BrainTrust Query: Retailers Should Be Fast ‘AND’ Data-Driven

Through a special arrangement, presented here for discussion is a summary of an article from Applied Predictive Technologies’ Retailing with Confidence blog.

Recently, I had the pleasure of reading Jim Collins’s new book, Great by Choice, which prompted me to re-read his classic, Good to Great. In re-reading Good to Great, I noticed that many of the ideas that Mr. Collins espouses are applicable to retail organizations today.

One idea, "the genius of AND over the tyranny of OR," is particularly timely. A common refrain from senior leaders is that in today’s chaotic environment, there’s not enough time for structured decision-making; there’s not enough time for a test to complete before deciding on rollout. In other words, executives often settle for "we can be fast or data-driven … but right now, we just have to be fast."

This idea strikes me as poor thinking in two ways. First, decisions driven by "gut" or anecdotal evidence are more likely to be incorrect. In this "tyranny of OR," retailers are both more likely to make the wrong call, and doing so more rapidly. This is a double whammy which compounds small problems into big ones.

Second, I fundamentally disagree with the idea that you can’t have both fast decision-making AND data-driven decision-making. The solution to "fast AND data-driven" decision making is planning. To have the right answer in March, you need to be testing ideas in-market in January and February. To have the right answers in August, you need to be testing ideas in-market in June and July. So on and so forth.

Long-term planning around supply chain and merchandising issues is a discipline in which retailers excel in some areas. Retail organizations routinely order product six to 18 months in advance.

Testing to drive innovation is no different.

One related and pernicious argument against planned-and-orderly-testing is that the business environment is changing too quickly, and that today’s results aren’t useful tomorrow. Doubtless, trying to predict the general macro-economic environment is frustrating and challenging: Will Greece eventually default and drag down Europe and the global economy? Will unemployment recover? Will there be any "shocks" that drive up fuel prices and drive down consumption? I don’t think anyone really knows. However, I do know exactly which solutions retailers will rely on to overcome any challenge: merchandising mix, pricing and promotions, marketing message, marketing vehicles, customer relationship, etc. In other words, no one can predict the problems tomorrow will bring, but we know exactly what’s in the toolbox to fix these problems. Isn’t it beneficial to sharpen our tools for whatever problems the future might bring?

BrainTrust

Discussion Questions

Discussion Questions: Which of retail’s disciplines — merchandising, marketing, pricing/promotions, labor, etc. — could most benefit from a more disciplined planning process? What steps should retailers be taking to ensure more data-driven insights can be tapped in today’s frequently rushed planning process?

Poll

18 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
David Biernbaum
David Biernbaum
12 years ago

Sometimes data lacks foresight, vision, and forward-thinking. Steve Jobs didn’t rely on yesterday’s data to create some of the most important new consumer products this world would ever known. That said, solid executives in retail and inside brands are extremely well staffed and networked with people they can go to in a rush to help them make fast decisions based on experience, insight, and intuition.

Ben Ball
Ben Ball
12 years ago

I’m afraid the real answer to this question is all too obvious. Real planning takes effort. Defaulting to “we have to make a decision now so we’ll go on experience” gives executives two deadly, but desirable, benefits.
1. It is easier.
2. You can do whatever you want to. Data be damned.

If this sounds like the voice of experience speaking — well, it is.

Dr. Stephen Needel
Dr. Stephen Needel
12 years ago

All of these can benefit from a better planning process. Over any rolling 12-month period, nothing changes very much; therefore the need for immediate action is minimal. There is no excuse for not planning ahead, analyzing the data, testing, and so forth.

Ryan Mathews
Ryan Mathews
12 years ago

Data is historical and history isn’t always the best tool for forecasting, especially in rapidly changing times.

That caveat firmly in place let me take issue with the question.

Why does the question assume that these areas, (along with things like real-time social media input) shouldn’t all be integrated into one seamless, real-time, business intelligence driven retailing model?

By continuing to look at these areas as though — at their core — they are too distinct to be integrated, we automatically limit the potential value of any data.

Integration would eliminate some of the frenetic aspects associated with today’s decision making process.

But we also have to find new ways of filtering, mining and reorganizing data — otherwise even the kind of integration I’m suggesting will sub-optimize over time.

But another way, it’s time for retailers to make that big scary jump from the 17th Century to the 21st — before it’s too late.

