Latest / Shiny New Object - a Marketing Podcast / Episode 304 / Danilo Tauro / Aperiam / Managing Partner
Transcript
- 0:00Danilo Tauro: Start using tools, start comparing tools. Getting
- 0:03your hands dirty is what I would recommend to learn about the
- 0:06Speaker: Are you a fitness loving foodie? Then defi snacks
- 0:06space.
- 0:15is for you. Defi offers real high quality chocolate you love,
- 0:18combined with functional ingredients, giving you as much
- 0:21as 26 grammes of protein per bag. So you can be both a little
- 0:24bit bad and a little bit good at the same time as a female owned
- 0:28brand, defy dedicates a portion of its profits to other women
- 0:31owned businesses currently sold in the US. You can find us on
- 0:34defi with an i snacks.com and follow us on social at
- 0:39defisnacks.
- 0:45Tom Ollerton: Hello and welcome to the shiny new object podcast.
- 0:49My name is Tom Ollerton. I'm founder of the ad tech platform,
- 0:52Automated Creative the creative effectiveness platform, and this
- 0:56is a regular podcast about the future of data driven marketing.
- 1:00Every week or so, I have the pleasure and the privilege of
- 1:03interviewing one of our industry's leaders about the
- 1:05future of data driven marketing, and this week is Cannes week. So
- 1:08I'm lucky. I'm here in the boiling heat, not great for a
- 1:11ginger man. I'm in the reception of Martinez hotel, and I'm here
- 1:15with Danilo Tauro, who is managing partner at Aperiam, but
- 1:19also a senior advisor at McKinsey and Company. So Danilo,
- 1:24can you tell the audience a little bit about who you are and
- 1:26what you do?
- 1:27Danilo Tauro: Of course, thank you for having me. So I like to
- 1:31say and think that I've seen the advertising industry from many
- 1:35angles, because I've been at P&G. So on the buy side for
- 1:38almost 10 years, I've been at Amazon ads on the tech side, on
- 1:42the product side, for three plus years. I've been at Uber. I was
- 1:46the general manager of Uber Eats advertising. So on the growth
- 1:49side of the industry, and right now, I'm basically on the
- 1:52investing side with Aperiam and with McKinsey as a senior
- 1:55advisor. So like to think that I've seen different angle of our
- 1:59pilot industry.
- 2:00Tom Ollerton: Well, the downside of having an impressive pedigree
- 2:03is the expectation of the content on this podcast so very
- 2:07high. But thank you. Thank you for your time. So quick one if
- 2:11you're going to give a smart, driven student some advice in
- 2:15how to get into and succeed in this industry, if they're doing
- 2:18all of the all of the basic stuff, but how did you get
- 2:21ahead?
- 2:21Danilo Tauro: Yeah, I think as I became more senior or older,
- 2:26depends on how you want to define it. I think one thing I
- 2:29realised is when I started my career. So I'm sharing this
- 2:32because hopefully that could be interesting for you know, young
- 2:34and motivated people. I was thinking, Okay, if you have
- 2:37knowledge that, what will get you there? But then, as I was
- 2:41progressing through career, I realised that knowledge is just
- 2:44a portion of what you need, but you also need network. You need,
- 2:48basically, as one of your assets, you need a network of
- 2:51people that support you, that give you opportunity, even that
- 2:55you can just pick up the phone and call whenever you need to
- 2:58know and understand something. So what I'm trying to say is,
- 3:01like the two assets I would focus on as a young student are
- 3:06knowledge and network. Because by building those two, I think
- 3:10you get to get a much steeper start with within the industry.
- 3:14Tom Ollerton: So acquiring knowledge is more
- 3:16straightforward than it has been in the past. But how do you
- 3:20acquire a network? What's the practical advice for how to do
- 3:23this?
- 3:23Danilo Tauro: I remember when I joined P&G, so I'll share a
- 3:26story. The very first training that I took was called the PIE
- 3:30framework, where PIE stands for performance, image and exposure.
