What is video analytics Coffee and Conversations by Cisco

Show video

Good morning, good afternoon, good night.  Thanks for joining us on today's podcast   of Conversation and Coffee with your  co-host Gary Senna. And I am Danny   Vicente. Thank you for listening whenever,  wherever, and however you are joining us. Conversations and Coffee is the place where we  share a cup of coffee and allow our curiosity sit in the driver's seat and explore topics in your industry. Everything from technology to leadership,   to innovation, and so much more. So  grab your favorite cup of coffee,   sit back, laugh with us while we dive  into the topics. Keeping you up at night.

Well, good morning, good afternoon, and  good evening. Thank you for joining another   conversation and coffee. I am your host  Danny Vicente, and as mentioned last time,   we do have a new host on the podcast.  This is her second go round. Sabrina,   why don't you say hello to everyone. .

Hey y'all, thank you all for joining us today. We're really excited to kind of  go and dive into what I do for   my day job at Cisco, which is on the retail side. I love it. I love it. And some of you may,   if you are watching this podcast  on YouTube versus just listening,   once you hear his voice, if you're listening,  you're gonna know exactly who this man is. We are joined again by Mark Scanlan, who  the infamous Mark Scanlan is joining,   and we have a new voice to the podcast,   and I'm gonna let Rafi introduce him.  And, and what brings you to the podcast? Sure. My name is Rafi Vartian. I'm Vice  President for Business Development at meldCX.  

We're a partner on the computer vision side with  Meraki specifically on the intelligent cameras. We're on the podcast because we  focus predominantly in retail   for our verticals. And I've got a  background in it and I talk a lot   for a living. So I think that's part of  it as well. I'm pretty used to talking , you, you and I both. And Mark, actually,  as we're going through introductions, even   for those folks that do know you, why don't we  remind them what you do and, and why you're here. Sure. And I may even have changed roles  since the last time we talked Danny. So  

I am the global Lead for retail in the  industry Solutions group here at Cisco. Love it. Love it. Well, you are always  the global lead in my book, right? . So sweet talking Devil . Well, mark so, so those of you that don't  know Mark and I have a long history of   doing recordings, whether video or audio with  Cisco, and I am notorious for going off script. So I do want to show him, I have the  script here. We will remain on script  

Mark. There's an absolute no way  I'm going off script whatsoever.   And , so those of these that don't know,  we did a, we did a recording in New York   for NRF and that is actually how I broke the  ice with Mark. We had a script down. I threw   the script in the air and we winged it , and it  was one of one of the best videos we've done. So we're gonna do a very similar thing here, mark.  

So I'm gonna, I'm gonna kick it off with a very  easy question, guys. And that is, that is what   is meld. What, what are we gonna talk about today  and what's its, what's its motion in, in retail? Sure. So I don't want to do the elevator  pitch and stop at every level,   so we'll go as high level as we can go. meldCX is headquartered in Australia,  founded about three and a half years ago,   focusing on technology within the  retail environment. One part of our  

business focuses. Device, peripheral  and application management. So we've   got a middleware application and the  other part focusing on computer vision,   which is a very broad term that t that basically  takes video signals and turns it into data. And that's a, that's also a pretty big bucket,  right? So in, in the retail environment,   we're learning that there is a very large gap.  Retailers have of what's actually happening  

within their environment. They've been operating  maybe many for dozens and dozens of years. But   they have found that during the pandemic in  particular, that there's a massive lack of   understanding of what's actually happening  on an ongoing basis in their environment. So we're trying to teach them. To  utilize intelligent technologies  

to learn more about how people  behave, how their own associates   behave in those environments. How we can  do things like look at in stock or out of   stock. Cuz that's been a massive problem  within the grocery market in particular. And so we're kind of trying to uncover all of  the data that's hiding in plain. and you can't   do that with some level of intelligence,  without some level of intelligence at the   edge. And because it's going into the camera,  we found out that we can get very scalable very   quickly. And so we've been in some very  high level conversations incredibly fast. Hope that, hope that helps. Is that  a good overview? Absolutely, it does.

