Read an Excerpt
Behind Every Good Decision
How Anyone Can Use Business Analytics to Turn Data into Profitable Insight
By Piyanka Jain, Puneet Sharma, Lakshmi Jayaraman AMACOM
Copyright © 2015 Piyanka Jain and Puneet Sharma
All rights reserved.
ISBN: 978-0-8144-4922-6
CHAPTER 1
Analytics or Die
THIS CHAPTER WILL TALK ABOUT:
Why analytics? Because you can't depend on chance.
What is analytics?
What is hypothesis-driven analytics? Imagine swimming the Pacific Ocean to find gold. Wouldn't you rather narrow down the most likely areas where gold could be, so you wouldn't have to cover every square inch of water from China to America?
Similarly, data is an ocean, and hypotheses help to narrow down where to find the most likely answers. Hypotheses are generated by human intuition based on the collective intelligence and experience of stakeholders and their understanding of the business and their environment. Data validates the hypotheses to come up with a convergent solution. The strength of the solution then will lie in the best of both—data and hypotheses.
How does hypothesis-driven analytics strategy work?
It starts with answering a set of plain sounding questions, "What is happening?" and "Why is it happening?" These simple questions have helped companies like P&G identify actions that can be taken to address a situation. Action without insight means decisions are made without any perspective on which paths would yield results.
In 2010, P&G CEO Bob McDonald strategized to digitize the company's processes end-to-end to achieve business optimization and decision efficiency using analytics. But P&G was simultaneously tasked with cutting costs. By 2012, it had eliminated 1,600 nonmanufacturing jobs, and its IT organization had cut over $900 million in total expenditure. Despite reducing headcount and expenses, P&G managed to score an increase in gross revenue.
How? P&G invested in analytics expertise to develop its capabilities to make wise and timely decisions in unpredictable and stressful environments. Its analytics provided predictions about P&G market share and other performance statistics up to a year into the future. At the core of this capability was a series of analytic models that reveal events in the business as they occur, including peaks and valleys in performance, identify reasons why they occur, and point to actions P&G can take to leverage or mitigate effects from these events.
P&G's analytics is driven by a cultural mindset to ask questions and propose hypotheses before delving into complex data analytics. "What" monitors areas like shipments, sales, and market share to keep an eye on the key metrics for the business. When exceptions occur, "Why" drills down to the drivers—country, territory, product line, and store levels—to understand the specifics behind what is happening. Analytics then identifies "action levers" that P&G can pull, such as pricing, advertising, and product mix, and provides estimates of what that action can potentially deliver. By efficiently establishing the whats and the whys, decision makers at P&G have been able to jump right into solving a problem, radically increasing efficiency and the pace of conducting business.
This strategy surely works for giants like P&G. Now, how can You, facing an ever-shrinking budget and ever-brimming datasets, take on your company's health and prosperity like the goliaths of the industry? How can you execute your business strategy with optimal efficiency and make the smartest business decisions possible?
You can, by asking the right sets of "Whats" and "Hows" and by using a simple structured process—a hypothesis-driven analysis, followed by testing of insights—identify strategies to implement.
It isn't rocket science, but rather a commonsensical secret. Like P&G, just ask these plain questions through the hypothesis-driven data-led process outlined below (see Exhibit 1-1). A hypotheses-driven analytics strategy can provide significant benefits, including greater insight and faster time to information and actions. The actionable insights generated from a hypotheses-driven approach are precise enough to allow for clear implementation.
A complete hypothesis-driven process is illustrated in the figure below.
A NEW NORMAL
Here's a bold statement. Every company is exposed to the very same factors to virtually the same degree. It is not the challenge that distinguishes the success of an organization, but its response to the challenge. Companies that ride the economic and competitive tsunamis with ease can survive and thrive. On the other hand, bad internal decisions made in response to market challenges can weaken a company to the point of collapse.
The reality of a weakened economy since 2008 and its limp, slow recovery has complicated business worldwide. The effects of plummeting sales, shrinking budgets, and demanding customers have elicited varied reactions from organizations—some made smarter decisions and rode the waves, while others drowned in the wake. That said, numerous factors contribute to the success or failure of an organization. Intense pressures from external market conditions can impede profitability. The rise and fall of consumer confidence, investments, and economic turmoil can create overwhelming obstacles. Competition and rapidly emerging technology can wreak havoc on the best-laid plans. This landscape is forcing a paradigm shift in tools and approach to business problems.