David Slavick
David Slavick
12 years ago

The answer is — ALL. All can benefit from discipline, not just more discipline. The push to achieve immediate results vs. test/learn is endemic to today’s retailers. Powering the data is a theme we take very seriously and many of the consultations are focused on best methodologies to build plans, source analytical tools/models to gain insight and predict response leading to communication plans that segment, allocate resources/offers and vary the delivery frequency/stimulus according to customer potential or real worth. All of this requires cooperation within and without the organization and its not easy.

Merchant Planning and Analysis traditionally have the greatest discipline, but in some organizations they are operating off of tried and true methodologies that could benefit from…let’s call it a re-fresh. They are loathe to change, but can — in moderation. If there is a 2012 New Years resolution to be made, it’s this — commit to test/learn and take your time to read the results, refine with hindsight and in the end impact on EBITDA will be significant!

Matt Schmitt
Matt Schmitt
12 years ago

Huge amounts of data are being gathered, but maybe the problem is still a sorely lacking emphasis on pinpointing the valuable streams of data to tap for decision making. Is the old problem of “data rich, information poor” actually increasing in complexity? Unfortunately, probably so more often than not.

Ben Sprecher
Ben Sprecher
12 years ago

I started out thinking I was going to agree with this article, but I got snagged on the line “The solution to ‘fast AND data-driven’ decision making is planning.”

If the author is saying that taking an agile test-and-learn approach today allows you to make better big decisions for the months ahead, then I agree. But “planning” is often the culprit in stretching tasks into projects and projects into “strategic initiatives”.

A healthy business will certainly be data-driven in much of its decision-making, but it needs the technology to make the data available quickly, easily, and broadly so that it can support business agility too. But, failing quickly and inexpensively on a small scale can be more important to long term success than “getting to the right plan” up front. Certainly in software engineering, the trend has been away from “waterfall-style” project management and towards fast, iterative, “agile” development, and the resulting improvements in productivity, quality, and innovation have been enormous.

Raymond D. Jones
Raymond D. Jones
12 years ago

The problem is not so much with the data as with the implications and decisions drawn from it. People often make the mistake of selectively accepting or rejecting data based on how it supports their opinions.

Retail predictions are further complicated by factors like seasonality, geography, and demography as well as shopper psychology. The fact that an item sold to young women in Philadelphia in September may or may not predict that it will sell to a similar group in Florida in November.

Most data simply provides hindsight. The key is turning it into insight and foresight. That requires a recognition of the types of data available and a keen sense of how to apply them.

Camille P. Schuster, Ph.D.
Camille P. Schuster, Ph.D.
12 years ago

Saying we do not have time to use data because we have to make decisions quickly is an excuse. It is possible to do AND is evidenced by some retailers who are constantly working to improve their processes.

To be successful, historical data as well as current consumer insight data needs to be gathered and analyzed to inform decisions. Getting data into a flexible warehouse, creating relevant algorithms for quantitative data, and incorporating qualitative data into the warehouse and algorithms makes it possible to have the perspective and wisdom of the past combined with the identification of trends when making decisions about the future. This process is not easy and requires time to get all the parts in place. The organizations that do this have an advantage over their competitors.

Ted Hurlbut
Ted Hurlbut
12 years ago

I agree completely with the points that Dacheng makes, but I would extend them. Yes, planning is essential, but planning cannot be static, it must be dynamic. Success requires a culture of continually assessing progress against benchmarks and updating forward benchmarks. It requires a commitment to continually assessing data, and keeping the benchmarks relevant. Second, planning is not just about what you expect to happen, but it’s also about what you will do. I’ve long believed that as important as any plan is, how you plan to execute that plan is of far more importance. This relates directly to the supply chain. More important than determining how much merchandise needs to come in during a given period is determining the schedule for actually committing to that merchandise, how many dollars to commit upfront, and how many dollars to release at key points thereafter.

The choice between being fast or data-driven is a straw man. The challenge is to be fast and data-driven, which means the data must be immediately at hand, with benchmarks for evaluating that data. To be merely fast is a rationalization for being lazy and flying by the seat of your pants.

ron kurtz
ron kurtz
12 years ago

All retail disciplines can benefit from a strong planning process that anticipates the need for change rather than reacting to change.

I am surprised by the apparently limited use by retailers (and their vendors) of survey research that can provide insight into the mood and interests of the consumer.

Good research, when interpreted through the lens of experienced and knowledgeable executives, can be a very powerful tool to achieve success.

David Zahn
David Zahn
12 years ago

I am of the opinion that decision-makers struggle with finding the balance between strategy, tactics, planning, and executing. We can all agree that you need more than just one of the above in the recipe. The question soon becomes, how much of any of them and in what order?

