Earlier this week I shared an article with a colleague entitled Can Analytics Helps Colleges Graduate More Students?, by Tanya Roscorla. The article presented some excellent case studies where universities are leveraging analytics to test innovations and help students identify the best path towards course completion and graduation.
While she appreciated the article, I found her response compelling (and obviously the inspiration for this post). She noted that the article stated “Colorado universities send about 20,000 people a year to community colleges because they aren’t ready for college-level math and English. But they’re losing many of these students in the process.”
Her response: “It’s interesting that people see this as a problem with colleges and universities – why are high schools graduating people who aren’t ready for college-level math and English?”
The question is appropriate and indicative of a much larger issue that I won’t discuss here – what are the roles of high schools, lower schools and parents in terms of getting kids ready for college?
But she got me to thinking…it’s great that Texas, Hawaii, Maryland are others are leveraging analytics to help students perform better once on campus…but isn’t that simply working on the symptoms of a larger problem and not really understanding or solving the root cause? And if we don’t solve the root cause, we’ll continue to mitigate the symptoms perpetually.
Let’s think about this in terms of a recent business example. A retailer reported sales at a particular store that did not meet expectations – it missed it’s annual sales target by over 20% (approximately $1M). Initial inquiries looked at the usual suspects…lack of traffic, ineffective marketing, insufficient labor, weather, etc. Sound familiar? And while each of these likely played some role in the missed forecast, after further inspection and questioning, a senior executive found the root cause – the initial forecast was flawed! It used a predictive model based on a different geographic area that was not comparable. So while the retailer could (and should) do everything it can to help improve traffic, marketing, labor, assortment, etc…the reality is that the store was meeting expectations based on the actual geography and demographics.
Both the business and education examples above cause emotions to run high – people are passionate about the topics and have vested interests in the outcome. Roscorla closes her article with this keen insight from Mark Milliron, co-founder and chief learning officer of Civitas Learning:
“One of the most important things to keep in mind when dealing with analytics and interventions is to balance out the hype and the skepticism. True analytics believers promise too much and don’t deliver enough, while skeptics fight back against innovation and change. As a result, true believers and caustic cynics hijack important conversations, which are complicated and require tough mindedness…[organizations] need to calm the caustic cynics, temper the true believers and create a space for learning analytics to move forward to the next level.”
So use analytics to identify symptoms and trends…then ask the 5 Whys (see cartoon below) and use analytics some more to get to the root cause…then make a data-driven decision to solve your problem.
People often have fears, real or perceived, that if not addressed properly can adversely impact the successful implementation of any business transformation. Analytics is no different. Effective change management is a critical component of any Analytics Strategy.
Consider the three fears below recently shared with me and approaches to mitigate them:
- “Analytics won’t produce an answer.” – While this is a viable outcome, it should not be a fear or concern. If the best data scientists analyze the most relevant data with the best tools and methods available…and don’t identify a pattern in the data…that is your answer! The answer may be that the data won’t help – there is no pattern. Your best course of action is to use your instinct and experience. But at least you’ve explored whether the data could help and have ruled that out as a possibility. And you’ll be making a better decision!
- “They won’t tell us that they cannot find an answer.” – This concern goes something like this: the data scientist won’t find a pattern, but in order to justify their rate/job/existence, they’ll fabricate an answer. And while some disreputable analysts might exist, the majority are reputable, inquisitive and curious by nature with a huge appetite to find an answer. They’ll keep digging and analyzing as long as you allow. And when you call it off, they’ll let you know that no pattern was found. So check the references on your hires, be sure you can trust them. Then let them at the data, and be comfortable that they won’t fabricate the result.
- “We’ll think we are smarter than the data and we won’t listen.” – This fear simply comes down to leadership and culture. Does your organization have top down senior level advocacy of a data-driven decision-making culture? And will they hold others accountable? If yes, then this fear is rather easily managed by evaluating your decision making performance. If specific individuals “don’t listen”, consistently decide against the recommendation of the data, and turn out to be wrong, then leadership must hold them accountable. On the other hand, if they turn out to be right most of the time, make sure that you adjust your models, data sets or methods.
