Skip to content

The Deciding Factor: book review

The Deciding Factor: The Power of ANALYTICS to Make Every Decision a Winner (2009) by Larry Rosenberger and John Nash with Ann Graham.

I’ve recently decided to deepen my understanding of analytics, in part to think about my website and how to better promote it, and secondarily to better understand this new way of thinking about business decisions. So I thought I’d start with this book and a broad introduction to this field.

The Deciding Factor is just that — a broad introduction written for executives by two guys from the Fair Issac corporation (the group that invented the credit score.) The essential gist of this book is that we can use analysis of the tons of data collected about customers/consumers to better understand how to not just make, but to automate decisions. If you’ve ever read the book The Numerati by Stephen Baker, it is the more detailed dive into the new world of data mining and decision making that is governing many business and political decisions.

Quoting Lowell Bryan, managing partner at McKinsey & Gary Hamel author of The Future of Management, the authors bring to our attention that “increasingly the work of managers won’t be done by managers. Instead it will be pushed out to the periphery. It will be embedded in systems.”

The greatest benefit will be seen in retail operations (credit cards, banks, retail stores) — where there is a lot of data on consumer behavior, and where the primary business decisions are being made on the front-lines by retail employees or call-center folks. One of the most interesting examples the authors give is about Best Buy and how they used analytics to better understand the types of customers coming into the store, created personas for each of these behavior types: Barry, the affluent techy enthusiast; “Jill”, the busy suburban mom; “Ray” the price-conscious family man; “”Buzz”, the young gadget fiend. Best Buy then rearranged the layout of the some of its stores to better serve these types. In addition, they trained their front-line employees to ask life-style questions to uncover the needs of each of these types (“how are you going to use this product?”, “Is it for you or someone else?”.)

The three essential components of using analytics is:

  1. Developing a Rules-Based System – automating high-volume operations decisions to make the decisions more consistent and increase control (such creating and understanding personas, and creating a set of questions to ask the types.)
  2. Using Predictive Analytics Models – creating decision models and frameworks to mathematically evaluate the trade-offs among conflicting objectives then execute decisions; using adaptive control (the process of making the best possible decision to control a complex system based on current knowledge and learn more about how the system behaves.)
  3. Connecting Decisions Across Multiple Dimensions – also known in part as cross-selling.

The use of both descriptive analytics – the process of finding relationships/patterns among data (i.e. figuring out the personas as described in the Best Buy example); and of predticive analytics – using what you do know to make informed decisions aobut what you don’t know to predict what might happen in the future – it expresses the future in terms of odds and probabilities.

The descriptions of how to make business decisions using analytics sounds similar to the way one imagines business decisions are currently made. However, there is a difference of degree and of control. In expert-driven decision making, one uses the experts and their experience. In data-informed decision-making, the experts interpret the information into a report and it informs the decisions. In data-guided, a predictive model replaces the unproven assumptions with objective information, and the model provides advisory decisions (methinks this is what happened in the investment banks). Finally in data-driven decision making, the running of analytic models and execution of decision are completely automated. The control is in the rules created to describe the system, whether that be a bank, a electronic store, or credit card business. The problems come in when the system is described incompletely and something unexpected happens.

The gist of this book is that decisions based on analytics are the way of the future. Decisions on how much credit-card increase to give people, whether to approve them for a loan, or what kind of other products can be sold to them are all automated. The rules are set by the business and math people, who interpret the data and create rules based on this interpretation. This feels very much like game-design.

To effectively use analytics you need organizations that have tons of data, and that collect the right kind of data. From this you can derive patterns and understand how to better make decisions, make suggestions based on these analytics. But you have to have good data, and thoughtful people making the rules, otherwise you may end up creating a system that controls you, rather than informs you. It’s a new way of thinking of understanding the world.

This book is a basic introduction to this world — at this point, I feel I need more in-depth thoughtful exploration of how it works best in the web world. Onto the next read.

Posted in book reviews, business.

Tagged with , , .

4 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Mike says

    I’ve been interested in personal analytics, given my obsessive ruminatory nature. Kevin Kelly has an interesting site at that’s all about tools people can use to gather data about their moods, habits, etc.

    A Mac/Windows program to look at in this regard is Optimism (

    I didn’t know you’d started a new blog! Found a reference to this post on LinkedIn.

  2. Sam Bruce says

    Rani, this string relates to a question that many in the corporate development world are pondering now – where is learning and development headed? As corporations continue to ratchet up their expectations for an explanation regarding how their L&D spend translates into value created for the enterprise, professionals in the field continue to struggle with what to measure and how to measure that value created. The Holy Grail in L&D these days is the L&D RoI.

    Getting back to my point, this book about Analytics could be seen by some in the L&D field to support their thought that L&D is moving more and more to the consultancy side of the continuum in which larger commercial firms with proven methodology will evenutally prevail against the academic business school providers who provide a more theoretical or in some cases customized approach to the client’s problems. What’s interesting to consider in this debate is that sometimes companies consciously choose the customized approach because they don’t want the consultant’s approach which often forces the problem into Framework X, Y or Z.

    Analytics can get you only so far. I thought it was interesting also to read last week about all the video footage that our military is gathering from unmanned airborne vehicles above Afghanistan. It seems they are barely keeping up with processing everything they are getting as it comes in and they now have a great deal of video footage that they could review again to make better links between the data they have already but they don’t have the resources to do so. They are drowning in data.

  3. Rani H. Gill says

    I agree that analytics can only get you so far. This is why you need the business people or learning people in the picture to give the analytics meaning, to interpret the data. I just think we have to be careful not to frame the debate as either/or. Too much of any one thing can be limiting – diversity is good.

    On the military — there is an interesting site — that shows how they are using cultural, immersive, video training that is visual, emotive, visceral. There are essentailly role plays that train the troops on how to interact with Iraqi people in the war zone — how to ascertain a threat. Contrast that with the other approach of video data collection — it may be overkill — but I think we need both approaches.

    Thanks for the comments Sam!

Continuing the Discussion

  1. Tools vs. Research, Write, Think, Design | wander@will linked to this post on January 17, 2010

    […] Research the learners. Are they novices? experts? do they have different roles/needs? can you create personas from these needs? Is it possible to actually collect data on them? How technically savvy are they? How do they get their information? How do they interact with their LMS? Do they interact with the LMS? Examples of defining personas and how to use personas can be found on the Cooper Journal website. One can think of personas as meaningful customer segmentation made real by colorful descriptions — see the description of how Best Buy uses personas in designing their stores and interacting with their customers in my review of The Deciding Factor. […]

Some HTML is OK

or, reply to this post via trackback.