Analytics is your key to making smarter decisions based on your very specific internal business dynamics. Not surprising is that a recent study based on data from 60 case studies found that analytics pays back $10.66 for every dollar spent.
"But really, is analytics for me? Can analytics help me double my business with a shrinking budget?" Yes and yes.
ANALYTICS IS FOR EVERYBODY
Yes—everyone from daily managers to leaders to scientists to one looking to buy a car. If you aspire to run your business with maximum efficiency and effectiveness and make the smartest business decisions possible, if you are part of an enterprise whose tolerance for decision error is getting slimmer by the day, if you are looking for new sources of advantage and differentiation for your enterprise, if you are struggling in a highly complex global environment with increasing competition, expedited time to market, and highly selective customers, if you nurture any hopes of increasing your business in spite of all external challenges and an ever-shrinking budget, and even if you're just looking to buy a car, then it is time for you to say hello to analytics.
CHAPTER 2
What Is Analytics?
THIS CHAPTER WILL TALK ABOUT:
What is analytics?
Business analytics versus predictive analytics.
Business analytics versus business intelligence.
Big Data versus business analytics.
What is growth hacking?
Scene 1
You are Ben, the Quality Control Manager at Bright Sun Solar Cell Manufacturing Company, and you have just received a note from your testing group that the last batch of solar panels has had an alarmingly low rate of product efficiency. With your reputation at stake, you need to move full throttle into action to understand what went wrong. Questions race through your mind. "Why was there such a high failure rate on the last batch?" "How will we fix it?" "How can I ensure that future batches don't have the same problem?"
The clock is ticking. Each incremental failure adds to the revenue losses already incurred, costs that are creeping into the millions.
Scene 2
You are Andreas, a treasure hunter, the Captain of the Moon Princess. Your dream is about to come true. A private investor has just offered you a large grant to search for lost treasures in the Pacific Ocean. Better yet, she offered you a special bonus if you find the sunken 1715 shipwreck, dubbed "Plate Fleet" for the silver plates on board.
But the grant is only sufficient to cover you and your crew for 30 days at sea.
WILL YOU BE CHRISTOPHER COLUMBUS OR SHERLOCK HOLMES?
We'll start by saying that in these two very different situations, treasure hunting and solving manufacturing defects, the path to the solution, actually the efficient solution, would be very similar.
If you were like Christopher Columbus, you might set sail into the ocean with your crew and submarine in tow and just start looking for the treasure wherever it lay. This would be an enjoyable route and you would, no doubt, find yourself enthralled with the sublime beauty of the ocean—the shimmering green waters, the dolphins leaping playfully in your wake, and the ancient, remarkable coral reefs.
Meanwhile, the clock is ticking on your grant money. Are you doing it as efficiently and expeditiously as possible? In 30 days, will you have found the Plate Fleet? Probably not, as your actions of exploration are independent of your goal to find the Plate Fleet. Had you been tasked to look for killer whales or a kind of edible seaweed, your actions might have been exactly the same. You have taken the explorer approach.
Now, if you were like Sherlock Holmes, what would you do? You might identify potential areas where shipwrecks have occurred and research specifically where the Plate Fleet may have sunk. You might look at historical trade routes and records of wrecks. Then, using depth information, you would eliminate possibilities. And, you don't want just any shipwreck; you want to pinpoint where the Plate Fleet sank. You would look at the records of 1715 hurricane paths. By using clues and facts, you would begin to identify potential areas where the treasure could be found. Once you have identified a dozen or so potential locations and prioritized the top three, you would send your submarine or deep-sea divers down to investigate.
You have just multiplied your chances of finding the Plate Fleet, and in a much shorter time, just by engaging in a focused search based on appropriate information. Should you fail, you would still have time to strategize again and attack the problem again. This is the detective approach.
As Ben, your approach to identifying and fixing the problem that caused the solar panels to fail is no different than the treasure hunt. With Columbus's explorer approach, you could be collecting an incredible amount of data points in the hope of finding the cause of the failure. But, as you can imagine, with multiple assembly lines, each one with a multitude of processes and equipment, the chances of stumbling into the problem area quickly are going to be slim.
Maybe the problem is simply a single valve malfunction, but imagine the probability of finding it among the hundreds of things you would have to look at! Again, how lucky are you, and are you willing to trust your career to luck?