Are these fears common in your organization? Likely.
Are there others? Likely.
Find out. Talk to key business leaders and partners, especially those resistant to analytics. Ask them and take time to understand their concerns. Then go about resolving them. Only after ensuring that the people aspect of an analytics initiative is address, can you effectively implement the process and technology changes necessary to implement the strategy.
Over the course of my career, I have authored and reviewed hundreds of business cases. And nearly every one of them included the same qualitative benefit: “better decision making”. For most of them, this was nothing more than filler – an additional bullet point in a list of items that really didn’t matter.
But Analytics is different – better decision making is the entire point of it! How this is achieved can be defined by the simple equation below:
Data + Gut > Gut = Better Decision Making
Let’s start by defining the two variables involved:
- Data – represents the quantitative analysis performed on a relevant set of metrics
- Gut – represents the cumulative experience, functional/industry knowledge and instincts of the decision maker
Now as a decision maker, let’s assume that you have a choice between alternative A and alternative B. Consider the two scenarios below:
- Scenario 1 – Your gut/instinct/experience tells you that alternative A is preferable. Perhaps you’ve seen it be successful more often – maybe you have personal experience in a prior similar situation. Or maybe alternative B has never been tried or proven in your field or industry. Regardless, alternative A just feels like the right decision.
- Scenario 2 – Same as scenario 1, but you also have data to support your decision. After analyzing the appropriate data set, using advanced quantitive methods, your analytics support team advises you that alternative A has a higher probability of a successful outcome than alternative B with a high confidence level.
Scenario 2 clearly represents a approach that will drive a better decision-making process. You’ll have more confidence in your decision, as will your boss, customer, and business partners.
There are two additional questions to think about:
- “What if analyzing the data takes too long or costs too much?” – This is a valid concern an done that the decision maker needs to assess. Is there timeliness or urgency to your decision? How big of a decision is this – a $25k or $25M opportunity? What are the risks of making the wrong decision? In cases where the risk is small, urgency is high, and cost/duration of analysis is also high – then the decision maker should strongly consider going with their gut. But even in this example, you’ve explored the option of using analytics and determined that it was not optimal due to these other factors.
- “What if the data doesn’t show a pattern or indicate a higher probability outcome?” – This is a common fear, but in reality, not a concern. Even if your analytics team comes back and tells you that “there is no pattern in the data” or “we cannot recommend alternative A or B” – this is still information and improves your decision making process. You’ve looked at the data and determined that is will not influence the decision – the probability of success with either option is 50/50. So go with your gut. Knowing this is still better than simply going with your gut and after the fact, wondering if the data might have helped you make a better decision.
But there is a catch – there is no guarantee that alternative A will in fact turn out to be the better choice. It very well may end up being alternative B. Business happens. Unexpected events occur. Surprise variables influence the outcome. Or it is just a matter of probability – sometimes the long shot beats the odds.
Regardless of the actual outcome, A or B, scenario 2 (data + gut) is clearly the better process. You’ve leveraged your gut/instinct/experience and used it effectively and you’ve also leveraged additional information. You’ve been more comprehensive in your assessment. You’ve made a better decision.
Now you can feel good about that bullet point in the analytics business case…and move it over the the list of quantitative benefits (more on that in a future post)!
Many retailers are currently struggling to understand how to define a strategy to effectively leverage advanced analytics. More so than the omni-channel imperative (which arguably is simply about keeping up with each other), there is a window of opportunity during which effective use of analytics can be a significant differentiator and a source of competitive advantage. In December 2013, Accenture released an outstanding research report, entitled “Seamless Analytics – Three Imperatives for the Retail Digital Marketplace“, where they state “the choice is clear: the digital marketplace is the future, and the decisive difference is analytics.” That assertion, as well as the three imperatives identified are spot on. However, there is one very important imperative that was missing…more on that later. But first, here are the three imperatives, with my insights added.
