Latest / The Retail Podcast / How Eagle AI Delivers 7:1 ROI Through Personalization & Gamification — with Cédric Chéreau, Co-Founder

Transcript

How Eagle AI Delivers 7:1 ROI Through Personalization & Gamification — with Cédric Chéreau, Co-Founder · The Retail Podcast

• 0:00
  And I think what is very, very important in what we're doing is
• 0:03
  the core philosophy of what we're doing is remaining
• 0:07
  customer centric. What is really interesting for
• 0:10
  customers, how will they naturally behave?
• 0:12
  Because we're not here to force them to change their behaviour,
• 0:16
  We're here to incentivize them to do a little bit more of what
• 0:19
  they're already doing. Hello and welcome back to
• 0:33
  another special episode of the Retail Podcast.
• 0:36
  Now Eagle Eye, who will meet one of the cofounders in a minute.
• 0:39
  Cedric, I've seen all over the place.
• 0:41
  I've seen them for years at NRF, I've seen the literature and I
• 0:45
  also we have a shared customer that we have both worked on.
• 0:48
  And so when I have the opportunity to interview Cedric,
• 0:51
  I thought what a great opportunity to interview someone
• 0:54
  that I pretty much can bet my bottom dollar that the problem
• 0:58
  they solve is a massive problem for retailers right now.
• 1:02
  So Cedric, why don't we just start with you?
• 1:04
  I know you're the Co founder and obviously managing director of
• 1:07
  Eagle Eye, but one, just tell us a little bit about yourself and
• 1:10
  then tell us a little bit about Eagle Eyes to set the frame for
• 1:14
  the conversation. Yeah, absolutely.
• 1:16
  So hi, thank you very much for having me today.
• 1:20
  I'm Cedric Chaho, I'm French, I live in in France and I'm the Co
• 1:26
  founder of Eagle AI, which is one of the department of Eagle
• 1:29
  Eye. What we do at Eagle Eye globally
• 1:31
  is to help retailers to scale personalization and to make it
• 1:36
  really efficient for them, for the customers and also for the
• 1:41
  suppliers who are usually participating in all this.
• 1:45
  And that's really the mission that we have.
• 1:46
  How can we make it easy and effective for retailers to
• 1:51
  operate personalization once again at a very, very large
• 1:54
  scale. That's I don't.
• 1:55
  You have, can you do when we talk about personalization, can
• 1:59
  you frame that for us? Because sometimes people use
• 2:02
  personalization for so many different things.
• 2:04
  What? What is that?
• 2:05
  If if we're just going to take another layer down and say what
• 2:08
  personalization means to you, an eagle eye, Well, how would you
• 2:11
  describe that? It, it, it's a very good comment
• 2:14
  and I think you're absolutely right.
• 2:15
  Personalization sometimes for some people means being able to
• 2:19
  put the first name of someone in on top of an e-mail.
• 2:23
  That's that's it. It's but it's not really
• 2:26
  effective. What we mean by personalization,
• 2:29
  eagle Eye is to be able to define the right parameters of
• 2:34
  an offer to make it really efficient for me.
• 2:37
  For instance, which brand, which product are the most relevant
• 2:41
  for me as a customer, probably product that are different than
• 2:45
  than your favourite brands. So we all have very specific
• 2:49
  product and, and if even if we have the same product, maybe we
• 2:54
  can adapt the threshold that I need to reach to get a reward.
• 2:59
  And that's what we mean by personalization, is being able
• 3:02
  to identify, predict the natural behaviour of each customer,
• 3:07
  define the potential of each customer to be able to really
• 3:12
  define the threshold that each customer individually need to
• 3:16
  reach to get a specific reward. And even the reward can be
• 3:20
  personalised because we all have a promotional sensitivity that
• 3:23
  is a bit different. So if we can identify all these
• 3:26
  elements, at the end you have, you could receive an offer that
• 3:29
  is completely unique that no other customer will receive,
• 3:33
  completely adapted to your need, your affinities with the
• 3:37
  different brand. And that's what we mean by
• 3:38
  personalization. OK, I got you.