- 3:35So that's a training that P&G gives to basically new hires.
- 3:39And the message was to be successful within P&G, you don't
- 3:43need just the first letter, which is the performance the P,
- 3:47but you also need image and exposure. So what does it mean?
- 3:50Image is you need to build a brand for yourself. So make sure
- 3:54that when they think of Danilo or Tom, when they think of
- 3:57yourself, you clearly have a brand. You stand up for
- 4:01something. And then, of course, it's important that the right
- 4:04people know about your brand. So there is the component of
- 4:07exposure. So don't get me wrong, you do need performance, but
- 4:10then image and exposure are equally important. So that was
- 4:14when I really joined P&G, and then as I moved to the US, so
- 4:172018 I think the concept of networking and building a
- 4:21network as an asset became even clearer, not just within P&G,
- 4:25but outside of P&G. And somebody recommended me this book, which
- 4:28I I recommend to a student, which is Never Eat Alone by
- 4:34Keith Ferrazzi. And politically, he explains the mechanics of
- 4:38first, why make why networking is important, but also how to do
- 4:42it in a smart way. So, long story short, that was a bit my
- 4:45journey. So from the PIE framework to Never Eat Alone, I
- 4:49think it was really about being intentional, building genuine
- 4:53relationship with people. And so I think that was at the base of
- 4:56how I personally built my own network.
- 4:58Tom Ollerton: And if someone isn't going to read, Never Eat
- 5:01Alone. What are the three things to take out from that?
- 5:04Danilo Tauro: So maybe I'll go to one, which for me was super
- 5:07interesting when I read it, which is he realised that people
- 5:12that are successful have relationships that support them
- 5:16throughout their career, their life, and so on and so forth. So
- 5:19then he started to realise that actually relationship is so
- 5:22important, as we said, it's an asset. So I started to say,
- 5:25Okay, if I need to get from point A to point B, what are the
- 5:28relationships that can help me get there? And so basically,
- 5:31from there, he would build a relationship Action Plan, which
- 5:36basically what, what he would work on. So I think it was a way
- 5:40to share how to be very intentional about building those
- 5:44relationships that can help you achieve your objectives in life.
- 5:53Tom Ollerton: That's really nice. However, coming back to
- 5:55PIE, which is nice and easy to remember, well done P & G, it
- 5:59sounds like it's been written by an extrovert, right? Because
- 6:02someone like me, like I'll cross the room, say, Do you want to
- 6:05come on my podcast? I love meeting people. The more people
- 6:07I meet, the happier I am. I love it. But there's, you know, I've
- 6:10worked with some incredibly talented people who don't want
- 6:13any exposure, they just want to do a good job. So do you think
- 6:16that PIE framework is focused on the neurotypical folk, if you
- 6:20like, or how would that work for someone who would rather die
- 6:24than get exposure to the wrong people? They just want to be
- 6:26brilliant.
- 6:27Danilo Tauro: I love this question because I think many
- 6:29people think that I'm an extrovert, but the reality is
- 6:32I'm a big introvert. So what I'm trying to say is, as an
- 6:36introvert, as I learned that PIE is important, then the question
- 6:41is, how do you get yourself to kind of get to know people? And
- 6:47I think there are different ways. Because, of course, you
- 6:49may say, you know, you can step in a room and start to talk to
- 6:52everyone, which is, personally not my style, but there are
- 6:55different ways, like, for example, engaging one to one,
- 6:59getting introduced to people planning in advance. And I think
- 7:03it's very important for introvert to understand that the
- 7:07end goal is building an asset, which is your network, your
- 7:11connections, and then from there, finding ways that feel
- 7:13natural, also to introvert people. So as an introvert, I
- 7:17would truly, you know, recommend considering this, just because
- 7:21it's something that, yes, it's easy to think networking equal
- 7:24an extrovert person, but again it's super important to do it
- 7:27and find ways that feel natural, also for introvert people.
