The business world is finally starting to understand AI’s true potential: not a magical solution to all our problems, but a useful complement to our existing strategies.
For Francis Brero, co-founder and CRO of MadKudu – a marketing operations platform that helps companies identify, qualify, and engage with their leads based on how likely they are to convert – this actually opens up a whole lot more possibilities.As a former mathematics tutor and data scientist who eventually converted to sales, Francis is obsessed with leveraging data to improve the way companies conduct their revenue operations. And so naturally, he’s got some strong opinions about automation, AI, and how they can help you create better experiences for both potential and existing customers.
We recently spoke with Francis to learn more about how you can supercharge your conversational support efforts with automation, the limitations around AI, and putting in the time with your customers. Oh, and we’ve also chatted about what our favorite Nicolas Cage movie (yes, you read it right) can teach us about life – and ultimately, ourselves.
Here are some of our favorite takeaways from the conversation.
1. Stop automating everything
Automation is a great way to improve your workflows and customer satisfaction, but only if it makes the customer journey easier. If you’re automating things for no other reason than to save you the trouble of answering phones, or if it’s creating friction and making customers jump through endless hoops when they’d rather just speak to a support agent, then, as Francis points out, it’s actually a problem, not a solution. Figure out where conversational support and automation can have the highest impact, but always make sure to include backstops so your customers can get human support when they need it.
“I’m seeing a lot of people and companies out there who are essentially patching their lack of strategy with an abundance of technology. When they don’t really understand what the customer journey is, or what it should be, or where their potential roadblocks lie, they’re just slapping on technology and saying, ‘Okay, let’s just put some conversational support and everything’s going to be fine and dandy.’ I don’t think that’s the right way to approach this. It’s critical to always start from the viewpoint of the customer and identify friction points, roadblocks, and potential accelerators where conversational support might be helpful.”
2. Adjust expectations around AI
AI is not the silver bullet it was once thought to be: it can’t write the perfect response for each customer, give us the perfect insight on each persona, or fix all the issues in our services. (That’s a good thing – too much AI can be, well, creepy.) AI is only helpful when we have a specific problem in mind that could benefit from a bit more scale. Everything else is still very much up to us.
“Start with something that works but needs to scale and use AI and automation to increase the scale. The most important part is to spend time doing user discovery: meeting with your CS team, reading through the tickets, doing all of that discovery work to identify the common threads you’re seeing in there and how you can make the process of addressing them frictionless. Ultimately, it goes back to identifying where AI will have the highest leverage in solving a problem.”
3. Put in the work with your customers
One of the things Francis credits for his success at MadKudu, and something he recommends to everyone, is putting in the work with customers. When they signed one of their first big customers, Segment, Francis would actually go and work from their offices for an afternoon a week. This gave him the opportunity to build deeper relationships, hear their feedback, and understand what the levers of adoption and retention were really going to be. While an onsite relationship with customers might not be so easy to forge these days, the key point still stands: listening to, and investing in your relationships with, your customers pays off.
“The main thing is creating space in your relationship with the customer to make sure that you can talk about things that don’t necessarily directly relate to your product, and going a little bit beyond what your product does to really understand what the other questions they have that somewhat relate to your space. First off, to understand how important you might be in the grand scheme of things, but also to build empathy and build the relationship. A lot of good things come from it.”
Caught your interest? We’ve gathered a list of articles, videos, and podcasts you can check out:
- The Ultimate Guide to Conversational Support
- Strategies for modern, conversational support
- Are automation and AI bullshit?
- Why MadKudu pivoted their company’s focus from churn risk detection to conversion prediction
- Accelerate sales with data
- Lead Qualification Using AI with Francis Brero of MadKudu, on Serve Don’t Sell
This is Scale, Intercom’s podcast series on driving business growth through customer relationships. If you enjoy the conversation and don’t want to miss future episodes, just hit subscribe on iTunes, stream on Spotify, or grab the RSS feed in your player of choice. You can also read the full transcript of the interview, which has been lightly edited for clarity, below.