It does. Absolutely. And. So I, I,  I simplify it even more. So if we,   if we think about the e-com world we've always  known exactly where the customers come from,   where they go on the website, what they  look at, how long they look at it what,   whether they consider it and put  it in the basket, whether they. you know, move on to something else. Does a  promotion on the webpage make a difference?   You know, to, to conversion All of that  stuff has been kind of intrinsic to the   platform within e-commerce, but not so much  in the physical environment. We knew when   people walked through the door, or rather  how many people walked through the door,   we didn't know anything about them, and we  know how many people checked out at the end.

and that was pretty much it. It was a black  box inside the store, unless you had some   intern standing with a clipboard doing random  interceptions, you know scribbling things down.   This changes the game from, from a retailer's  perspective. And there's, there's studies that   show that personalization within the shopping  journey can drive a, a 40% larger basket.

To be able to personalize. You first  need to understand the consumer,   and that's what computer vision and, and  meldCX is able to do for us and with us. Definitely Mark, and you and I have  also discussed more of like that video   analytics portion, which is like that  use case mainly that meldCX does offer. So with more of that video analytics,   but also that behavior analytics that  goes kind of behind with that software as.

Well, let me take a, let me take a bit of  a step back to say one thing very clearly,   because there have been implementations of  video within retail environments before.   It's predominantly been focused  on things like loss prevention. Okay? Loss prevention is a very nice word, which  is are peoples stealing things? . And if so,   how do I identify that they're stealing  things? And then can I go backwards in time   and bring that to law enforcement? So video has  traditionally been backwards looking effectively,   that says if something has happened in the  last 90 days and I've recognized that some   semblance of crime or something that bad has  happened in my environment, can I go back and. Effectively. Right. So, and that, and that's  a, that's absolutely the right use case for  

folks that are looking at loss prevention  from like a law enforcement perspective,   which a lot of LP professionals are, right?  Former F B I, former law enforcement,   things along those lines. The question  really is, is what could you do if you   had information in real time about what  was happening, but you weren't looking for? individuals, meaning that we don't look at the  face of the person. Right? But if you understood   more about metadata around an individual, and  by metadata I mean is the person wearing a Nike   shirt or basic demographic information,  right? Or the times of days they come in,   all the rest of those things, does that data  actually provide things that are valuable? And what we're finding is the answer to that  is yes, it does. So we're at that point. That,   that proof of concept stage with  a lot of retailers where they're   I, I don't believe that this can actually be  real. And then we say, well, we'll prove it   to you. Right? And they, they go, okay, prove  it. And we say, okay, well what are we proving? That's the first thing we have to define  what the thing is that we're proving.  

Is it that we're gonna utilize cameras to  look at cars going through a drive-through,   right? And look at that. That use case, is it  gonna be, that we can bring, we can look at   gaps in the shelf to do planogram,  compliance, things along those lines. And so we try to take a narrow approach  and say, let's prove the narrow thing,   but then we can look at the data. Around  that narrow thing that gives us insights   about what's actually happening within that  environment. And Mark, you speak a lot about   this and I think it bears repeating about the  idea of going from kind of insights to action. Maybe you want to kind of touch on that for a  second cuz I think it's very valuable. Yeah, sure.

It was a, a phrase or a series  of words, I guess that somebody   used in a meeting a couple of years ago,  and it, it stuck in my head. Visibility,   insights, and action. So we're able  to gain visibility with a camera. Fairly straightforward. It is no  different to any other sensor. It  

just happens to have a lens on it. It's a  super sensor, if you like. The AI models   that meldCX produces provide the insights  from what it's seeing. So from the camera's   perspective, it's, it's an object. It doesn't know  whether it's a car or a person or a can of soup. But the AI models help interpret  that data and make inferences.   So I'm trying to think of a good example. But  being able to look at a situation and derive a.   Need or, or concern that it highlights. And  then there's the action piece. Okay. So what  

are we gonna do about that? Actually Raffi  mentioned looking at cars in a drive-through. If the, if the line is growing, what does that  tell you? and what can we do to prevent the,   what's eventually gonna be a drive  off bulk or abandonment of the line.   So what action are we gonna take? Are  we gonna dispatch an associate with a   handheld to, to try and triage the line? Are  we gonna, you know, offer promos to get them   the customer to select certain things that are  gonna, is gonna shorten that service duration? Or are we gonna try and divert them to curbside  and triage the line that. So that visibility,   insights, and action, a a approach  I think can be applied to many,   many things. Obviously video  analytics is what we're talking   about today. Right, but it, it is  something worth banging in mind. There's no shortage of data in retail.  Is it relevant data? Can we derive some  

insight from it? And then what the heck  are we gonna do about it once we do? So guys, I have to ask, because every time I, I  hear these type of things my immediate response   is, oh, who, who the heck wouldn't do this? But  I'm sure there are people giving you pushback. I'm sure there are people saying, that's not  for me. What, what, what, what are, what are   you hearing from customers that are saying that?  Sure. What is our, what is our counter to that? Sure. Well, I think the first pushback is   cameras are creepy. I think that's  the first one that we hear a lot.