- 7:30Tom Ollerton: Beautiful. So gonna move on now, I'm looking
- 7:33for a bit of advice, single bit of advice to make someone be a
- 7:36better data driven marketer.
- 7:38Danilo Tauro: Yeah. So I think one thought is today, data
- 7:43driven marketing is also based on tools that can help do data
- 7:48driven marketing, and there are so many tools, sometimes it's
- 7:51overwhelming. I was at the P&G alumni reunion two days ago. We
- 7:56are discussing exactly about this, which is the space is
- 7:59becoming so complex that people don't know where to start. So
- 8:02back to your question. How do you do it? My point is, just
- 8:06roll up your sleeves. Start practising. Start using tools.
- 8:10Start comparing tools. AB test so run pilots. So I think really
- 8:15getting your hands dirty is what I would recommend to basically
- 8:19learn about the space.
- 8:20Tom Ollerton: That's such good advice. When you mean AB
- 8:24testing, you think, right, we need a email service provider,
- 8:27so you run trials with three or four of those providers. And
- 8:31what? Why would you say that over just running a pitch, for
- 8:34example?
- 8:34Danilo Tauro: Yeah, so I think I'm sure we're all familiar with
- 8:37the concept of almost paralysis by analysis. So the point is,
- 8:42when there is so much out there, we try to build the perfect
- 8:46plan. We try to discuss and litigate options. And I think
- 8:50the point is, you never get to doing stuff. And I think this is
- 8:54actually slowing down the learning and the ability to
- 8:57become a better data driven marketer. So what I'm trying to
- 8:59say is like, really start with this approach of a roadmap of
- 9:04proof of concepts of test to get the learning and the strategy
- 9:09can emerge later on, after I think people really learn hands
- 9:13on what it means to run campaigns, email campaign, paid
- 9:16media campaigns and so on and so forth.
- 9:24Tom Ollerton: This episode of the shiny new object podcast is
- 9:27brought to you in partnership with madfest. Whether it's live
- 9:30in London or streamed online to the global marketing community,
- 9:34you can always expect a distinctive and daring blend of
- 9:37fast paced content, startup innovation pitches and
- 9:40unconventional entertainment from madfest events. You'll find
- 9:43me causing trouble on stage, recording live versions of this
- 9:46podcast and sharing a beer with the nicest and most influential
- 9:49people in marketing. Check it out at www.madfestlondon.com.
- 10:01So we're gonna get onto your shiny new object now, which is
- 10:04AI agents. So I think a lot of people listening to this will
- 10:09probably know what that is. But what is an AI agent to you? And
- 10:13why is it your shiny new object?
- 10:15Danilo Tauro: Yeah, yeah. I know it's very plain to talk about AI
- 10:18agents today, because everybody is talking about them. But I
- 10:22think the thing that is very relevant for me is like, how do
- 10:25we move from intelligence to activation? So to your question,
- 10:29what is an AI agent? It's an intelligent system that can
- 10:33activate and work across different tools. So if you think
- 10:38chatgpt, so that's intelligence. You ask a question, you get an
- 10:42answer. So that's intelligence, if then you have somebody that
- 10:45is intelligent and can activate tools, for example, through APIs
- 10:49or UIs, and take the output from one tool plugging into another
- 10:54tool, and basically manage a full workflow with different
- 10:57type of operations, in parallel, in series, and so on and so
- 11:00forth. I think that's basically an AI agent team that can
- 11:05basically get can accomplish many different things, and why
- 11:09it's my shiny new object. Because I think we talked about
- 11:12the power of big data, advanced analytics for the last maybe 15
- 11:17years or 10 years, and I think AI agents are bringing all of
- 11:20this to life, because now finally, you can break the silos
- 11:24across the different systems. You can make sure that things
- 11:28get done at scale fast. You can iterate. So AI agents are
- 11:33basically bringing to life a vision that existed 10 years
- 11:37ago, but can become reality today.
