Small Business Trends: Maybe you can give me a little bit of your personal background.
Guy Yehiav: I have about 22 years’ experience in supply chain applications. Built up a company called Demantra that does demand management, trade promotion management, supply chain optimization. Implemented a lot of different projects with different types of companies and industries; retail, CPG, complex manufacturing, consumer electronics etc. Then I sold the company to Oracle and was the head of supply chain strategy for four years. And then I went back to entrepreneurship which I love.
What we did here with Profitect is try to be a different software company helping our customers in a much more efficient way and be much more effective, bring them live quickly, and focus on their problems mostly.
Small Business Trends: Everywhere you look today we’re talking about analytics, machine learning, artificial intelligence, deep learning. What you’re talking about is prescriptive analytics. Can you tell us exactly what that is and how that may differ from some of the other areas that we’re hearing about?
Guy Yehiav: All of those technologies that you just illustrated, Gartner aggregated all of them and said those are smart machines. They ingest data. They run different mathematical models, different dimensions and then they spit out stuff to generate value. And if you talk with customers they say hopefully it generates value. Sometimes they don’t. Sometimes I need to hire a lot of talent to then generate the value out of those smart machines.
All of those smart machines at the end of the day generates typically a report, a title of a report, a workflow sending those reports etc. Prescriptive analytics is based on all the years of working with a lot of different types of companies. We’ve learned that we would like to democratize the data differently. Giving not the data itself to a lot of users and people, but guiding them on the actions. And this is prescriptive analytics.
We have a smart machine that ingests data as you can imagine, use machine learning algorithms to automatically cluster things based on behavior, looking for anomalies etc. etc. But the beauty of it is when we find the value, what we call the opportunity, we then send it to the right users using workflow with a descriptive insert in plain language. So we tell them what they need to know and then we also prescribe an action; what they need to do in order to present an optimal outcome.
Small Business Trends: So it’s like you find the insight, and then not only go over the insight but deliver the next best action.
Guy Yehiav: And it could be multiple steps. Let’s say that we identified that the hottest Halloween costume is being sold in a specific store on an average every four hours. If it’s been about 8 or 12 hours since this costume was sold, and it was sold in other places, the engine will identify what is the probability of that costume not being available for consumers to pick it up from the shelf. That’s ultimately what the statistics tell you. In the next four hours you should on average sell at least one. If it’s been eight hours — ‘hey you’re lagging’.
Now, since we triangulate all of the information the machine knows … you do have it on hand. So how come you have it on hand? Other stores are selling it. You used to sell it and suddenly you stopped selling it. There’s a probability that it’s not available for consumers to pick up and so the action would be very simple — ‘Hey Mr. associate this specific costume you do have on hand for some reason you’re not selling it. Can you check that it’s on Aisle eight, shelf three? And if you don’t see it on the shelf, you do have it in the back. Go and move it to the front.
Now that allows people that are not analytically savvy to generate value using a smart machine that is very analytical and relies on statistics. But the outcome — the user interface, is in layman’s terms for everyone to understand. It will tell you what [is] the action you need to do and what’s being said.
Small Business Trends: We’re coming up on Halloween which is a big season for people to buy things from a retail perspective, with about seven billion dollars spent annually on Halloween related products.
How does using prescriptive analytics impact a retailer who’s trying to sell right now, versus a retailer doing things without having all these insights and next best actions available to them?
Guy Yehiav: When we speak with the store managers during the year, they always tell us that they’re getting tons of reports; shelf availability reports, average price reports, labor spending reports, based on budget etc. When they get all those reports, they either go in the back room and start analyzing it or they’re asking … different associates to analyze it and they will come with ideas on how to solve them. What happened is now you’re spending time in the back of the room and what you’re not doing is you’re not converting traffic to happy customers. Retail traffic is a major term that is being heard all across the net because you have more and more online traffic or online consumption of goods. And then you have people going to the malls and going to the stores which everyone is saying over the last four or five years traffic in malls is going down.
However, when you look at analysis, most of the traffic going into stores go with a better intent. When someone goes to the store rather than buying online they have an impulse buy. They are going to buy something, and you can even convert an upsell even more using impulse buy. And so if you do have enough people in the store to give the right service, rather than read reports in the back room, you actually have a better way of improving your sales.
More than that, if you use prescriptive analytics you’re not relying on talent. You will have better service and you will have better on shelf availability.
During the holidays and especially on Halloween, there’s a lot of pop up stores. Usually the retailer is just renting a certain square for about a month or two months for the whole of the holiday season and then they hire temporary labor in order to sell and service the customers.
Now the temporary labor is not familiar with the culture of the specific retailer. And you can imagine that different retailers, like any company, they have mission statements and values and they have different type of cultures that makes them who they are. So if you have a temp, they don’t know retail, they don’t know what is on shelf availability. You’ll give them a report. They will not know what to do with it.
The whole idea here is to just tell them what the insight is and what they need to do, because they do want to do the right thing. So now during the holidays it’s even more important, because you want to convert the traffic if you want to be able to create the online orders as fast as you can. You want to service the customer the best, but on the other hand you continue to get a lot of those reports. And so prescriptive analytics during the holiday season is even more critical than the average day of the year.