You know, I got one in my  face right now, I can attach There you go. Exactly. It's, it's, it's looking at  me right now. Luckily this one's not intelligent,   it just has a lot of pixels. But the,  the idea that someone is watching,   right? That idea that someone is, is,  is kind of viewing it kind of live,   if you will, and trying to look about you  and trying to figure out what you're doing. It's an interest. Approach or it's an  interesting kind of feedback because  

we all have phones in our pockets that are  giving an enormous amount of data away on   an ongoing basis. The reason why retailers  have the opportunity to create personas,   I think that we've talked to, you probably  probably talked about this before, right? We all have lots of personas that we're  associated with, right? There's a whole   sub-practices within agencies that talk about  persona mapping effectively against individuals,   right? If you buy. Like I have, and I'm  a proud minivan owner, I should say,   right? . So if you're a proud minivan owner  like I am, you are likely to have children. You are likely, if you're in the Midwest that  like I am and you've got an all-wheel drive,   you're probably traveling a lot,  right? You've got family in Michigan,   right? So there's all these kinds of things  that you can assume. Based on those personas,   and because we have those devices that  are essentially broadcasting an enormous   amount of information about a on ongoing  basis, that there's a lot that the agencies   and marketing technology companies have  been able to build off of our profiles.

Now, a lot of this has changed because  if you are an Apple user, right   you are partially responsible for the loss of  about 60% of Facebook's market capitalization   over the last. because you hit that button  that says, do not track me across multiple   applications. Okay. It's getting a  little in the weeds, so I apologize. I'll, I'll bring it back to, to, to, to something  that's a little bit more relevant here in a   second. But They used to have ability to  say, if you log into Facebook on mobile,  

on your phone or whatever, you could, they could  track you across all kinds of applications. So   that's why if you go and, you know, look  at something on Amazon or whatever mm-hmm. And you pop over in the other browser to  cnn, all of a sudden there's an ad barking   at you. That is the thing that you just  looked at. What the heck is that? Well,   yeah, because they're tracking you  across multiple applications. It was   true on your phone as well until Apple.  change things and Googled it as well.

It says, do not track across my, my phone  effectively. Right? So, although it's a little bit   less intrusive, there is nothing more intrusive  in the world that's gonna be as your phone.   effectively, right? Because it's got location data  and all the things around you. So that's why we go   to great pains to explain to customers that the  camera is a sensor and we treat it as a sensor. So although it is video and it is being  deployed predominantly for loss prevention,   i e video storage, so that folks that are  in law enforcement can look backwards in   time to see if something has happened.  What we're doing with our AM AI models   is in. the video in real time to  drive out those insights, right?

And to take video terabytes of data and  turn it into ones and zeros kilobytes of   data. Effectively, right? So we are not  tied, the way that the technology works   is it's not like we're pulling and taking in  tons of raw video on our side in the cloud,   and then have a lot of people looking at it  and trying to figure out what's going on. We're quite literally doing it at the camera  where we can't even see the video, so we don't   see it on our side. All we see is metadata  being extracted out of that video, and then   we're being able to inference that information  as. There was a bit of a tortured explanation,   but I think it's important to say why cameras can  be great sensors and also be privacy compliant. Yeah, there's, there's always  been. This concern around privacy,  

whether you're talking about, you know, your  digital assistant at home, your, your Alexa,   or your Google, I've gotta be careful what  I say here cuz something's gonna spark up.   And, and there's this misperception, I think  as with cameras, that somebody's sitting there. watching it or listening to it or whatever  it may be. And, and the reality is it's a,  

an ai that's looking for specific  triggers or, or patterns.   And really it's not invading in your  privacy. You know, somebody isn't listening   to your personal conversations that you're  having in front of your digital assistant. So it's, it is one of those  perceptions we have to overcome. Yeah. Yeah. So privacy is one of  the biggest pushbacks I would say. Guys, we are at the .