From churn to acquisition
Dee: Francis, you are so very welcome along to Scale by Intercom today. We’re delighted to chat with you about your work at MadKudu. Would you mind giving us just a quick bit of background about yourself and how you ended up founding that company?
Francis: Absolutely. Well, thanks a ton for having me. I’m excited to be here. Long story short is I studied engineering back in France and came to the US for an internship at a startup. I was actually supposed to start in Connecticut, but before I started, they told me the company had moved to California to start raising VC money. So I joined, and about eight months later, we raised our Series A with Sequoia. I kind of fell in love with the whole startup ecosystem and actually met my co-founders at that company.
Essentially, one of the things we were realizing is that we were a predictive analytics tool working for a lot of retail companies, trying to help them figure out which customers have more potential than others and what kind of products to surface to them. And one of the things we found was that even though we were doing a lot of predictive analytics and just analytics in general for our customers, we were doing very little for ourselves, and we were not really doing a good job at using data to assess the health of our customer base.
And as I was starting to look at the tools out there to help with that, in those early days of Gainsight, Totango, we found that while a lot of those tools didn’t necessarily have the strongest analytics behind them, they had really good processes. So we started MadKudu with the idea of figuring out how could we help B2B SaaS companies leverage all of their data to better assess the health of their customers. Anyway, we slowly migrated away from this kind of churn, retention, prediction play, and more towards an acquisition play.
“Once people are exposed to the product, you want to make sure they are onboarded properly and understand what’s going on there”
Dee: My next question was going to be about that shift from churn prediction to sort of conversion predictions. Have the events of the last year affected that at all? I think there’s been a big shift in focus to retention is the new growth and nurturing existing customers rather than capturing new ones. I’m curious to know if the pandemic has had a knock-on effect on what you’re doing there.
Francis: I wouldn’t say that it has on our side. I think what’s been interesting is to see that the teams are typically fairly different. One of the things we’ve seen is more of a shift at least towards product-led growth and more product-centric means of monetization and acquisition. Which, to some extent, kind of plays into more of the CS part and how to do support. Because once people are exposed to the product, you want to make sure they are onboarded properly and understand what’s going on there.
Because there’s been this big shift on the marketing side of not being able to run field marketing and not being able to do events and webinar fatigue, there was a question of how to get people to engage with brands and with our products. And so, what we’ve seen for a lot of customers is that there’s been a strong push towards a PLG approach. And that, I think, has put support on the forefront again, at least on the go-to-market side, because now it’s raising a lot of questions of how to make sure the initial onboarding is frictionless and how to make sure we’re providing the right level of support for the right people at the right time.
Automation, but make it frictionless
Dee: It’s funny you should mention webinar fatigue because we met recently at a webinar, which was around Intercom’s Ultimate Guide to Conversational Support. I’d love to hear a bit more about your thoughts around conversational support and how you’re implementing this at MadKudu. And what were some of the challenges that you were trying to overcome with it?
Francis: I think conversational support is a very interesting topic. One of the challenges that I run into, both at MadKudu and in general, when trying to help our customers, is that you always want to make sure that it stays helpful without becoming robotic. And, at the same time, it has to be helpful without being harmful to productivity. So, I guess there’s kind of two levels of challenges. One of them is really conceptual, and the other one is kind of a people process.
If we start with the more conceptual part, I always view these things through the intersection of this trifecta that anything you build has to be desirable, feasible, and viable to be successful. And I think, from a viability standpoint, conversational chat and conversational support have become fairly straightforward. Tools like Intercom do a great job at making it fairly cheap to do, and it’s feasible because you have all the tools to do it. Then there’s a question of how do you make sure you have the right data in there.