- 11:39Tom Ollerton: So give me an example of that working. So
- 11:42conceptually, I get that :disparate data sets, big data,
- 11:47large volumes, and that means teams talking to each other, but
- 11:51you stick an AI agent over the top, and you can say, x, y, z,
- 11:55help me with that, and it pulls together the disparate data
- 11:58sources. So in in your worldview, now, how should data
- 12:03driven marketers be using that?
- 12:05Danilo Tauro: Yeah, so let's take, I love the question,
- 12:07because we can take a very specific use case. So if we
- 12:10take, for example, the media supply chain. I call it supply
- 12:13chain because there are specific steps. So you have strategic
- 12:17planning, you have tactical planning, you have buying, you
- 12:20have optimization, you have reporting, you have measurement,
- 12:23you have financial reconciliation. So these are
- 12:25kind of high level the steps of a media supply chain. So then
- 12:29the question is, how can different steps rely on
- 12:33different data sets and different tools? So now the
- 12:37question is, if you need to manage all of this, you can
- 12:40either do it manually, with an agency, you can in house. But
- 12:44these are different people that work on different teams that
- 12:47need to pass data in different ways, ideally in a more
- 12:50digitised way. But sometimes it's like emails and Excel
- 12:53spreadsheet. So what I'm trying to say is like, if you have AI
- 12:57agents that are able to connect the different systems,
- 13:00understand the KPIs, understand the relationship about all of
- 13:04them, and understanding how to optimise the system end to end.
- 13:07The end, they can basically work through APIs or through UIs. My
- 13:11point is you can tell the AI agent what is the expected
- 13:15outcome, and then the AI agent will take care of the full
- 13:18workflow end to end. And this is what Mark Zuckerberg, in one of
- 13:22the latest earning calls, said, Oh, as Facebook meta, you will
- 13:26tell us exactly what you're trying to achieve. We'll take
- 13:28care about the rest. So this is basically what he has in mind.
- 13:32Tom Ollerton: So it's an interesting one. So let's go
- 13:34deeper on the Facebook pitch there. I mean, that's that's
- 13:37Facebook saying how great their their black box is, right So,
- 13:41but isn't the advantage really, with agents to pull across
- 13:44different black boxes and different data sets, whereas,
- 13:48if... I didn't see that, but if Mark Zuckerberg said, don't
- 13:51worry about it, we'll take care of it. That should ring alarm
- 13:54bells to most brands.
- 13:55Danilo Tauro: Yeah. So I think clarification. So that's the
- 13:59vision. I think, if you ask me from a marketer standpoint. So
- 14:03my recommendation would be to be very clear about who owns
- 14:07inventory and who manages your workflows. So what I'm trying to
- 14:11say is like, if some companies own media inventory and manage
- 14:17the process of allocation of budgets across different
- 14:20channels, there may be basically a conflict of interest, but if
- 14:24you have media owners and separately you have basically
- 14:27people that can manage your planning, your buying, I think
- 14:31you reduce the conflict of interest. So again, the
- 14:33Zuckerberg was an example, but the recommendation would be to
- 14:36have a clear separation between the technology for the planning,
- 14:40the buy and so on and so forth, and the media inventory.
- 14:43Tom Ollerton: And so who, for you, is creating the best agents
- 14:47at this point in time, like you described, a very specific use
- 14:51case and a philosophy which I'm on board with. But who's doing
- 14:54that? Are you building those yourselves, or is Have you seen
- 14:57them provided by certain suppliers?