BEN, MEET DATA-DRIVEN DECISION MAKING
Play detective. Pore over the data to decipher the clues that will tell you what went wrong. This is Business Analytics 101. Start asking questions. Where did the failure happen? Are all the lines producing unusually faulty products, or was this an isolated incident? When did the problem start? Where exactly are the faults in the product and to what processes or equipment does it correspond? You get the gist.
The most important thing to note is this: You don't have to know all the answers to the guided questions you are asking. You can construct a solid hypothesis based on what you have already learned, partial as it may be, and then use that hypothesis to identify the most likely problem suspects.
By being Sherlock, you quickly discover that high temperatures in line ten caused the fault in the product. Temperature issues are often caused by raw material quality issues or by hardware, such as a malfunctioning heat exchanger. By identifying the top things to look for, you can easily narrow down the cause of the problem, which in this case is a faulty valve on a heat exchanger. You have saved time and resources getting to the root of the problem, and your supervisor recognizes you for it. At least that's what we hope!
IS IT ROCKET SCIENCE?
The word analytics does conjure up a vague image of a long complex equation from a statistics class, but analytics as applied to business is not that complex. The reality is that a very small number of statistical techniques learned in statistics class can actually be applied in a business scenario. From the techniques that can apply to a business problem, only a very small percentage meets the business constraints of ROI, explainability, maintainability, turnaround time, and scalability. So, simpler analysis techniques win hands down in a business situation.
In 2006, Netflix announced a data mining competition for a prize of $1 million to improve its movie rating prediction by 10 percent. A classic business meets statistics scenario. A very large number of data miners competed. A year later, the progress prize winning team reported it had spent more than 2,000 hours to achieve an 8.43 percent improvement. Two years later, an ensemble (a combination of models) finally delivered the 10 percent improvement to qualify as the winner. The winning solution was a blend of hundreds of individual models, but Netflix is not using it. Why? Because the solution was operationally expensive to implement, and the gain in revenue from the better recommendation didn't offset the cost of scoring and maintaining the model.
As it turns out, the most useful analytics techniques in business are contained in a small set of simpler techniques that can be learned by most professionals. You'll also be surprised to know that it is actually the people skills needed for bridging the gap between business and math that make or break the deal.
ANALYTICS IN BUSINESS
Analytics is only useful when it drives impact. Depending on your business, impact can be revenue growth, process efficiency, and improved offerings.
Analytics for impact = Data science + Decision science
Analytics for impact has two components:
• Data science: This is the technical track, designed to derive insights from data.
• Decision science: This is the business track, designed to align stakeholders so that the valuable insights produced using the data science track can be inserted into the decision-making process and converted into action.
Great analytics is not just about cool and complex models. It is also about using soft skills, understanding the business, and presenting relevant insights useful in the context of business to drive business impact. Unless analytics drives business impact, it is not analytics. It is just statistics; it is just data science. Often, analysts get so mired in the data to focus on getting the coolest insights that they miss the human element.
A successful business analytics professional is not a statistician (although statisticians make great analysts), but someone who engages enthusiastically and appropriately with business counterparts who approach them to perform data analysis. True, the analyst needs to choose the right technique for analysis and deliver insights. But a successful analyst uses influence and soft skills to build alignment with the stakeholders or business counterparts. This ensures that when the golden nuggets of insights are mined, the business counterpart is ready to act on it by turning those insights into business impact. More information on this will be presented in Chapter 4 and Chapter 10.
Using question-and-answer sessions with their counterparts, the successful analyst arrives at a crisper definition of what is being asked of the data (the real business question) and what might be some of the clues toward the answer (hypotheses). They can use these answers to dig into the right data, do the appropriate analysis, and make actionable recommendations.
This is not to say technical analysis and modeling skills are not important, but without the rudder of business and people skills, the ship becomes lost in the ocean.
THE 80:20 RULE OF ANALYTICS
The analytics landscape is still in its formative stage. So definitions and terminology are in flux and will probably remain so for some time. For the purpose of this book, we will use these terms:
Business analytics: The use of simpler analytics methodologies on past data.
Advanced analytics: Everything else, including predictive analytics.
(Continues...)
Excerpted from Behind Every Good Decision by Piyanka Jain, Puneet Sharma, Lakshmi Jayaraman. Copyright © 2015 Piyanka Jain and Puneet Sharma. Excerpted by permission of AMACOM.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
