Even in today’s world of multi user role play computer games the Rubik’s cube is still a great seller for Christmas stockings.
Back in 1981, when it hit the market for the first time, a 13 year old boy wrote a book with the title “You Can Do The Cube”. I still have a copy somewhere. In it he provided a foolproof solution to making sense of the 43 billion possible combinations of the exasperating craze of that year – the Rubik’s Cube. To further infuriate those of us for whom the multi-dimensionality of this toy was a source of almost limitless frustration, he went on to describe some of the fascinating patterns you can make with it once you had achieved the simple task of getting all sides the same colour!
Retail finance and merchandising executives today could be forgiven for feeling a similar level of frustration when confronted by the mass of multi-dimensional data provided by today’s EPOS systems. Retail is a multi-dimensional business and the cube analogy is a good one as we try to dice and slice the data that we collect according to time, location and product.
How often do you receive 2 dimensional reports on paper that seem better designed to increase the income of the paper supplier than that of your own company?
How often do you say to yourself “If only I could see this piece of data for all branches at once” ? – or maybe you need to see it for one branch by all periods of the year – or was that all branches together for one period? The permutations are legion and we are only dealing with 2 dimensions here – time and cost centre. Maybe you want to compare value with volume, actual with plan, or how about seeing the index to last year, or the variance. The Rubik’s Cube starts to look pretty easy at this point.
If we take the simple example of sales value in a merchandising report we can see how multi – dimensionality works. We might define sales value by the following attributes : Product (SKU, line, sub product group, department, total), cost centre (branch, area, total), analysis (actual, plan), period (day, week, month, season, year).
Each of these attributes will, in turn, have a series of possible values so branch could be Regent Street or Oxford Street, area could be North or South etc..
The classic approach to providing information relies on either a sort of lowest common denominator method of providing something that might please some of the people for some of the time, or the shotgun method of peppering you with everything you could possibly want if only you could lift the report onto the desk.
A multi-dimensional (OLAP) approach, however, allows the executive to interact with the system to select parameters for reporting without the risk of slowing down all of the data entry clerks whilst the central system compiles the response.
Using the example of Sales Value again, he could ask for sales by department by area, actual and plan for the current season to date or any other permutation from the list and see it almost immediately on a screen in front of him. Seeing an interesting piece of information he might change a parameter to view the same data set, but this time for all branches within an area.
Suddenly the users of the system become empowered, or as one embittered IT colleague darkly observed, “The inmates take over the asylum”.
So far we have only looked at what the dimensions of data mean in terms of access. Do they have any further practical application? To answer that question I will give a brief overview of 2 case studies.
In the first a food retailer replaced a monthly paper based branch profit & loss account report with a networked multi-dimensional database. Executives were immediately able to view time series analyses, exception reports based on values of the dimensions (e.g. branches with wage bill more than x% of sales) and to dice and slice the data in ways that previously called for scissors, glue, and at least a gold Blue Peter badge. Leaving aside the cost savings in preparation time and raw materials, the timely, relevant and accurate nature of the information delivered dramatically enhanced the control of branch expenditure.
In the second case, a clothing retailer wanted to analyze the relative profitability of goods depending on their supplier. By appending the attribute “Supplier” to his style data he was able to cross tabulate the information in such a way that he was constantly aware of not only the level of business that was being placed with each one, but also what gross margin his buyers were achieving. The company’s negotiating position was immediately improved, as was his ability to assess the performance of his staff.
Maximising the leverage that you can get from your existing data can call for some inventive thinking and we are necessarily constrained by the limitations of current technology. We should all be on the look out for tools that allow us to break out of yesterday’s constraints and to innovate. Francis Bacon said “Knowledge is power”. The more we focus our knowledge the more powerful we become.