• 3:40
  How does gamification fit in into that?
• 3:43
  Because again, sometimes I feel that's misleading as a as a
• 3:46
  subject in terms of people talk about.
• 3:48
  Yeah, tell, tell me what? Yeah, I, I, I your perspective
• 3:51
  on. That yeah, absolutely.
• 3:53
  I think we never design the solution that we worked on as a
• 3:57
  gamification platform, but the gamification came on top of it
• 4:01
  and it was a little bit of a surprise.
• 4:03
  Like the cherry on the cake, it was designed to be
• 4:05
  personalization. The idea was to for each
• 4:08
  customer to receive different offers with threshold to reach,
• 4:11
  and if I reach this threshold, I will get a reward.
• 4:14
  By definition, it's a little bit gamified.
• 4:16
  It's not designed to be just fun.
• 4:19
  It's designed to incentivize me to be a little more loyal to
• 4:22
  that brand of that retailers. So there's there's a nudge,
• 4:26
  there's an incentive here. So therefore it's a little bit
• 4:29
  little bit gamified, but it's not fun per SE.
• 4:32
  I would say it's not a pure gamification, yeah.
• 4:35
  OK. Thank you.
• 4:35
  You're right, I think it's, it's far the solution.
• 4:38
  It's part of the solution and it's it's especially in grocery
• 4:41
  retail, it's important to keep in mind that what is important
• 4:45
  is for customers to be able to earn something and to earn a
• 4:48
  reward. Having fun, playing a little
• 4:49
  game is always nice, but in grocery report, it's not the
• 4:53
  main mission of the of the of the retailer.
• 4:56
  The main mission of the retailer is to provide something that is
• 4:58
  clearly bringing value to the customer at the end.
• 5:01
  How does when, if we stick with grocery or or or maybe not?
• 5:05
  We can however you want to find answer the question.
• 5:09
  But can you share like a pragmatic approach to AI where a
• 5:14
  sort of simple, well defined use case has outperformed big flashy
• 5:21
  AI initiatives which you you see across across the industry?
• 5:25
  I think that the multiple usage of of AIS is something very
• 5:30
  pragmatic, probably the most basic for me, but it but it
• 5:33
  everything is in the detail. The most basic is being able to
• 5:36
  identify the five most important items for each customer.
• 5:42
  When the retailer is sending an e-mail to to to an e-mail or
• 5:46
  notification to their customers. I want to re interest all my
• 5:50
  customers with the five best offers we have installed this
• 5:53
  week. Just being able to do this is
• 5:56
  difficult because you need to understand the affinities
• 5:59
  between the products and the customers.
• 6:02
  You also need to look at the data.
• 6:04
  If I bought my laundry detergent last week, there's no need for
• 6:07
  you to put that on top of my list because I will not buy it
• 6:10
  again. I will probably buy it again in
• 6:11
  a few weeks, but not tomorrow. If you have products that are,
• 6:15
  that have a facial value that is more important than others,
• 6:19
  maybe they should rank higher in, in, in the, in, in the
• 6:23
  e-mail. So all these elements, if you
• 6:25
  use AI, you're able to combine all these KPI's to be able to
• 6:30
  really offer the best, bless you, the best 5 offers for each
• 6:36
  customer. So it's a simple, very simple
• 6:38
  use case. But to answer it very
• 6:40
  efficiently you you need the power of AI to to to accomplish
• 6:46
  this. And so I don't want to put words
• 6:48
  in your mouth, but if I understand what you're saying to
• 6:50
  me is that you look at all the signals post purchase to make
• 6:55
  sure that communication is in line with those signals, which
• 6:59
  is the example that you gave. If I bought detergent, don't
• 7:02
  send me another offer on detergent to five, sorry to Five
• 7:06
  products or Five Points. If you send the communication to
• 7:11
  a customer and you say come to a store next week because we have
• 7:14
  a lot of offers. Among all the offers we have
• 7:17
  here are 5 that are really interesting for you collect the
• 7:20
  best of I for the customer. We look in the past, We look at
• 7:23
  what's happened in the past. What are the signals of a
• 7:25
  female? And that's what Eagle I are
• 7:27
  doing. Absolutely perfect.
