If you are a manger who is adamant at relying on your ‘golden gut’ or intuition to make decisions, this book is not for you: ‘Analytics at Work’ by Thomas H. Davenport, Jeanne G. Harris, and Robert Morison (Harvard). The authors concede that sometimes intuitive experience-based decisions work out well, but they hasten to warn that often such decisions either go astray or end in disaster.
The authors cite examples of the most extreme cases of ill-informed decision making, such as executives pursuing mergers and acquisitions to palliate their egos, neglecting the sober considerations that create real value; banks making credit and risk decisions based on unexamined assumptions about always-rising asset values; and governments relying on sparse intelligence before deciding whether to wage war.
And examples of situations where non-analytical decisions do not lead to tragedy but leave money on the table include: product or service pricing based on hunches about what the market will bear rather than on actual data on what consumers have been willing to pay under similar circumstances in the past; hiring people based on intuition, not on an analysis of the skills and personality traits that predict an employee’s high performance; and maintaining ‘comfortable level of inventory,’ in the place of ‘a data-determined optimal level.’
The eBay example
As if to dispel the myth that easy things like making changes to your home page do not require elaborate deliberation, there is the discussion of eBay’s approach in the opening chapter. The company – which enjoys more than a billion page views per day, for the 113 million items for sale in over 50,000 categories at any given time – ‘undertakes extensive and varied analyses by performing randomised tests of Web page variations before making any change to the Web site or the business model.’
To make sense of all these tests, eBay built its own application, called the eBay Experimentation Platform, to lead testers through the process and to keep track of what’s being tested at what times on what pages, one learns. “In addition to online testing, eBay considers change to its Web site using a variety of analytical approaches: the company conducts extensive, face-to-face testing with its customers, including lab studies, home visits, participatory design sessions, focus groups, and iterative trade-off analysis. eBay also conducts quantitative visual design research and eye-tracking studies, and diary studies to see how users feel about potential changes.”
Lot of data
Companies, governments, and nonprofits, in both developed and developing economies, generally collect and store a lot of data, from a variety of sources. “The data may come from transaction-oriented applications such as ERP (enterprise resource planning) systems from software vendors such as SAP and Oracle, scanner data in retail environments, customer loyalty programs, financial transactions, or clickstream data from customer Web activity. But what do they do with all this information? Not nearly enough.”
In this context, the authors recount a conversation with the managers at a retail grocery chain on what they did with their data. “One manager said, ‘Well, we sell it. In fact, we make more money selling data to retail data syndication firms than we do selling meat.’ We dutifully said that result was impressive, but the firm’s managers also admitted to a less impressive fact – that they later buy back their own data, mixed with that of local competitors.” What else did they do with the data? They stored the data on disk, on tape, under a mountain so that it is safe from nuclear attack!
Internal information
The ‘analytical DELTA’ that Davenport et al. propose has ‘data’ at the start, followed by enterprise orientation, leadership, targets, and analysts. A chapter devoted to ‘data’ has many examples of how companies can benefit from information sourced from internal operations or customer relationships.
For instance, the Royal Shakespeare Company in the UK could increase the number of ‘regulars’ by more than 70 per cent by examining ticket sales data of over seven years. Olive Garden, an Italian restaurant, is another example; it uses data on store operations to forecast almost every aspect of its restaurants, such as ‘for staffing and food preparation down to the individual menu item and component.’
You may perhaps be aware of the Nike+ program which uses sensors in running shoes to collect data, and upload to the runner’s iPod, and then to the company site. Do you know that “through analysis of this data, Nike has learned that the most popular day for running is Sunday, that wearers of Nike+ shoes tend to work out after 5 pm, and that many runners set new goals as part of their New Year’s resolutions. Nike has also learned that after five uploads, a runner is likely to be hooked on the shoe and the program”?
Cubes, arrays, nonnumeric
The authors speak of three methods of structuring data for analysis, viz. cubes, arrays, and nonnumeric. Data cubes, as they explain, are collections of pre-packaged multidimensional tables. While these may be useful for reporting and ‘slicing and dicing,’ they are less useful for analytical exploration, ‘because the variables they contain are limited to what some analyst thought should be in the cube and in the resulting report.’
Arrays, the second option, are common in the form of rows and columns of spreadsheets. This format allows for the most flexibility, but may be confusing to non-technical users who don’t understand the structure of the database or the locations and fields of the data within it, caution the authors.
And the ‘last frontier’ for data analysis, the unstructured non-numeric data, can be in a variety of forms, such as ‘the vocal tone of your customers during service calls,’ and ‘blogs, Web pages, and Web-based ratings and comments,’ as way to understand consumer sentiments about the company.
“Firms are also increasingly interested in mining text in internal databases – like warranty reports and customer complaint letters – for customer service issues, ‘reason fields’ (for example, in denying credit), and product descriptions (for example, to reconcile multiple product hierarchies following mergers and acquisitions).”
Analytical culture
Towards the latter part of the book is a chapter on building an analytical culture, where the authors recommend the integration of analytics with other cultural priorities. If, for example, you are like P&G, zealous about developing new products, a complementary analytical culture can encourage the development of new product metrics to assess customer reactions and measure how new products are faring in the marketplace, they note.
The case of Hotels.com, a business unit of Expedia, Inc. is described elaborately in the book. Web activity and financial reports showed solid growth and increased sales, but the truth of the customer experience waited to be unearthed, till the company instituted a serious ‘voice of the customer’ program combined with Web analytics.
“The program allowed customers to indicate problems at any time in a session, using software that records every screen presented to a user and which mouse clicks a customer makes.” Interestingly, the company even created separate phone numbers (more than 700 in total) that dynamically appeared based on the page and how the customer got to the site, so that when customers called a certain number, it would be obvious where they encountered problems, the book informs. As a result, Hotels.com was able to identify problems that would otherwise have escaped notice.
The real cultural shift, however, happened when the company discovered that a large percentage of customers who had made it all the way to the end of the checkout process did not complete a transaction, due to ‘a combination of unclear messaging, user flows, database issues, and outright bugs.’ All the relevant groups came to the table and, operating collaboratively at an above-normal pace, resolved the problems, to improve the conversion rate and bring immediate additional revenues, apart from creating customer goodwill…
Imperative study for the data-minded business managers.
**
Tailpiece
“We have two Co-CEOs, one with a good voice, and the other with a photogenic face, and so…”
“You are well-equipped both for the radio and the TV?”
“Yes, and for the employees, we are creating an animated version of a boss with combined features!”
**
BookPeek.blogspot.com
