A Watson for retail may not be as far away as you think. Intelligent, automated retail technology is actually here, and some of the smartest retailers are already using it.
LISTEN TO PART 1:
speaking with Ashley Goetz, Director of Marketing and Social Media • April 2011
Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.
Audio Transcript:
I’m Dan Brown, from Mulberry Marketing, San Francisco, and I’m here with Ashley Goetz, Quantum Retail’s Director of Marketing and Social Media.
Thanks Dan.
To get a little more of an idea of the advancements some retailers are taking advantage of, can you give us the inside scoop – who’s ahead of the game?
Some of the biggest names in retail, like Guitar Center and Marks & Spencer.
Why aren’t more retailers taking advantage of the most advanced technology on the market?
That’s a good question. But mostly it’s a fear of a something new – a different approach, but in reality – it pays for itself in just a few months…
Months?
Mmm hmm. See – when you are a supply chain behemoth and you reduce your overstocks by say 20 or 30%, that can pay for itself almost immediately. And then the long-term benefits are just plain no brainers.
Because it’s predictive?
Mmm hmm. And those predictive analytics are actionable and automated. In fact automated technology like this is so efficient that it can take in real-time data from every SKU in a chain, perform predictive analytics on it, and automatically execute optimal decisions on it in –every– location.
Every item – in real-time?
Yes.
That completely streamlines merchandising doesn’t it?
Absolutely.
What your saying is twenty years beyond what most retailers are doing today.
Well once retailers take the time to understand what our solution really does, we usually jump out as their vendor of choice.
So what are they using instead?
Well typically – it’s old school technology that they spent lots of money on. These systems have become their foundations – and many retailers feel more comfortable sticking with the same old thing, or they feel like they can do it themselves.
But much of those foundations were created in the 80s or 90s – right?
Right – and they continue to slap band aid fixes on them.
I remember technology of the 80s, computers with the blue and white screens…
Like on Dougie Houser.
Yes – and Atari.
Yea –you could ping-pong across the screen and even that was slow.
And then Al Gore invented the internet.
Ha ha.
And pagers were a huge advancement.
Oh yea – and brick cell phones.
And that is the type of technology that was around that laid the foundation of the big systems out there?
Yes – well when retailers paid so much money to buy or build a system – they continued to just add on updates to user interfaces in the nineties – but it doesn’t change the premise of the system. You basically have a big spread sheet of data – and the system lets you enter parameter or rules and algorithms in it – and it runs those algorithms on the timeframes of data that you tell it to. But you’re still trying to deal with every item and every store. Manually…and that’s impossible when you have thousands of stores and hundreds of thousands of SKUs.
Yikes.
And so today – you hear a lot of people talking about this concept called localization. Basically meaning that every store would have its own assortment.
Right.
People don’t have the same tastes, cultures, trends, sizes, or habits in every area. Even just one example – think about high schools – Let’s say you’ve got a high school whose colors are purple and blue – and you’ve got your highschool girls going out and buying out all the XS tank tops and pants and shoelaces, and eye shadow and glitter – you name it – in those colors – well that’s just a high school, and that’s probably just a few stores of your chain that would have high demand on those colors.
Right.
Well you think about it on a larger scale and every region has its own trends and neighborhoods and cultural pockets with their own preferences –they’re all unique. Every store is unique. But the closest to “localized” that these types of systems can get is generalized clusters, unless you can afford to manually manage every store or every department. And the average large retailer has near a thousand stores and hundreds of thousands of SKUS. Even if you assigned one person to every store – they’d have hundreds of thousands of SKUs – or if you assigned someone just to one department – they’d still have thousands of SKUs behaving differently at thousands of stores. And they can’t even afford that many merchandisers.
So typically what do they do?
They generalize. They put those stores into clusters, usually based on size or location, and they treat them like one store. They may find a few things within each of those stores that they do localize on – like sporting teams, or gifts right. And some retailers are good about that – like at Kowalski’s Super Market you’ll see trendy local magnets with Uptown and Lake Calhoun, and other Minneapolis locales.
Right.
But it’s that general merchandise and those normal products that are so difficult to predict. You look at the recession – and you’d expect people would downgrade to lesser brands right?