That's a short answer. Yeah, we've got more pushbacks   I can give to you, but if, if you wanna  move on to another topic more than happy to. No, I mean, I, you know, I'm,  I'm, I'm came to go as deep as we,   we all feel comfortable and, and, and want to. Okay. Folks, I, I do wanna remind everybody  that There are gonna be a numerous links down  

below this video or this audio podcast  that you are listening to. So if you   have any further questions or you want a  deeper look into anything you are hearing,   please click those links below.  And, and, and feel free to browse. There is also going to be an email address  down below, mark. This is something new that   wasn't on the previous podcasts that you  know about. But if you have any questions,   please feel free to shoot an email  to us with us with those questions,   and we will try our best to answer  that on an upcoming podcast. Thanks, Rafi, and, and I kind of wanted to  also touch on that license plate usage to   recognize the cars as well. So we'd love  to go into that and get your thoughts on.

Sure. Well, there's two lines of thinking  that we've been solving for, for retailers.   One of them is to utilize license plates to be  able to create a more frictionless system of. Grocery pickup or license plate to pay  in a drive through. Things along those  

lines. Right? So there's a line of thinking,  which is, which is a, a, a thing that we're   pursuing and we're implementing, which is  that co consumers want convenience. Right,   and they're willing to show, tell you what  their car is and what their license plate is,   and things along those lines, and verify their  identity so that when they get to their grocery   store or they get to their drive-through,  they can effectively just drive in. They get recognized and they drive  out right. You have to ensure that   the security is protected and there's  a lot of things that you've gotta do in   order to meet the security requirements  that the retailer has put aside. That's  

absolutely a use case that we can kind of  go through, and there's ways of doing it. There's all kinds of technical ways  to do it. Real time, near real time,   all kinds of different things. But  there's another line of thinking where   there is. Wanna collect data again,  metadata. So there's a customer that   we're working with that had required us to  not use license plate information. Okay?

So going back to this idea of training our  models and utilizing an ethical approach to   artificial intelligence, we said, okay,  we're gonna basically fork off our model   where we can look at the make and the model  of a vehicle, but not look at the license.   And we're gonna use the pixels that are  available to us to take a picture, right? And then be able to track that picture from  camera to camera in a drive-through to be   able to get those statistics and that  information. But we basically can see   the license plate. I e, the camera sees it,  but all it sees is a collection of pictures   that it's sort of like a fingerprint more than.  And we bring those through in the drive through. We collect a little bit of  data. We ensure that that is   tokenized. So it's encrypted and secure kind  of at the edge. And once that car leaves,  

that token is destroyed. And we don't  have that information. We don't know who   that person is. So we can kind of solve for  both use cases and both use cases are valid. And it just kind of depends on what  the customer's really asking. Whatever.

In a lot of cases, it's, it's  gonna come down to whether the,   the, the consumer opts in if they see  value, hundred percent. Generally there's a,   there's a, a, something called the Plain Site  Doctrine. If, if you can see it in the streets,   then you can use it to identify somebody and,  and license plates, generally a public record.

However, I completely understand that  some retailers and other industries,   may have privacy concerns around that and moving  jurisdiction to jurisdiction not just in the US   but globally. It can vary. So it is definitely  something to be cognizant of, but being able to   identify, make, model, and color of a vehicle  it's, it's perfectly ethical but also as, as   Rafi was suggesting before you can actually infer  certain things from, you know, types of vehicles. Oh, my camera just shuts off. Sorry. Am I still with. Yeah, see you, but we can see you fine,  Uhhuh. Sorry. Sorry about that. No worries. Oh, I ran outta space on my hard drive. Excellent.

How did that happen? Your deck before you start? Right?   Well, it's a, it's a, it's a, it's a great  point. Markets, it's this idea of what are   people prepared to. . Mm-hmm. . Right? And the  risk that they're, that they, that they think   about when they look at kind of technology  deployments. This is really net new stuff.