To me, the most challenging part is really on the desired part and making sure that the automation and the workflows we’re building are actually helping the customer and that they’re making the journey frictionless, not actually adding friction. I think that’s one of the things where I see a big pitfall, where oftentimes, people are just pushing conversational support because it is feasible and it’s more viable than having someone talk to their customers, which I think can be very dangerous. There are some cases where it’s a very simple question, and conversational support is the right way to assist them. But there are some cases where the white-glove approach of having a human pick up the phone, do more discovery and spend more time with the customer is going to be critical.
“I’m seeing a lot of people and companies out there who are essentially patching their lack of strategy with an abundance of technology”
And I do think we live in this era where there’s a bit of the famous paradox of choice of not knowing what to do. Products out there have more and more features. If we take Intercom as an example, there are so many kinds of sub-products within Intercom, and there’s always a question of which sub-products are the right fit. Even if I think of MadKudu, there’s a lot of things we could be using, but there’s a question of, should we actually be allocating mind share to using them at our given stage? And I think that’s something that benefits from interacting with another human being, with an expert of Intercom, and that person telling us, “Look, at this stage, this is going to be better, but then, as you grow, this feature might be helpful.”
And that’s where I think the conversational support kind of hits its limits because you’re not going to have the same bandwidth and interaction, and everything is going to be a lot slower, and it can be very frustrating on the customer side. So, from that conceptual part, it’s critical to really figure out what is the right channel to help customers.
And then, if I think more of the challenges on the people process side, we live in a time where shiny tools are getting all the attention, and I really worry that this could be very detrimental to the customer journey. What I mean by that is that I’m seeing a lot of people and companies out there who are essentially patching their lack of strategy with an abundance of technology. And so, when they don’t really understand what the customer journey is or what it should be or where their potential roadblocks are, they’re just slapping on technology and saying, “Okay, let’s just put some conversational support and everything’s going to be fine and dandy.” I don’t think that’s the right way to approach this. It’s critical to always start from the viewpoint of the customer and identify friction points, roadblocks, and potential accelerators where conversational support might be helpful.
And that’s where you want to implement it. I’m not saying that there’s nothing you can’t discover with conversational support because I think you can. But in the vast majority of cases, it’s important to have a clear problem and to implement conversational support towards that problem.
Dee: That makes a lot of sense. And it’s something that Kaitlin Pettersen, our Global Director of Customer Support, talks about quite a bit, that idea of making sure that at whatever stage of the conversational support funnel you’re in, you’re getting the right content or the right response to that customer at the right time. And that could be an automated response, or it could be, as you say, the white-glove human approach. You, as an organization, have shifted from churn prediction to conversion predictions. How has this shift towards conversational support benefited you so far?
Francis: It’s been helpful in many cases. In some cases, it’s been helpful, and in some cases, it hasn’t. I think what it has done is it has forced us to figure out what are the different categories of questions and the different categories of support customers are looking for.
“I spent 10 or 15 minutes going back and forth with this support bot when I could have written my question in 30 seconds if I had a form”
So, one of the things we’ve done is we’ve run some analysis to figure out what are questions that come up frequently, and what are questions that actually don’t require that much context or don’t require very strong human interaction to get more discovery and more solutions. And those can be automated to some extent in conversational support. But it’s been a big question of, “How do we align our support infrastructure with the customer expectations?” I think that’s the most important part at the end of the day. It’s all about the viewpoint of the customer, understanding the customer journey and adding tools that actually help make that customer journey easier.
One thing that we’re seeing is that people are becoming increasingly used to leveraging conversational support when they’re looking for something. So, that’s a part that makes sense. But oftentimes, what we’re seeing is people don’t know what question to ask, and that’s when it’s really important to have some kind of a backstop to make sure that you move out of the typical journey of the conversational support and go towards an agent or an operator. Otherwise, the whole system can become very detrimental to the user experience. And this actually happened to me a few weeks ago. I was on a support website for one of our providers, and I was trying to add a new credit card to our billing page.
Dee: Which should be a simple enough function to do through self-serve.