• 7:29
  That's OK. You're doing understood.
• 7:30
  We're identifying which product you bought in the past to
• 7:33
  predict which product you will probably buy in the future you
• 7:37
  provide. That as a service or I guess is
• 7:39
  it a? Solution.
• 7:41
  It's a SAS solution and all the algorithm that we've built over
• 7:46
  the years, we encapsulate them into different use cases just to
• 7:50
  make it very pragmatic, very plug and play and easy to use.
• 7:54
  We received the customer data, what happened over the last 12
• 7:57
  months. We looked at all the interaction
• 8:00
  between products, customers, stores and everything and we
• 8:03
  predict what will happen in the future.
• 8:05
  And we look at the potential of each customer if they did this
• 8:09
  three months ago, how will they behave in three months.
• 8:12
  And based on this, we are able to identify which products,
• 8:15
  brands, categories are the most relevant to them.
• 8:18
  And I think what is very, very important, what we're doing is
• 8:22
  the core philosophy of what we're doing is remaining
• 8:26
  customer centric. What are what is really
• 8:29
  interesting for customers, how will they naturally behave?
• 8:31
  Because we not here to force them to change their behaviour,
• 8:36
  we're here to incentivize them to do a little bit more of what
• 8:39
  they're already doing. I'm not asking customers to
• 8:43
  switch from brand A to brand B. I'm here to tell him if you're
• 8:48
  more loyal then you should be rewarded for that.
• 8:50
  Yeah, OK. I don't know why you've not
• 8:53
  taken over the industry because I mean that that sounds like the
• 8:56
  ideal thing when when, when, when I talk to retailers and the
• 9:00
  problems they face. But one of the things I read
• 9:02
  somewhere in your literature, I don't know if it's a promise or
• 9:05
  an example, but you look at sort of like a A7 to one return on
• 9:10
  investment, which if true is remarkable by any standard,
• 9:14
  which I sort of understand from the way that you've just
• 9:17
  described it. I was wondering what, what are
• 9:19
  the hidden mechanics behind performance?
• 9:22
  And then what misconceptions do retailers have about achieving
• 9:27
  this return? I can be even more bold.
• 9:32
  It's old here if you say it. Oh, there.
• 9:34
  OK, go for it. The the the 7 to 1 can be 12 to
• 9:38
  one. It could be a 20 to 1 because
• 9:41
  everything is in the way we build their parameters of the
• 9:44
  offer. The seven to one is based on the
• 9:47
  best practise. Once again, the idea is to
• 9:50
  create more loyalty from customers.
• 9:52
  So I'm going to incentivize you, Alex, with some offers on your
• 9:56
  favourite brands and I'm going to ask you to spend a little bit
• 10:00
  more on your favourite brand at this specific store.
• 10:03
  And if you do a little bit more than you will get a reward.
• 10:06
  Yeah, because it is incremental by design.
• 10:08
  I'm asking you to do a little bit more than what the natural
• 10:12
  behaviour is, then it is profitable.
• 10:14
  Is it 7 to 1? Because I'm asking you to do a
• 10:17
  little bit more. And because you're doing a
• 10:19
  little bit more, you're buying a lot more products at the end of
• 10:21
  the week, at the end of the month, or should I ask you a lot
• 10:24
  more from you? If I ask you a lot more than the
• 10:27
  ROI will be 20 to 1. But The thing is on the long
• 10:31
  run, you will be discouraged to participate in this and you will
• 10:34
  say, well, that's too hard for me.
• 10:35
  I'm not participating and not engaging in this anymore.
• 10:38
  So it's, it's your fine tuning of finding the right parameters
• 10:42
  on each offer to make sure that we keep a very good level of
• 10:46
  engagement from customers on the long run.