You know, we, we are at the, the, the  cutting edge of this, this market right now,   but it's only gonna increase. I say that, but  I think that there's barriers. and I think that   there's some unethical behavior that's out there.  There's some problematic technology that's been   implemented. So I think that what we try to focus  on and the reason why we're so invested in our   relationship with Cisco is can you get any more  trusted from a network perspective than Cisco? and I think the answer is, is no. Right?  Like, you know market leaders and we   want to be seen as a market leader. And we  also, during our process, when we talk to,  

you know, some of the, the top end,  you know, fortune 500 is we are here   from the grace of the platform. We are an  application that sits on top of the platform. If we do anything that violates  anything from a Cisco perspective,   they can turn us off and turn us off.  So we, we, we are disincentivized from,   you know, having any kind of bad behavior  or running afoul of any kind of security   protocols cuz it would cripple our business  if we did it that way. So that's a pretty  

good incentive to stay in the, stay  in the lane and over the long term. So I, I have, I have a, a question I, because both  you and Mark have touched on this you know,   you talked about delivery systems, but  both of you have touched on the. You   talked earlier on about people being in,  in line with their cars and then taken off   because it was taken too, too long. So,  so how are we optimizing drive-through?

I would imagine that's something that we're doing. Yeah, absolutely. Well, I think the first part  is we're collecting data about what's going on,   and we're showing insights that  we're previously unavailable.  

So the, there's a big shift that  happened in the pandemic, a lot of.   Kind of traditional kind of QSRs, if you will,  like the McDonald's of the world and folks that   have been at for a long time, they've always  had this 70 30 split, or about 70% of their   businesses going through drive-through  about 30% is kind of going in retail. It turns out now that that's effectively  industry-wide, if not higher in some places.   So where people have expected a 50 50 split, it's  gone to that seven 30 and it's really not coming.  

So what, there's a lot of folks that found  themselves earning an enormous amount of money   throughout that process because they had  that drive-through that was available,   but they also don't know what's going on out  there because their point of sale data is   basically an on-off switch, somebody  ordered and then somebody delivered. Right? And all the data in in between is just the  big, big blind spot. So the first thing that we're   doing is we're identifying what the blind spots  are, and then secondarily, we're looking. , what   are those balk and abandonment rates? When does  the congestion happen? Right. So it's almost like   there's this idea wasn't an economist, but I'll  have to look it up, of, of the accordion effect. I don't know if you've guys have heard about  that, but it's, it's related to traffic and the   accordion effect is effectively, if you've ever  done a long distance trip and there's an accident,   you know, a quarter mile down the road, but  there's nobody blocking. The lanes at all,  

you're gonna incur counter a slowdown cuz  the first person looks in rubber necks and   then the person behind them has to hit  the brake and then they rubber neck,   and then it creates this congestion and  then that congestion opens up ahead. So this kind of back and forth,  which is an accordion, right?   So that's what we're seeing in the drive-through.  We're seeing that accordion effect where you got   a tremendous amount of congestion and then it  opens up. But then we're starting to look at,  

okay, what, what were the  drivers behind those things? and it turns out that the drivers could be a  lack of an efficient payment system. It can be,   believe it or not. Large dogs that are  in the car that come out and try to poke   their head out the, it's, it actually  happens, believe it or not, right?   That, that then the person that's serving wants  to interact and they're having this great.

you know, interaction human, you know, canine  interaction, and then it, it slows things   down. Right? But the biggest one is like, when,  when things become inaccurate or you've got a,   a , I'll go back to the minivan here. You got a  minivan with like six kids in it and everybody's   got a different order and then it slows it down,  and then you people start blocking away, right? Yep. So there's a, there's a lot of things  that are going on that we can say when we   look at Congest. Go back to the example of  kind of vision and insights and actions.  

What are the actions that we can take?  Some of the actions that we're looking   at is integration with many board systems  where you're promoting not items that are,   you can traditionally think of,  quote unquote upsell, right? I'm gonna do something that's a little  bit higher margin. You're trying to   figure out what are the areas that are the  easiest to make of which you have the most   inventory to be able to get that line  moving. because throughput is critical,   right? In those really kind of peak times  within those restaurant environments.

So it's a, again, it's a long answer. I, I'm sorry  I'm not giving you good sound bites here. They're,   they're longer answers, but there's a lot of  complexity within those answers. I love it. I love it. Raffi's example you know, in terms of the,  the, the promo, if you've got line growth   and you have a, a bunch of people in the line  that are gonna order a, a hot sandwich that,   you know, maybe is pretty packaged, but  it's still got to sit in the microwave   for 60 or 90 seconds and you can send a  promo targeted to the, the individual who,   you know, excuse me, you know, is in the  line already for, you know, a cold danish. Straight outta the cold cabinet, you've just  collapsed that line by 60 or 90 seconds. And  