- 14:59Danilo Tauro: Some. We at Aperiam, we are not building
- 15:03agents, but we are investing in companies that are actually
- 15:07building AI agents. And I think the secret is invest in
- 15:11companies that build AI agents for or that start at least from
- 15:17very specific use cases, because I think the competitive
- 15:20advantage is not in the underlying technology, which, to
- 15:24be honest, is almost getting commoditized. It's the ability
- 15:27to really understand what's the use case, how to prioritise, how
- 15:31to make sure that the outcomes are good, how to make sure that,
- 15:34basically, the different KPIs you may want to optimise for,
- 15:37like latency, outcome and so on and so forth. You can really
- 15:41optimise for that. So, long story short, we're not building
- 15:44AI agents, but we're investing in companies that are building
- 15:47agents that solve specific marketer problems.
- 15:50Tom Ollerton: So an interesting situation I'm seeing at the
- 15:55minute. So I use otter.ai transcribe all my meetings,
- 15:59everything, everything, everything. I love it. I love
- 16:01not having to take notes and have all in one place. But then
- 16:04otter has a chat function in it, right where you say, oh, what
- 16:07you know? What were the main points? What did Danilo have to
- 16:10say? But I didn't write thoseprompts. I didn't write how
- 16:16it responds, whereas what I can do is take that out of otter and
- 16:20put it into chat GPT, where I can build a project and build my
- 16:23own agent, really a meeting, a meeting agent, if you like, but
- 16:27I design that to a degree on top of what chat GPT has built. So
- 16:31how do brands and data driven marketers fight against the
- 16:37situation where there'll be like an agent, an agentic layer over
- 16:40the top that they didn't write. So I'm, I know someone at P&G,
- 16:45and I'm, yeah, I'm using this agent, but I didn't design
- 16:49what's inside that agent. How can I be sure that it's not
- 16:51hallucinating? Like, what's your view on that? Because it is,
- 16:55it's, it's kind of putting a black box on top of a black box.
- 16:58Danilo Tauro: Yeah. And I think this is a fantastic point,
- 17:01because the point is, it's important to understand the
- 17:05space to be able to lead, versus being led by technology. I think
- 17:10this is a very important point, because by not understanding the
- 17:14space, things turn into a black box, and then there is no
- 17:18control, there is no governance. There are risks with privacy,
- 17:22with ethics, and so on and so forth. I think understanding the
- 17:25space gives you a chance to turn that black box in something that
- 17:29is much better controlled, where governance is in place. So long
- 17:33story short is knowledge gives you the power to control and
- 17:38avoid the black box effect that you just mentioned.
- 17:47Tom Ollerton: So one of the frustrations I have with the
- 17:49industry at the minute is everyone's getting very excited
- 17:52about using AI to do what they were doing before, but faster
- 17:55and cheaper. And there's a lot, there's a lot of people patting
- 17:58themselves on the back, thinking, Oh, I used AI to save
- 18:01myself 20% production efficiency and so on and so forth. But I
- 18:05argue that everyone could do that, everyone can get a licence
- 18:09to and off they go. Whereas I think with AI, the opportunity
- 18:13is to do something you couldn't do before. And what you're
- 18:15talking about here is something we couldn't do before. So I I'm
- 18:19on board with what you're saying. So if someone's
- 18:23listening to this podcast and they're thinking, Danilo is
- 18:26right, how do I get started? What's the what's the entry
- 18:28level move here?
- 18:30Danilo Tauro: So one, so two thoughts that I shared before,
- 18:33which is, one, get started. You know, talk to AI agents,
- 18:39company, understand what they're doing, compare them. Try to do
- 18:42some pilots. So really roll up your sleeves and do something.
- 18:45Because I think again, if we start with the strategy, we may
- 18:48lose a lot of time and waste time with this paralysis by
- 18:51analysis. So really get started to understand better. And then
- 18:55the second thought is, before I mentioned the media supply
- 18:58chain, so I called it supply chain. And I think it's very
- 19:03important to understand how things are done today, and
- 19:06categorise all the different steps, the sequencing and so on
- 19:10and so forth, because that supply chain mindset is, I
- 19:13think, a great fit for AI agents, because it's very
- 19:17important to understand not just the technology, but also how to
- 19:20structure, connect, integrate and synchronise the different
- 19:25agents. Now, people talk about synthetic teams, so humans and
- 19:29agents. So having an understanding which is supply
- 19:32chain oriented, I think it's very important. So that will be
- 19:36Tom Ollerton: And then, what is your advice to suppliers that
- 19:36the combination I'd recommend to get started.