• 10:49
  The customer lifetime value is a very important concept because
• 10:52
  we don't want to have customers for one week and see them
• 10:55
  disappear after that. We want to keep them on the long
• 10:58
  run. So it's asking customers to do a
• 11:01
  little bit more, but even if we're asking them to do a little
• 11:04
  bit more, that will bring the full basket on average at the
• 11:08
  end of the month and that is the 7th 1:00.
• 11:10
  I could make a comment, but I don't want to.
• 11:13
  Now make makes sense, I'm going to flip this question around.
• 11:17
  What is the ideal customer profile?
• 11:19
  Then for you, what when we are really, really efficient is for
• 11:25
  every retailer where you have a lot of customers, scaled is
• 11:28
  always a challenge for them where promotion and loyalty are
• 11:32
  playing a role, an important role in the business.
• 11:35
  And when there's hard competition because at the end
• 11:38
  of the gain, what you want to is to gain some market share from
• 11:43
  your competitors. So when you took all these
• 11:45
  elements into account, grocery is perfect for us because you
• 11:48
  have frequency, you have a lot of customers, you have scales
• 11:50
  and you have competition and promotion is playing an
• 11:54
  important role. Pharmacy, health and beauty
• 11:57
  retailers also completely core for us.
• 12:02
  And then all the enterprise retailers, I would say once
• 12:06
  again, we have frequency and competition.
• 12:08
  Okay, all right, I'm, I'm okay, pharmacy, can you just expand on
• 12:14
  that, 'cause that's a how, what does that look like as a use
• 12:18
  case? Because I'm, I'm thinking like,
• 12:21
  are you meant like pharmacy like in Boots, as in like the whole
• 12:24
  beauty suite? What is that?
• 12:27
  Not the the the medicinal part where you you get you you like
• 12:31
  Nelson, you see Well, it does prolative school where you want
• 12:34
  to buy your shrimp food, your vitamins, your your OK, your
• 12:37
  space, all these. Followed in a barrel someone is
• 12:40
  a great example. Exactly.
• 12:42
  Exactly. OK.
• 12:43
  Got you. So when you, when you look at
• 12:45
  the future of innovation and you look at for example, AI, dynamic
• 12:53
  driven pricing, supply chain optimization, real time
• 12:57
  individual engagement, what do you believe will have the
• 13:01
  greatest impact on a retailer's strategy from your experience?
• 13:07
  Because sometimes I think they think one thing, but you
• 13:10
  probably see another. What I'm just curious on your.
• 13:13
  Thoughts Innovation is very important and I think we're just
• 13:16
  at the beginning of what AI can bring.
• 13:18
  We, we, we do a lot of predictive AI, being able to
• 13:22
  predict what the customer behaviour will be in the future.
• 13:25
  If you combine this with agentic AI, it will be our with the
• 13:29
  interaction of the customers and how can we create and Gen AI on
• 13:33
  top of this, you can create individualised personalised
• 13:36
  content for them that will bring to, to to something that is very
• 13:41
  specific, very unique, very innovative for the customers.
• 13:44
  And obviously we're looking at this on a very regular basis.
• 13:47
  All retailers should, should remain in contact of all these
• 13:51
  innovation because this is just the beginning of it.
• 13:53
  But to answer your question, I would say that what is very,
• 13:56
  very important for for for retailers is not to look for
• 14:00
  what will become the next great innovation is how can we make
• 14:05
  sure that we're not missing the train that is right now in the
• 14:08
  station. Yeah.
• 14:09
  How can we make sure that even if it doesn't look that cool and
• 14:13
  fancy, we hop on this train Because if we miss this train,
• 14:17
  if we don't put AI and and more personalization and more
• 14:21
  engagement from our customers right now, we might not be able
• 14:25
  to do what will be super cool in a few years.
• 14:28
  And, and, and I think that's, that's one of the biggest thing
• 14:31
  that I'm sharing with my clients is don't think necessarily too
• 14:35
  big, too cool, too, too innovative, but make sure that
• 14:38
  you're using what is already available that you can launch
• 14:42
  within a few weeks. Make it a reality in your stores
• 14:46
  because it will start the whole journey of change and you will
• 14:49
  be able to, to, to reach the next innovation more easy.