if you, if you can do that multiple times  shorten the line overall and then you will   see a re reduction in those people pulling out  of line. And, and while we're talking about   drive-through, it's actually no different  to the, to the line in the grocery store. Exactly. You know, we're talking  about cars, people are objects too.   That probably sounds bad. . I am not an  object. perfectly don't objectify me. But  

you know, it, it's, it's from, from the video's  perspective, it is an object with a predictable   movement pattern and predictable actions that can  come out of you know, what the camera's seeing. So, you know, you walk up with.  cart with $300 of groceries in,   and it's five deep at the, the register. You,  oh, forget this. Walk away. And you're not just  

losing the revenue that's in the cart. You've now  gotta pay somebody to put that back. And there's   also the potential for spoilage. You know, if  there's frozen goods in there and it takes them,   you know, an hour to actually wrestle that cart  and, and get the stuff back to where it should. Potentially it's defrosted, you know? So, and,  and you can apply this across multiple segments   of retail where you can look at the behaviors that  are gonna occur. I, I've, I've done this in, you   know, big box stores where I've, I've walked away  from a cart because I couldn't make a phone call.   So there's all sorts of reasons  where you can potentially.

Things like Abandonments not just in  the drive-through or at the register,   but somewhere else in the store. If a cart  is there and it's been sitting there for,   you know, 15 minutes, chances are somebody  has walked away and just left that product   there. And you know, you can dispatch  somebody to, to go and deal with that. Mark, I think additionally from that is  even brand loyalty, right? Like if I,   if I know that that store is gonna  have a long line every time for lunch,   I'm gonna go somewhere else for lunch or I'm  gonna go do my groceries somewhere else. Yeah. And, and, and thank and thank you. Because  this is an area that's so frequently.

Skipped. You know, we, we look at  the, the, the direct costs. You know,   what's the revenue lost and what's the cost of  labor to put that product back on the shelf.   But you've got the soft costs as well. And if  you go to the marketing department, they will  

be able to tell you what, you know, a, a point  of loyalty costs for that particular retailer. And each time I abandon my  car to walk out the door   you know, how many points are being lost  in that process. That's gonna impact not   just the single shopping trip, but the lifetime  value of the customer. If they decide to take  

their wallet to the, the grocery store  or the home improvement store down the   street because they're not happy with the  service they're getting on a, on a regular basis. And we've all, we've all done that, right? I know  I have. Oh yeah. You know, I've, I've been to,   I've been to a Home Depot and I've looked  at the quality of the lumber and I'm like,   no, wait, this is ridiculous. You're gonna  charge me for that. And then I go to Lowe's,   and then the next 6, 7, 8 times I'm going for it.

I'm thinking about Lowe's first, then  I'm Home Depot. Right. I'm just using   that as an example. It's not to say that the  retailers are, they're both great retailers.   Thank you to, to both of them if  you like . But, but the point,   the point is, is that perception is real,  right? And it's hard to figure out how you   measure those different things and what the  long-term impact kind of on your businesses. We're working with one retailer that   they, they were saying we, we, we always ask  the question of what are you currently doing?   because they look at like, what's the art of the  possible with this new technology, quote unquote   new technology? And then we say, well, what are  you doing now? And they go, well, we do survey. They go, oh, interesting. Right, okay. You're  doing surveys for people and how long they're  

waiting, and they're like, I, we go,  how accurate do you think that is? And   the answer is not accurate at all because  there's a perception about the amount of   time that you spend in a line or whatever,  and that perception could be colored by,   I didn't have enough caffeine this morning, or  I need to use bathroom, or, you know, whatever. Right. There's all kinds of things  that can modify perception and what   we're trying to show is extremely.  , right to the subsecond level of   data and what does that data give you and  what actions that you can take on those.

And, and with surveys you're trying to  extrapolate a very small sample size. Mm-hmm. and the thing to remember with, with  retail, particularly in large geographic areas   like the US or Europe The how the brand is  perceived is gonna vary significantly from   area to area. So a format that works particularly  well in the northeast of, of the United States.   Put that in the deep south and you'll see  a completely different behaviors in there. I remember. , years and years ago when I first  got into into retail somebody telling me that  

when you see promotional signage outside stores,   there has to be varied by market  because two, for the price of one   versus 50% off when you buy two is perceived  differently. It's the same math, but it's   perceived differently by different demographics,  different geographic locations. And so on. And this is just a quick reminder for everybody  listening. If there is anything that you want a   deeper dive in, be sure to check out  the links below. They will have lots   of information on everything you are hearing  here. Guys, I, I have a, I have a question for  

