18 Oct Podcast: Leveraging Analytics For Omni-Channel Success
Podcast: Leveraging Analytics
For Omni-Channel Success
Integrated Solutions speaks with Mark Krupnik, CEO of Retalon, a company that provides retail predictive analytics solutions, about how retailers can leverage analytics to help with their omnichannel initiatives, including best practices you can apply today.
Play Podcast: (11:45 Min)
Bob Johns: Forecasting, business intelligence, and retail analytics have gone through many stages since the 1960’s back when IBM built Inform. Now, some alterations were only band aid solutions that proved to be temporary and quite ineffective, while other methodologies have stuck around for decades and have been accepted by most retailers today. Analytics in the supply chain have come a long way especially in the past few years.
What are some of the ways that large retailers that you have worked with are using analytics to improve the supply chain today? And how does the technology fit in?
Mark Krupnik: There are several directions in which retailers try to change their approach to business these days. One of them is they try to manage inventory inside, sometimes they use a term, turning all my stores into one large distribution center rather than buying or catching a next purchase order from the vendor. Another thing that retailers are trying to do today is to monitor inventory in different locations.
Traditionally, all the replenishment allocation processes were one way streets. Where people calculate how much they need to allocate, when and how many pieces to replenish to stores, and maybe on sporadic occasions they would recall it back to the distribution center warehouse.
The reality today is they can do much more by balancing inventory between locations by satisfying online requests of customers directly from different locations and that basically allows them to use inventory much more efficient.
Other directions that I should mention here is retailers are trying to establish relations between inventory levels and other drivers such as prices and promotions. It’s been almost a joke what has happened in retail business for years where retailers will declare a promotion or announce a markdown, 20 percent off, and the stores will not have any additional inventory to support it.
So instead of selling more merchandise, retailers would sell the same quantity that they have in store but at a lower price and lose money. With ability to use specific analytics and predictive analytics, they can forecast how much inventory they need at stores in order to satisfy certain promotional activities, media activities, or price changes and obviously take it into account either as a constraint or as additional goals.
Finally, the last thing but definitely not the least, retailers are planning their business way ahead of the time, giving long lead times, giving anticipation of the market they need to plan their business sometimes up to 18 months forward and when the planning is not supported by specific sets of options how it will be executed is just a good intention.
You can plan to grow a business 10 percent but if predictive analytics models are not involved, if you don’t know what will be the changes in store structure, new products, cannibalization between products promotions and markdowns, this planning in again just a good intention that retailers would not necessarily be able to fulfill.
Bob Johns: Interesting. A lot of the times the way most retailers look at their analytics data is misleading and it costs them millions of dollars every single year. What are some of the tricks of the trade on how to not be trapped by common misconceptions, trends, or even too much data?
Mark Krupnik: There is not too much data. It’s really how you use the data that you have. It’s clear that there are certain factors that influence retail process. Everybody knows about seasonality in retail, everybody knows about price.
If you change your price, you may affect your demands. The devil is in the details, the problem is that when you calculate your retail characteristics, they do not exist on their own. They depend on other factors and characteristics.
For example, when you measure your seasonality or your responsiveness to price change, you might be out of stock last year at that time. And if you use these numbers to forecast your next year data, you’ll basically plan yourself to be out of stock next year.
Retailers should really be able to derive from this information, influence of different factors such as low sales, such as assortment wins versus assortment debts, they should account for cannibalization between the product’s affinity and substitute. The effect of different promotions and a lot of customer-specific factors like we have for example a company who runs university bookstores and apparel through these stores as well, and they obviously have their own specific factor — such as enrollment rate of students or beginning of school year. Another company in fashion that specifically plans to have terminal stock at the end of the season. So how can we account for all these factors?
Basically the idea is that predictive analytics enables something that was the idea for the retail world for years to have one integrated solution. When you change the promotion, the system should know about your inventory level.
When you plan your inventory levels, you need to know what is the cannibalization between products. Which products will replace each other, what is the assortment, when you do any forecasting or inventory transfers for purchases, you need to know the prices.
Unless you have one integrated predictive analytics platform, it’s very difficult to calculate each of the factors separately. But once you do have it all integrated on one platform, it suddenly becomes very simple because the number of exceptions, the number of errors goes down.
You don’t need to look at the exceptions of each factor separately anymore, the system knows about all these factors and automatically adjusts and calculate their proper balance and interrelations between the [inaudible] [0:07:19.1]
Bob Johns: Perfect, now e-commerce has been a large disruptor in the retail supply chain. And as more retailers sell in multiple channels and try to become omni-channel, they continue to struggle to create a seamless experience for the customer. How can a retail supply chain help push retailers into an omni-channel reality?
Mark Krupnik: Retailers sell their merchandise through multiple channels. Today it maybe is the web and yesterday it was a catalogue. It doesn’t matter, the idea has always been the same. To have the right product, at the right time, at the right place. And the right time really means just before a customer comes to buy it.
Predictive analytics actually does that and not only that. Please be clear that in general terms, distribution network includes not only stores or warehouses and distribution centers, but it prolongs all the way to third party logistics, vendors and customers.
You will find a story that when installing an inventory balancing system at one of our jewelry retail customers, they decided to pilot the system. The very first report had suggested to move a sufficiently large diamond to a new location. Next day, I’m getting an email from this retailer saying that two hours after the diamond was moved to a new location, the customer came and bought the stone.
Overall however their results were not less impressive. In three months our customer reports a 20 percent increase in sales with almost 40 percent reduction in inventory. This is the kind of a result retailers can expect when they manage the inventory distribution among the omni-channels using predictive analytics.
Bob Johns: That’s an amazing story. Now what about the future of analytics in retail? What do you see changing in the next 5 years and what do you think will be around for a while?
Mark Krupnik: My belief is that different solutions will merge. We observed that the large players such as SAP and Oracle and JDA, they have a number of solutions that represent different sides of retail such as inventory management, pricing, promotions, assortment planning, so all the technology if I see it correctly, moves in the same direction to create something that is called integrated solution for retailers. So interestingly enough your magazine is called Integrated Solutions for Retailers. So you guys foresaw this happening years ago.
Predictive analytics has finally made it possible and this is what we offer to retailers today. It looks like we are all on the same right path especially given the recent Gartner report stated that by 2016, which is just three years from now, 70 percent of the most profitable businesses will manage their processes using real time predictive analytics.
Bob Johns: Perfect! Well Mark, I want to thank you for joining us today.
Mark Krupnik: Thank you very much for the invitation.