Right.
But they’re not, they are choosing what they feel is the most value – for the lesser price – which is typically the trendy item that’s on sale. But they won’t stick with it, they’ll follow the sale. In most cases they’re just buying less. They’re being smarter. Organics are increasing in sales – even though budgets are tight. Gas is going up and less dollars are going into stores – but what dollars do go into stores are constantly changing. Because you still have your marketing team putting out promotions but you then you also have people socializing products – like on Twitter – and Foursquare and Facebook – you have the potential to get what seem like random surges. Or you have some celebrity like Charlie Sheen wearing Nike on Sheen’s Korner – and who knows what that does to demand – and you’ve got Groupons skyrocketing business – but how long do those effects last? It makes consumer behavior really unpredictable for traditional systems.
Are retailers starting to get it?
I think so – it’s beginning to hit home more and more. But they don’t know what to do. It’s overwhelming. There are so many channels to monitor. But to put it really basically, you have your stores-and-you have your e-commerce.
Right.
And your e-commerce maybe is a little more complex for distribution, but really it’s a lot easier – it’s like one store to manage the demand. So if you can forecast that correctly – you shouldn’t need to worry – but most retailers haven’t been able to.
And you still have online products out of stock – that doesn’t make sense – but it’s because you pushed all your inventory out to the stores – and didn’t think your online demand would be so huge – and now you’re waiting for your vendors to come through. Or vice versa – you launched a product online first to assess how it would do and now you have none of that product to push out to stores!
So you have all these channels, and all this data, and how the heck can you analyze and predict it and execute it when you have thousands of stores with hundreds of thousands of SKUs?
Right. And lately I’ve seen publications calling for a Watson in the retail landscape – and that’s essentially what is necessary in this new environment isn’t it?
It is, and that’s what Quantum has created.
That’s amazing.
It is. You see — all that old technology that was built in the eighties and nineties or even just after the turn of the century– that’s old school.
Old School.
Right. And we’re different. And everyone says that. But we are. We’re New School. Ha.
I like that.
Really – you should talk to some of our employees – we know how to have fun. But really we know what works. And the old stuff doesn’t kick it anymore.
Not in the least. And your founders got that.
They did. They said let’s throw out the box and build something new. And smart. And they are some smart cookies. Retail scientists. Doctors. Seriously. And they got together and built a new school solution, called Q that wraps up BI, continuous learning, and automation, what some would actually call AI.
Artificial intelligence.
But that scares some people.
But it won’t for long – it’s necessary today. That’s where technology is headed.
Exactly, it’s not like Google or Amazon have people back there recommending their favorite books to you when you buy something – the system analyzes patterns.
Right – and those patterns are continuously evolving.
Exactly – and that’s the sort of thing you can do when you build a new solution from the ground up on technology of this era – I mean they were building the first iPhone when Q was being developed – that’s what New School is.
But it’s really more than that too – because Quantum looks at the business from a whole and asks your customers to re-evaluate how and why they merchandise.
You’re exactly right. When you meet with Quantum, we don’t shove our name and our list of clients down your throat. We ask you what your problems and opportunities are – and we help you build and define a strategy for your merchandise. Every item needs to have a role and a goal. That’s what makes Q so unique. Our Watson doesn’t just keep up with demand – it keeps up with each item’s strategy.
Right – go into that further.
So every product is given a role and an objective, like being a traffic driver or profit generator – they are also given constraints – like minimum presentation requirements, etc. And the system monitors products and their strategies to make sure that they are –always- upheld. That’s why we say our system is profit-driven. It’s not just about keeping items in stock – it’s about keeping up the strategy. And sometimes those strategies change. But Q monitors that and it will show you –based on how products are behaving in each location – what the most profitable strategy is – you may only be keeping a certain amount of stock of one product – because it’s seasonal – but it’s performing like a basic – or you may be stocking on a typical size ratio, but this store has it’s own size ratio. Or maybe there are economic changes, or weather changes that you didn’t forecast for.
You have a truly unique story about that.
