While the above are in the context of developed markets like the US, its likely that the Indian picture would be no different.<\/p>\n
Historically, most companies have used their ERP systems for generating analytics and MIS. When Business Intelligence (BI) tools were introduced, most retailers simply used the same metrics as in the past \u2013 they merely used BI tools to produce faster.<\/p>\n
\n
\u201cThe difference between ERP reporting and BI reporting is about converting information into actionable intelligence\u201d was Niraj\u2019s take \u2013 \u201cBut, not all retailers appreciated this and did not always have the support of good BI people to help them infuse BI into the DNA of the company\u201d.<\/strong><\/em><\/p>\n<\/div>\nAs an example of this, consider the difference between reactive vs proactive reporting:\u00a0 Most food retailers, have implemented an OTB (Open To Buy) system \u2013 i.e. the system automatically reorders on stockout of fast moving SKUs. A proactive, BI driven OTB system would be able to predict<\/b> sales and move to replenishing on the basis of predictive analytics. Taken to its logical extent, the system would also be able to provide strategic reports of interest to the CEO & the Board \u2013 example – GMROII (Gross Margin Return on Inventory Invested) reporting.<\/p>\n\n
\u201cImplementing a BI system will clean your data by default. Data cleansing is like entropy \u2013 the need for it will ever-increase.\u00a0 A BI implementation is the equivalent of \u2018spring cleaning\u2019 \u2013 in that it gives you a fresh start.\u201d \u2013 Ranjith<\/em><\/strong><\/p>\n<\/div>\nOn the issue of how much BI can help create a \u201csingle version of truth\u201d for key metrics \u2013 the view was that Businesses first need to get to internal agreement on what that \u201csingle version\u201d is. For example \u2013 sales figures are presented differently in the financial system versus the merchandisers system (with the main differences being tax and discount accounting).\u00a0 BI frameworks can present different variants of this \u201csingle moment of truth\u201d \u2013 across the company- but it needs agreement between Finance and Merchandising on how each of them will use and interpret the data.<\/p>\n
\n
\u201cOne of the fundamental building blocks to making BI happen \u2013 People \u2013 is expensive. \u201cReverse skilling someone from the business to become a Business Analytics engineer is a good idea. Over time, the investment in training pays off in terms of how the organizational DNA morphs\u201d.\u00a0\u00a0 \u2013 Vinod<\/em><\/strong><\/p>\n\u00a0\u201cIn\u00a0 Shopper\u2019s stop \u2013 we have a team of people who regularly scan the data looking for patterns \u2013 for example \u2013 one of the patterns we look for is based on how customers complete a \u201clook\u201d.\u00a0 So we can find out all customers who purchased trousers and then send them a targeted mailer for shirts, ties, socks, kerchiefs etc.\u201d was Ranjiths contribution to how businesses can use BI to create those \u201cmoments of truth\u201d.<\/p>\n<\/div>\n
How do you compute ROI on BI?<\/h2>\n
The panel was unanimous that \u201cyou can\u2019t\u201d.<\/p>\n
\n
\u201cIt\u2019s like trying to compute ROI on ad-spend\u201d \u2013 said Niraj.\u00a0 \u201cYou can be fooled by spurts in metrics without having a full handle on all the underlying cause \u2013 effect relationships.\u201d<\/em><\/strong><\/p>\n<\/div>\n\u201cRather than talking about ROI on BI \u2013 consider this question \u2013 how do you justify spend on BI?\u201d. He continued. \u201cIf an investment in BI can bring your shrinkage down by 0.1% – that puts your breakeven at a few months.\u00a0 Similarly, if BI can help optimize inventory costs because of predictive analytics \u2013 then your breakeven is likely to be a few weeks.\u201d<\/p>\n
\n
Wispey\u2019s view was that \u201cthe biggest impact of BI is on availability. How do you measure a return on a customer\u2019s feeling of \u201cThey always have what I want?\u201d. Rather than run after a ROI metric \u2013 look at measures like Forward cover (for stable SKU\u2019s) and a sell through rate (STR) for fashion or other fast moving items.\u201d<\/em><\/strong><\/p>\n<\/div>\n\u201cThe ability to use BI is about BI is a function of business and promoter maturity\u201d \u2013 was Subhodip\u2019s view.<\/p>\n
To the audience question on \u201cHow should BI treat unstructured data such as that from social media feeds?\u201d \u2013 the panelists felt \u201cBefore tackling unstructured data \u2013 retailers first need to harness what they get from their ERP systems.\u201d \u2013 this was Vinod\u2019s view.<\/p>\n
At the same time, there is a technical solution to the question. It involves using tools to \u201ctag\u201d common phrases and then analyse the grouping, frequency and intensity of the tags to arrive at conclusions.\u00a0 However, as both Niraj & Gunjan were quick to point out \u2013 \u201canalyzing feeds from social media would cost you 2 to 3 times more \u2013 so there is a cost-benefit tradeoff to be considered\u201d.<\/p>\n
In conclusion, this is what the panelists had to say:<\/p>\n
Wispey:\u00a0 \u201cWhen I started in retail \u2013 I was taught a few simple rules \u2013 Know your customer better than they know themselves. In merchandise \u2013 it was right product, right place, right price. Retail has grown to the point where many of us have lost touch with the basics. While there is much data, the reality is that the world is now a smaller place \u2013 feedback travels and becomes an international incident before you know it\u201d<\/p>\n
\n
Subhodeep \u2013 \u201cset stakeholders expectations \u2013 define your single version of truth and invest in good resources who can sustain the project\u201d<\/em><\/strong><\/p>\nNiraj: \u201cBuild a project design which is scaleable. Retool the team to understand business drivers.\u201d<\/em><\/strong><\/p>\n\n<\/div>\n