- “The Insights Imperative: Applying retail analytics to data for sharper customer insights.” – Retailers need to get a handle on their internal data and establish data governance processes to cleanse, manage and monitor the quality of their data. Most have shortcomings in their current data environments and this initiative can be long and tedious. It is important, but it is also not a complete impediment to leveraging analytics. The fact is that data will never be perfect and monitoring/cleansing will become a normal operational process. Measuring data quality and using quantitative methods to adjust for issues can be employed to enable effective analytics while data quality is being improved to acceptable standards.
- “The Actions Imperative: Turning insights into actions, at speed and at scale.” – Analytic insights must drive actions to be effective. Without a specific decision or action, analytics are merely interesting and not nearly worth the investment or attention. But once a decision point is identified, many retailers become overly reliant on their experience, instinct and gut. And for good reason…those are likely what has helped them be successful. But now with the availability of advanced analytics, decision-making is vastly improved. Consider the following equation: data + gut > gut. When you add an understanding of the data (assuming it was developed with a reasonable effort, cost and duration) to gut feel, the end decision is better than one made on gut feel alone. Those retailers that understand this equation will soon, and significantly, outperform those that rely on their historical methods.
- “The Outcomes Imperative: Focusing on business outcomes.” – To be truly transformational, analytics must be employed to support specific business outcomes and strategies. Once a retailer understands how they plan to differentiate, be it customer service, product quality, scale or something else, then metrics and analytics can effectively be employed to support the strategy. More metrics and more analytics are not necessarily better…but having a smaller set of the right measurements that support and drive the overall business strategy is imperative. And these metrics must become embedded in the ongoing operational mindset of the enterprise.
- The Missing Imperative – Leadership – In my opinion, this imperative is by far the most important, and without it, the value of analytics in the organization will be limited. Senior leadership, starting at the CEO and extending throughout the executive leadership team, must be strong advocates and active supporters of analytics. With that foundation, a retailer can change the culture from gut-driven to data-driven decision-making. Without it there is little chance of success. So the initial focus for any retailer must be on understanding these key stakeholders, educating them and confirming their commitment to analytics. Once this is accomplished, no small feat, then strategies and tactics around the other three imperatives can be defined and executed.
And then, as Accenture confirms, “when retailers routinely start making data-driven and analytics- supported decisions with the seamless customer experience in mind, they will increasingly see the impact in their bottom line.”
Welcome to the Cherry Advisory blog!
I have long been a strong advocate and practitioner of servant leadership – an approach that prioritizes the needs of others, helps individuals develop and perform as successfully as possible. Throughout my career as a consultant, it’s always been about making my client as successful as possible and enabling them to be successful once I’m gone. Early in my career a senior partner once shared with me that “the mark of a good consultant is that they make themselves expendable” through knowledge sharing to their clients. On the client side, a deep caring for my organization across all levels created a natural tendency towards servant leadership – my success would be measured based on the collective success of my leaders, peers and reports.
As my career has progressed, my views on servant leadership have transformed into a somewhat utopian view of leadership. In every organization, in every relationship, in every role…we should all act as if we had the same job. My job is to make your job easier. What’s your job?
Sounds simple enough, but how does it apply? If we each just took an extra moment to think before we act, “how can I make this easier for my business partner”?
- Do I need to win this argument? Or can I compromise?
- How can I best brief my boss on the situation, options, risks and recommendation?
- Should I share a great 20 page article? Or can I summarize the main points?
- How can I put my direct reports in a position to be successful? Can I be more explicit in my expectations?
- Should I demonstrate my expertise by making things sound complex? Or rather, prove my skill by simplifying a difficult subject?
That last point reminds me of one of my favorite quotes, from C.W. Ceran: “Genius is the ability to reduce the complicated to the simple”. Sometimes our most challenging problems can be solved with simple solutions. It reminds me of the myth that the U.S. spent millions on an “astronaut pen” that would write in outer space, while the Russians simply used a pencil. Although proven false (both countries initially used pencils), the moral of the anecdote still holds.
So with that introduction…Cherry Advisory is here to make your job easier by:
- sharing knowledge and experiences;
- providing simple insights into advanced analytics, omni-channel, retail, business and IT strategy;
- delivering value; and
- helping you and your organization become as successful as possible.
Thank you for reading and please share your feedback and questions, as I promise to personally respond to every one.