to demystify analytics. I'll start with the fundamental observation that terms such as big data and artificial intelligence are getting so much attention in the media that the bricks-and-mortar topics of everyday analytics aren't getting the attention they deserve: topics such as problem definition, data gathering, cleansing, analysis, visualization, dissemination, and knowledge management. Applying big data to every analytics problem would be like taking one highly refined chef's tool – a finely balanced sushi knife, for example – and trying to use it for every task. While very useful big data use cases have emerged in several fields, they represent maybe 5 percent of all of the billions of analytics use cases.

What are the other 95 percent of use cases about? Small data. It is amazing how many analytics use cases require very little data to achieve a lot of impact. My favorite use case that illustrates the point is one where just 800 bits of information saved an investment bank a recurring annual cost of USD 1,000,000. We will discuss the details of this use case in Part I.

Granted, not every use case performs like that, but I want to illustrate the point that companies have lots of opportunities to analyze their existing data with very simple tools, and that there is very little correlation between ROI and the size of the data set.

Mind+Machine addresses end-to-end, information-heavy processes that support decision making or produce information-based output, such as sales pitches or research and data products, either for internal recipients or for external clients or customers. This includes all types of data and information: qualitative and quantitative; financial, business, and operational; static and dynamic; big and small; structured and unstructured.

The concept of mind+machine addresses how the human mind collaborates with machines to improve productivity, time to market, and quality, or to create new capabilities that did not exist before. This book is not about the creation of physical products or using physical machines and robots as in an Industry 4.0 model. Additionally, we will look at the full end-to-end value chain of analytics, which is far broader than just solving the analytics problem or getting some data. And finally, we will ask how to ensure that analytics helps us make money and satisfy our clients.

In Part I, we'll analyze the current state of affairs in analytics, dispelling the top 12 fallacies that have taken over the perception of analytics. It is surprising how entrenched these fallacies have become in the media and even in very senior management circles. It is hoped that Part I will give you some tools to deal with the marketing hype, senior management expectations, and the jargon of the field. Part I also contains the 800 bits use case. I'm sure you can't wait to read the details.

In Part II, we'll examine the key trends affecting analytics and driving positive change. These trends are essentially good news for most users and decision makers in the field. It sets the stage for a dramatic simplification of processes requiring less IT spend, shorter development cycles, increasingly user-friendly interfaces, and the basis for new and profitable use cases. We'll examine key questions, including:

● What's happening with the Internet of Things, the cloud, and mobile technologies?

● How does this drive new data, new use cases, and new delivery models?

● How fast is the race for data assets, alternative data, and smart data?

● What are the rapidly changing expectations of end users?

● How should minds and machines support each other?

● Do modern workflow management and automation speed things up?

● How does modern user experience design improve the impact?

● How are commercial models such as pay-as-you-go relevant for analytics?

● How does the regulatory environment affect many analytics initiatives?

In Part III, we will look at best practices in mind+machine. We will look at the end-to-end value chain of analytics via the Use Case Methodology (UCM), focusing on how to get things done. You will find practical recommendations on how to design and manage individual use cases as well as how to govern whole portfolios of use cases.

Some of the key questions we'll address are:

● What is an analytics use case?

● How should we think about the client benefits?

● What is the right approach to an analytics use case?

● How much automation do we need?

● How can we reach the end user at the right time and in the right format?

● How do we prepare for the inevitable visit from compliance?

● Where can we get external help, and what are realistic cost and timing expectations?

● How can we reuse use cases in order to shorten development cycles and improve ROI?

However, just looking at the individual use cases is not enough, as whole portfolios of use cases need to be managed. Therefore, this part will also answer the following questions:

● How do we find and prioritize use cases?

● What level of governance is needed, and how do we set it up?

● How do we find synergies and reuse them between the use cases in our portfolio?

● How do we make sure they actually deliver the expected value and ROI?

● How do we manage and govern the portfolio?

At the end of Part III you should be in a position to address the main challenges of mind+machine, both for individual use cases and for portfolios of use cases.

Throughout the book I use numerous analogies from the non-nerd world to make the points, trying to avoid too much specialist jargon. Some of them might be a bit daring, but I hope they are going to be fun reading, loosening up the left-brained topic of analytics. If I could make you smile a few times while reading this book, my goal will have been achieved.

I'm glad to have you with me for this journey through the world of mind+machine. Thank you for choosing me as your guide. Let us begin!