- 19:40are feeding into those agent systems. So say, for example, in
- 19:45that media supply chain, you've got a you know, you talk about
- 19:48measurement, so optimization, you talked about deployment and
- 19:52so on and so forth. They all speak different languages, and
- 19:55they all output their data via an API in or not even in an API,
- 20:00might be decks or spreadsheets or whatever it is, they're
- 20:03they're all going to talk their language, right? You know,
- 20:06because as a supplier, we have our language that distinguishes
- 20:09ourselves from other ad tech providers, right? But that's
- 20:12kind of a blessing and a curse, in a sense, because our output,
- 20:15our language, our words, our our acronyms, are different to
- 20:18everyone else's. So what's your advice to people in that supply
- 20:22chain to make sure that what's coming out of their system is
- 20:26congruent with everything else that's needed?
- 20:28Danilo Tauro: Yeah, so I think this is an important point, and
- 20:32it goes back for me to the concept of orchestration. So I
- 20:35think right now that systems are more complex and more
- 20:39interconnected, I think it's going to be very important for
- 20:42every single player within those supply chains, which could be a
- 20:45media supply chain, a creative supply chain, would be very
- 20:48important to understand how to interact with all the other
- 20:51systems, how to make sure that the data have the right
- 20:54taxonomy, the right schema, they're readable, they're
- 20:57understandable, because if not, they're going to be left behind
- 21:00just because the different systems that we need to get
- 21:03input, we prefer to work with other systems. So. Long story
- 21:07short, if before as an entrepreneur, there was a focus
- 21:11almost single minded on your technology and the problem
- 21:14you're solving, I think now it's also important to understand how
- 21:18you fit the broader ecosystem to create an experience which is
- 21:21frictionless for the end customer, which is the buyer,
- 21:24which is the brand.
- 21:27Tom Ollerton: So we're coming to the end of the podcast now,
- 21:29unfortunately, but if someone wants to get in touch with you
- 21:32to discuss agents or anything else we've talked about today,
- 21:35where's a good place to do that and what makes a message that
- 21:37you'll respond to?
- 21:40Danilo Tauro: I'm very active on LinkedIn, so I love to get to
- 21:42know new people through LinkedIn. So please feel free to
- 21:45reach out and yeah, super passionate about anything
- 21:49advertising technology, marketing technology, AI is a
- 21:53big investment thesis for us at vaperium. So anything that is ad
- 21:57tech, martech with AI in it, I think I would love to learn
- 22:01more.
- 22:02Tom Ollerton: Brilliant. Thanks so much for your time.
- 22:04Danilo Tauro: Thank you,Tom.
- 22:11Tom Ollerton: hi just before you go, I'd really appreciate it if
- 22:15you could take the time to write a review of the shiny new object
- 22:18podcast on Apple podcasts or iTunes, whatever it's called
- 22:22these days, or whichever podcast provider you use, we're at any
- 22:26podcast, so it would go a long way for us if you could just
- 22:29share the Word and give us a bit of a support on those channels.
- 22:32So just be fantastic if you haven't got time. That's also
- 22:35cool. And yeah, if you could tell your colleagues about the
- 22:37podcast, and also, if possible, don't forget to subscribe, and
- 22:42I'd love to hear your feedback. If you'd like to speak on the
- 22:45podcast or be a guest, or you think I'm asking the wrong
- 22:48questions, anything. I'd be super interested to hear what
- 22:50you think. So please email me at Tom at, automated creative.net,
- 22:57that's Tom at, I'm not gonna bother spelling it anyway.
- 23:00You'll work it out. Thanks so much.