you and, and, and we're talking about all this  and, and, you know, simplistic Danny Brain says,   this is pretty cool and futuristic, but  there's gotta be something on the horizon. Where, where are we headed? What's,  what's, what's in the future for us? Oh, Mark's better at predicting than I am. If, if, if I did, that's,  I'd be in Vegas Raffi. Boy,   it is, it's a. Difficult one to answer  because technology is moving quickly.   AI is moving quickly. Typically, retailers  don't honestly move as quickly. We, we saw,  

you know, a very rapid forced innovation cycle.  During the pandemic. We saw lots of Retailers   running out to either deploy or enhance their  curbside and, and drive through capabilities. And, and to an extent, that's why Raffi and I  tend to probably focus heavily on, on the drive   through because we saw a problem during the  pandemic where, you know, lines were backing   out onto the street. And, you know, the, the,  the market was desperate for a, for a solution.   But , we we're seeing a slowing  of that innovation and kind.

Retailers are now taking a step back and going,  okay, we, we deployed a lot of stuff during that   period. Is it still valid? And how can we in some  cases kind of shore up or reinforce what we've   done? We've, we've decided it was the right. Now  how do we consolidate and, and refocus on that? So I think we're gonna see a lot of the use  cases we've talked about and things we've,   we've seen in the last two and three years  become more prevalent within the industry.   I think you'll probably get fewer. Innovators  kind of leaping ahead. But that said,  

you know Rafi, I, I don't know  how much detail you can go into. We, we, we were having an interesting conversation  the other day about Hyperpersonalization. Mm. How   we can get to a point where we're, we're not just  looking at, you know, historical spend data in   the, in the CRM or loyalty system but we can look  very specifically at how consumers are interacting   with the store, with fixtures, with the objects  in the store being able to do targeted Engagements   and I, I, I want to get away from thinking about  promos cuz you know, promos sounded like Yeah. We were just trying to sell you more. Right,  right. It's, it's, it's really about engagement  

because I think ultimately the, the, the winner  in this whole environment is gonna be the retailer   that serves their customer best. Mm-hmm. .  And it's going to, as Sabrina brought up,   it's gonna drive that loyalty because consumers  will pay a little more for a superior experience. And in a, in a very   vanilla world is those retailers that stand out  are, that are really gonna succeed to do that they   need to understand their customer better. Rafi,  I don't know if you have any thoughts on that. Well, I, I would say, you know, I mean it's,  it's, it's a dangerous business to be in long   term prognostication, but I would say short  term, the bet that we're making, , right? Is, is exactly to that point.  Mark is that if we're looking at   how everyone is looking at how  much square footage they own,   whether it's a retail environment or a corporate  environment or whatever, right? They're trying to   figure out how much of the, the space we  need. If you wanna take a grocery store,  

for example, how much of we need to be able to  carve out to make effectively micro fulfillment. So that we can keep the fresh stuff  fresh and then bring it out to be able   to put into a car that's actually gonna get  delivered, right? And that stuff just gets,   it gets wobbly that, that, there's  no straight lines in any part of   this business where it just kind  of goes, it goes month to month,   quarter to quarter, and consumer shopping  and behavior happens it's like a whiplash. It goes back and forth and happens  all the time. So our big bet is that   real time actionable data and having  that data available, Is really the,   the new kind of gold rush because that 10 billion  and the 60% market value that came off Facebook,   it's not like marketers are not spending  that, those dollars in different places. like we've seen, I used to be in the digital  signage business and, and to some extent we   kind of still are because we're looking  at, you know, the efficacy of signage. But   the digital out-of-home business in particular  is growing like gangbusters because when people   are outside, it's showing you, there's a big  screen that's in, on the side of the road.