Given the fast-moving times we are confronted with, I often rely on two thoughts: There is never time to do it right the first time, but we find time to do it over; and progressive performance over postponed perfection.

If we do the right things right initially — even if not perfect — the inertia and positive movement will aid us in making mid-course corrections. I am not one for dipping my toes in the water when I am looking to make waves. I opt to splash a bit, but at least do it in the water and not try to do it in the office board room poring over spreadsheets. I will look at the tides, maps, temperature — but then choose to DO SOMETHING and accept that there may be changes needed.

Warren Thayer
Warren Thayer
12 years ago

Pricing/promotion is the lowest hanging fruit here. Back in my days as a corporate ad director in NYC, I had a sign in my office that said “Cheap. Fast. Good. Pick any two.” That stopped a lot of plans for 100-page four-color catalogs by tomorrow afternoon at 2 for $25.” To a degree, that’s still true. But integrating data shouldn’t be rocket science. Using data — at least data that has been shown to successfully support correct decisions — should be a seamless part of the plan.

One reason I’ve loved being a volunteer firefighter for most of my life is that you simply cannot take the time to get bogged down with minutae. The first-due engine has to radio back immediately and say what we’ve got (i.e, fire showing out of first floor window on side B of three-story balloon frame structure), where it’s going (danger of spread to attic and adjacent home), and what you need to control things (send us a tanker truck and one more engine with foam.) You could micromanage any of of these things and get stuck in analysis paralysis, but the house would burn down. With a good plan and quick response integrating all basics, you’re in a position to win. I’ve always found the “what have you got, where is it going and what do you need” model really helpful in just about anything, no matter how complicated.

Jonathan Marek
Jonathan Marek
12 years ago

To Ben Sprecher’s point, an “agile test-and-learn approach” is exactly what is needed. You are right that old-fashioned project planning isn’t what Dacheng means. A fast technology-enabled, test-driven approach to BIG decisions gets these decisions made both quickly and profitably.

Ralph Jacobson
Ralph Jacobson
12 years ago

Using the right tools can enable fast, data-driven insights and decisions. The issue here is “big data,” I believe. All of the key business function can benefit from a structured planning process. Companies, CPG and retail alike, can utilize capabilities employed by leading firms around the globe.

The other challenge is addressing “long-term” planning within retail. As we all know, retailers, especially food retailers tend to live and plan by the moment. Just make it through this quarter, this month, this week… Black Friday.

There are great templates available to guide the “tactics” of strategic planning. Leveraging big data insights to drive the planning process needs to become more prevalent in our industries in order to capture the literally trillions of dollars left in stock-piled inventories worldwide.

Scott Welty
Scott Welty
12 years ago

I think the comments posted here are spot on. Like the saying goes, fail to plan…. A good, structured plan is pinnacle to the retail life cycle, this defines what good looks like. Once the plan is in place, retailers need strong demand sensing solutions that can interrogate demand signals from various channels, customers driven by varying pricing and promotion strategies and make educated predictions on future demand. In today’s volatile and fickle retail world, the retail discipline that can benefit the most is not listed here: customer satisfaction. Listening to your customer, analyzing demand signals combined with timely and accurate behavioral predictions capabilities that enable all retail disciplines will yield the best result of all. A loyal fan base, that is open and willing to continue the conversation about you.

Ed Dennis
Ed Dennis
12 years ago

Failure is often the end-product of analysis. You can’t drive a business to success with data. Data will, if used properly, increase your opportunity for success primarily by making sure you have the inventory the customer wants. But success depends on the interaction between the retailer and the customer. All of the “great” retailers (e.g. Nordstrom and Amazon) emphasize exceptional customer service. None of the great retailers ever involve themselves in gotcha or conditional service. They strive for one thing: “an exceptional buying experience.” Data can help but attitude is much more important. Given my preferences I would concentrate efforts on exceptional sales associates.

Ronnie Perchik
Ronnie Perchik
12 years ago

All of retails’ disciplines can benefit from planning, and I agree that it’s not impossible to incorporate data with speediness; it’s called quantitative reasoning.

How about utilizing digital media to blast your marketing message as a way to combat this? Digital is a relatively inexpensive, and certainly fast (almost instantaneous) media that can, if tracked and analyzed properly, can yield tremendous info on consumer behavior, product likeability, etc.

In my opinion, incorporating digital into your offline efforts as a form of market research, and ultimate launch of the message, can really help marketers and retailers in this predicament.