• 14:53
  I see what I mean. Yeah, Yeah, absolutely.
• 14:56
  Final question, what's the one question I should have asked you
• 15:00
  that I didn't? I like the question you asked
• 15:03
  before when you said why are you all over the market.
• 15:07
  That seems to be pretty simple and that that that's, that's a
• 15:10
  very good question. I don't know.
• 15:13
  You know, it's OK, you know, I mean, that means the other
• 15:16
  questions did their job. But I'm just because I'm not an
• 15:19
  what I'm always conscious of, although I've worked in data
• 15:22
  technology, AI for probably longer than a lot of retailers,
• 15:27
  I'm still conscious of the fact that I may not be the expert in
• 15:29
  the room or in the field. And so therefore my approach may
• 15:33
  lead to a bias in in in part of my thinking.
• 15:37
  And hence why I always have this sort of question to sort of say,
• 15:40
  have I actually asked the right question or am I?
• 15:43
  Or am I, you know, reading the headlines, going down rabbit
• 15:47
  holes and there's actually something else I need to be
• 15:50
  thinking about. And so that's it, I mean.
• 15:53
  But to, to continue on this to, to follow up on this, because,
• 15:55
  because I think that was the right question.
• 15:57
  And the, and the answer is I don't know why we're not all
• 16:01
  over the market. I mean, I just, I think, I
• 16:04
  think, I know, I think exactly for the reason that I described,
• 16:07
  it's sometimes difficult for retailers to hop on the train
• 16:12
  that is passing. What we're doing with the
• 16:14
  personalised changes, the personalised promotion, the
• 16:17
  personalised ranking of, of mass offers that we're doing for, for
• 16:21
  you, it seems obvious, it seems like a no brainer, but sometimes
• 16:25
  it's difficult to act on it. And, and I'm saying to my
• 16:28
  retailers, you could do this in five weeks.
• 16:30
  Yeah, but we're trying to build this on our own.
• 16:33
  We have a whole road map. We'll be there in two or three
• 16:35
  years. And we all know that in two or
• 16:38
  three years, the road map will change, the priorities will be
• 16:40
  different and you will be in a different situation.
• 16:43
  And I think that's, that's the main difficulty is people are
• 16:48
  looking too far ahead when, when this should be more operational,
• 16:52
  more pragmatic. Let's do it right now.
• 16:54
  Prove me that your solution is actually delivering A7 to one
• 16:58
  ROI. Let's fine tune the whole thing
• 17:00
  together and we'll talk about it in a few months, in a few weeks,
• 17:03
  not in a few years. And that's, that's that's where
• 17:06
  I think the retailers should sometimes be more action
• 17:10
  oriented. I look this, we can talk about
• 17:13
  this for a long time and I have to caveat because obviously I
• 17:16
  know what you what it says on the tin of what you do and what
• 17:19
  you've said in this podcast. But if you do and say what you
• 17:23
  can do for retailers that it's signal based, SAS based, it's
• 17:29
  like you've taken away all of the problems, all the reasons
• 17:33
  why a retailer traditionally in technology would say no.
• 17:36
  But it it could also be that the the guys in tech aren't the guys
• 17:40
  in merchandising, aren't the guys in product.
• 17:42
  And it's sometimes the the people in the business that need
• 17:45
  to get behind the AI and you know, again, but that's a that's
• 17:49
  a whole nother conversation into get get going through retail.
• 17:53
  It's Cedric. Thank you so much for giving up
• 17:55
  your time. And will you be at any other
• 17:58
  shows before the end of the year or you NRF is the next one for
• 18:02
  you? Actually flying to Las Vegas
• 18:04
  next week for grocery shopping. Oh, nice.
• 18:06
  APA is in Vegas. It's a great event.
• 18:09
  It's a fantastic, it's a fantastic event.
• 18:11
  We have tech for retail in Paris in November.
• 18:13
  OK. And NRF in in general.
• 18:16
  Fantastic. So I will probably see you at
• 18:18
  NRF. Absolute view pleasure.
• 18:21
  Thank you. Thank you, Alex.