People look at it and it influences  behavior. Right? But if you can't   really target and look at that phone and  be able to do those targeted conversions,   Buyers have been spending the  better part of a decade doing,   they're looking for other things that they  can measure to be able to show a, a real. Physical, click to convert, if you will, right?  A mouse click in a, a physical environment,   a mouse hover. Those are things like  where people's hands are and how they're   engaging. And do they put it in the  basket? Because we can look at say,   okay, do you wanna give us the point of sale  data? We can say where people have moved in,   the things that they've interacted with, and  did it actually end up being paid for or not? Right? And that level of data is, is much more  than you would get when you're online. So our,   our. In the, in the short term, and  I guess short to medium term is that  

data is the currency of retail  and that's where we wanna be. Yeah. And I, as you were talking there,  Rafi, I, I was coming back in my head to   how I generally open up a presentation  when I'm, I'm talking to a retailer and   it's really about how can  you be an Agile retailer? Mm-hmm. Because. As you say, it, it, it's,  it is the whims and vagaries of, of consumer   demand. But it's also the business landscape  nobody could have for foreseen. What was gonna  

happen over the last three years and. . So  you had the business landscape change and   on top of that and in part, main part because  of that the consumer demand changed as well. Mm-hmm. So being being able to get that  realtime or near realtime data helps you   be agile and respond to those customer  demands and, and changing landscapes,   whatever the next, that's way you've  made whatever the next thing is. Yeah. Yeah, exactly. Hopefully nice. Yeah.

Well guys, I know we are up against the  time slot and I wanna be conscious of,   of the, the allotted time that you gave us. And we are approaching my favorite time of the  podcast. And Mark, you've already joined the   podcast, so you know the secrets out of the  bag. But Rafi, this is new for you. And so I,   I like to tell everybody, if you are like me,  you join podcast and you the night before,   think in your head of all the crazy questions  that I may ask you or that Sabrina may ask you. And because we throw scripts on the  ground in the very beginning of podcast,   I don't ask those questions. And so this is your  open opportunity. What should the audience know  

that maybe I haven't asked you?  Or that one nugget? If there's one   nugget that they should walk away  from this podcast with, what is. And Mark, same question to you. So  get ready cuz we're coming to you. I would say that we are fanatically committed  to finding the places that are a little strange   where maybe a vendor hasn't gone before  and that we can spin up new use cases   around. We love to find those little niches  within markets that have been previous. Uncovered or unexplored, and we can find  them and then deliver immediate value out of.  

We have the ability to move very,  very, very fast. If the resources   were committed the right way, and  we can find those kind of nuggets   of information that allows us to unlock  a tremendous amount of business value. So we try to focus less on  product. and more on the solution.   And I think that that's what it is. I  mean, we, I spent a lot of time talking,   like I mentioned before, and there's a reason,  because if you don't have an honest communication   going on with your customers and you don't  ask about what problems they're facing,   and you can't get that level of trust, you're  never gonna find what your value proposition is. So that's, that's what I think is important about  us, is that we're veterans and that we really look   to find the places where we can drive the business  value that may not have been uncovered previously.

Beautiful. Well said. I, I should have known  it was coming. Danny and I, I didn't . So   I think I, I come back to  something we, we, we frequent. Say when we're talking to, to customers  and, and they're, they're trying to wrap   their heads around what the heck is  video analytics and computer vision?   And there's a lot of expertise out in retail but  it's generally not systemic or institutionalized.   I can't even say it. Institutionalized. So if you  were to take one of your most experienced manager,  

and set them up on a shelf at the, at the front  of the store and, and look down on the store. What would they see? What would they notice?   This gives us the opportunity to, to some degree,  institutionalize that knowledge so that it becomes   just part of the process. Nobody even thinks about  it or notice. If, and, and Rahi can correct me if   I'm wrong here, but generally , if we can see  something, then we can probably analyze it. And if we can take that knowledge from those  experienced professionals in, in the industry   and be, actually I should, I should add a piece  in here. You know, the number one challenge today  

in the, in the retail market is, So, you know,  not only are we losing associates, we are losing   the experienced, you know,  store leadership, et cetera. So we need to be able to take that  knowledge and replicate it through   the systems so that other people don't have  to. That's great. Love it. Love it, love it. Well guys, I, I want to  thank you for your time and,   and conversation. I think I  learned quite a bit, like,   I won't speak for Sabrina. She's far smarter  than I but God knows I learned quite a bit.

I think our customers did as well to our customer.  To the folks listening in the audience again,   just one last reminder. If you have any  questions, please feel free to email us   and we will try to get that in an upcoming  podcast. And if you want a deeper dive into   anything you. Click the link down below and  take a tour on whatever URLs are down there. Guys, thank you so much again, I hope to  have you both on the podcast. Once again,   Sabrina and I thank you all for listening.

Thank you very much.

2023-02-08

Show video