When someone says “this store has everything under the sun!”, of course it doesn’t. It happens to have the exact things they are looking for.
How did it achieve that?
Perhaps an astute storekeeper has observed the preferences of the customers who frequent that store, and has come up with clever ways of stocking products and managing display. It’s working so well that he decides to replicate this model in his other store as well. Somewhat frustratingly, the other store does not do so well. The same product stock, the same brands, the same display, but what works in one store doesn’t work in another.
Being the astute businessman that he is, he understands that he is missing something. His analysis is missing something. The data he is analyzing is missing something.
So he looks at the data. He is analyzing sales based on products. His collection of women’s evening wear in Store A is doing really well. So he orders more stock. It continues to do well. He orders even more stock and populates Store B with the same great collection of dresses.
It doesn’t sell.
Closer examination reveals to him that it’s not just evening wear that’s selling well in Store A, but a certain kind of evening wear – specifically, leopard print dresses. It so happens that all the stock he has ordered is of this kind. Customers of Store B, for some reason, are not attracted to leopard prints. Maybe Store B would do well if it stocked up on polka dot patterns. Or maybe it wouldn’t. How does one stay ahead of the game, or at least keep up with it?
The answer lies in a well-designed AI algorithm that is fed attribute-rich data. So your data on dresses would have – in addition to standard details like sleeve-length and collar type – free-form text that can hold the type of print, buttons, texture, or what have you – just about any attribute that the astute storekeeper may want to put in. The more he puts in the better the algorithm puts out. From an overall sales snapshot for dresses, to the type of dress that is doing well, and at which location, with the ability to drill right down to the type of print that’s doing well – this is the kind of AI that can provide insights to make best use of data. We are moving from looking at products as whole units, to breaking them down into high-level features, breaking them still further down into some very specific attributes that are catching the customer’s eye. We can do this faster than the customer’s eye moves on to the next interesting thing. We can see what this other interesting thing is, and how popular that is becoming. We can see graphs that show how trends change, and where. We can do this over time, across stores, and for a huge amount of data. Not just any old data, but data tagged with attributes. And that is how the astute storekeeper stays ahead of the game.