Yes – there are always signals –
-like when all the animals run from the forests before earthquakes…
– ha ha – sure – see Q can track the animals. Just kidding. But there are spending habits that are widespread – and when the recession hit – for example, Q told Guitar Center that it needed to cut 90 million dollars of inventory out, 2 months before the stock market collapsed. And they were really confused – they thought there was a glitch in the system – but no – Q recognized that something was happening and they had to get on the phone with their vendors and let them know, because these are high ticket items that are typically made when they send their order quantities. So luckily they had enough warning. And while other retailers suffered heavily from the effects of the recession – Guitar Center was able to minimize it.
That’s amazing.
But unexpected events bring unexpected demand. And Q responds immediately. Q also calculates out of stocks and even lost sales. So when you run out of an item, Q looks at where you are in the demand curve, how many days you are out of stock, and then knows how much money that out of stock cost you. It also keeps that learning so it can be used when other, similar situations happen with similar products in similar locations. But that becomes much less of an issue when the system ensures you always have the right amount of inventory in every store.
So what makes people afraid of a solution like this?
Well, one qualm that people tend to have with the term automated – is that they have to trust it. And when a system tells you one thing, it’s hard not to feel that you as a merchandiser know better than the system. And you go can go in and tinker with it however you want. But New Look, one of our UK fast fashion clients, has said that once they started just leaving Q alone – it worked even better. So there’s a trust issue there. But in reality – I don’t think most merchandisers want to be mathematicians and detectives for thousands of products. Especially in food stores – where the type of demand per product really changes from location to location. So if the system can find the patterns and show them to you – or even better – just respond to them every day – which gives store owners and vendors a sigh of relief, and lets merchants focus on the true art of merchandising – understanding what products people like – what products sell well – finding the new trends – promoting new products – and staying fresh – and that’s what Q is really good at.
So when you have a new Omega 3 fortified organic tomato sauce, or you’ve just added tablets to your electronic assortment, or you find a really great willow grey silk tank – you know, one with a twisted neckline and a few ruffle folds – but not ruffle sleeves…
Oh no, just a little detailing.
Right – Well Q will tell you where items similar to that product sold well in the past - and where they didn’t – it will assign a profile as near to the attributes of the product as possible – it will look at the life cycle of the product, the demand curve, the markdowns, and over 30 other characteristics – so you can make the best decisions on new items, and the most exact decisions on basics and general merchandise – but Q won’t stick with the past, it will learn from every new pattern of every SKU in every store, and every channel. Which you think would make it slower – but once it learns from the data – it doesn’t need to hold it any longer. It tracks demand in real-time because it knows how products behaved in the past – and when those circumstances arise in the future – it has already learned how to respond. Does that seem scary to you?
No, it seems incredibly efficient.
Right, I think that it lets the merchants be better merchants. And it actually increases the art or merchandising. Because the art is in the merchandise strategies, and goals, in the quality of the products, the brands, the colors, style, fabrics, fit, flavor, and everything else that makes up the art of retail. And Q works well for all categories from fashion to food to general, and specialty merchandise. Because customers tastes simply change from store to store. And they change daily. Keeping up with that – is a constant feat – but that is what Q is made for.
So in a nutshell, Q helps retailers do more with less to scale their people, processes and systems, translating data into actionable knowledge and automatically aligning execution to strategic goals.
But that’s just marketing speak –
It is –
- that doesn’t mean anything.
Haha…
The truth is in the results. Why don’t you go over some of the typical results.
Typical results:
· A 2 percent full-price sales increase (in fast fashion)
· A 6 percent sales increase (in general merchandise)
· A 4 percent increase in gross margin
· An 11 percent inventory reduction
· A 40 percent reduction in overstocks
And that’s just in a few months. Unlike other systems that degrade over time – Q just gets better. It evolves – and learns – and as Quantum continues to develop the system – you benefit from it.
Those are amazing results. Thanks for sharing. Well there you have it – the Watson for retail – is already here – and it’s called Q!
Thanks Ashley.
Cheers!
Join us next week when we’ll be talking with Irene Messier, former SVP of Planning & Allocation at Guitar Center.
Download a PDF of This Retail Life: Changing the Game in Retail»
