Shelf Real Estate & Data Analytics

In the whole gamut of retail supply chain, there is only one moment of truth! The time when the customer looks at the product on the shelf and decides to buy it. As they say, its all prep work till register rings. Business wise, shelf is the only real estate property for any retailer. A place where the retailer has that unique opportunity to persuade the buyer to buy the product.
So how does a retailer do it? Traditionally couple of thumb rules has been practiced in the industry like putting the most profitable items at the eye level shelves and pushing others to the bottom. However with astronomical rise in the number of items and increase in competition among the brands becoming fierce, it is the need of the hour to assign right shelf spec to the right set of products so as to meet the underlying business objectives. These objectives can be as varied as minimizing the cost, Maximizing the sales, Increasing the assortment or a combination thereof. The skill is in arriving at that right number of items and variety to be placed on the shelves to meet these rules.

Now it may all sound very simple but the complexity of the problem lies in numbers. Considering the number of items that a category manager has to carry, the possible combinations of facings of the items that can be placed on the shelves is mind bogglingly high to be analyzed by a human brain. This is where data analytics can help to process the information at incredible speeds to arrive at the desired solution. This is a perfect opportunity area for human – computer participation where the computer’s data analytical skills supplemented with category manager’s business know-how can do wonders!

Know Your Store

Perhaps every retailer understands that not all their stores are created equal. Due to demographical and many other socio-economical differences, sales pattern varies dramatically. So – why is it that most of the retailers still have one- size-fit-it-all assortment mix and merchandise plan?

The unpleasant truth is, its really hard to intelligently segment the stores into manageable chunks which makes business sense.

Guess what? With advent of technology in the field of Business analytics, it is now well within the boundaries of reality to achieve the holy grail of merchandising. However, to achieve it, there are two fundamental tenets that one must adhere to.

1. Data don’t lie

2. make sure your data is speaking the truth

However much these two may sound contradictory, they are both correct. Let’s see how.

When I say data don’t lie, what I mean is that data is not biased with pre-conceived notions. If intelligent analysis is applied on top of it, it can show key insights. For e.g. – a good data mining algorithm can identify the sales patterns and help group the stores which have similar propensity to sales in the same bucket. The benefit is, such type of analysis is completely unbiased and can at times help in correcting the conventional wisdom.

Now, that being said, it is also important that the data is captured in clinical condition and free of any type of corruptions like negative sales, missing stales information etc. If the data is not cleansed it is easy to understand that all the subsequent analysis will have inherent errors which may grow as we delve deeper. Always remember the GIGO law of computing – ‘Garbage in – garbage out’.

Now assuming that both the fundamentals have been adhered to, there are several exciting things that can be done. for the scope of this article, let’s just talk about store grouping. You can use the data mining packages to identify the group of stores which have similar sales profile. Once you have the list, you can use it for any number of things. Following are some examples that immediately come to my mind:

1. Specific assortment mix for each logical group of stores

2. Localized advertising plans for each cluster. for e.g. – Hispanic heavy ADs in Hispanic heavy stores.

3. Reorganization of distribution and item authorization plans to minimize the inventory and logistics cost

4. Focused merchandise plans for each chunk

It is obvious that the application of such a bottom – up store grouping process is only limited by the creativity of the users. So go ahead and figure out what you want to do with your set of store groups.

Retail Analytics 2.0

Retail analytics is nothing new. However with recession  tightening its grip on economy, retailers are leveraging more and more analytics to gain new insights and identify trends and patterns which define their business.  The idea to gain an ability to make smarter decisions and manage the business more effectively based on data was always an exciting one but recession has taken retail analytics from the realm of being ‘exotic science’ to a ‘must have business engine’ which is part of the survival kit in the current economic downturn.

I call it retail analytics 2.0. In my opinion, 1.0 version was the industry’s first brush with maths and science based data analytics championed by companies like A.C. Nielsen which helped them analyze their data to identify market trends and such using rule base engines. 1.0 was also pretty much adopted by the top tier retail firms  and largely ignored by tier 2 & 3 companies.

However 2.0 has a much wider footprint with small and medium size embracing it with both hands. There are two key factors in my opinion which has led to this change. One, retail analytics technology has matured significantly in last couple of years providing smarter and more intelligent insights thus proving their effectiveness to the business community. Two, the growth and success of early adopters has inspired enough confidence in smaller firms to pursue analytics with much more enthusiasm.

2.0 also means that the nature of firms who are in the business of working on data analytics domain has undergone a metamorphosis of sorts. Big firms are still the leaders but many new nimble players have entered the market with the idea to provide a service oriented approach to data analytics recognizing that each problem is unique and there is no one size fit all solution.

It would be interesting to see how this space further evolves but it is fair to say that retail analytics has successfully made the transition from being a supporting fulcrum to become the cornerstone of the business decision making in retail arena.