Francis: In the first place, it should be easy, right? For some reason, it wasn’t. I went to their support page, and I had a very clear question, I just didn’t know why it was not working. They had conversational support, and it kind of walked me through all the questions like, “Hey, what’s your first name?”, “Can we help you?” They had a list of all the different topics, and I spent probably 10 or 15 minutes going back and forth with this support bot when I could have written my question in 30 seconds if I had a form. In that case, the conversational support was highly detrimental to my experience, and in other cases, it can be positive.
“It’s forced us to have very deliberate and intentional conversations internally in figuring out where we want to have automation and where we don’t want to”
From our perspective, it’s really forced us to have very deliberate and intentional conversations internally in figuring out where we want to have automation, where we don’t want to, and how we make sure we have clear backstops so we don’t pull people down this kind of endless interaction with a bot that can be incredibly frustrating.
Dee: For sure. And again, that is something that Kaitlin kind of tries to hammer home as often as possible. On a more general note, knowing that you’ve pivoted in that direction, do you think it’s important for companies to be ready to shift or pivot like that in a moment?
Francis: Yeah, I think it is. The companies that win are the companies that are able to really adjust to customer expectations, and I think, ultimately, delivering the right product in the right manner to the people that are trying to get it is what’s going to determine winners.
I don’t necessarily agree on a philosophical level with how our society now interacts with Amazon, where we have this kind of need to have everything delivered to our doorstep within five hours of us ordering it online because we’re craving whatever. But ultimately, it’s a customer expectation, and it’s really a question of, if you’re not going to meet those expectations, you have to over-index on some other customer expectations to make sure you’re going to be able to retain the business. Otherwise, you’re going to lose that business to Amazon.
I think there’s something similar there in terms of figuring out how to support your customers and how to make sure you’re providing the value and making the journey frictionless. Because if not, that’s when you start opening the door for your competitors to come in and swoop by.
Investing in proactive support
Dee: When we did that webinar, Francis, you were explaining how MadKudu is developing more proactive, self-serve options to make your product stickier and to increase adoption. Do you want to tell us about that?
Francis: Yeah, absolutely. We spend a lot of time looking through the different user interviews we did, actually watching user sessions in Hotjar, one by one, trying to understand where people are struggling. What are the answers that customers are looking for but not necessarily finding in our app?
We’re also looking at our ticket systems to identify the frequent questions and using that as a basis for the proactive questions we’re going to ask. We can use some basic behavioral patterns to try and identify what someone might be looking for based on what they’re doing or what page they’re looking at, but also to know what are typical types of questions that customers might have.
“If we do a better job at making it easier to find the answers to those 20% of questions, we’re actually solving the problem for 80% of the interactions with our customers”
Like most companies, we have the 80/20 rule of 20% of the questions actually account for 80% of the volume. If we do a better job at automating or making it easier to find the answers to those 20% of questions, we’re actually solving the problem for 80% of the interactions with our customers, which can make the journey a lot more frictionless. That’s where we’ve been able to make adoption better. We’re making it easier for people to get the answers they’re looking for and therefore move on to the next step – adopting the product.
Dee: That makes a huge amount of sense. Staying on support, have you noticed a change in the customer’s preferred support channels since the onset of the pandemic? And how have you managed it?
Francis: That’s an interesting one. I can’t say we’ve seen a drastic change in the expectation of support from our customers. I think the main change, as I mentioned, was more on the acquisition side, where people have less of this in-person kind of events to meet up, may it be a rep going on-site or meeting up at an event. We’ve seen an increase in the need for interactions through Zoom or other mediums that can give you a high bandwidth interaction.
On the CS side, we haven’t necessarily seen anything that different, we’re still seeing a lot of customers use our ticketing system. One of the things that’s interesting, though, is that we’re getting close to the point where it’s pretty obvious that when customers open a ticket via a form, it’s indicative of less urgency in the resolution versus when they’re pinging us through conversational support that typically indicates that this is a little bit more urgent. That’s kind of an interesting trend that we’ve started seeing.
I don’t think it’s related to the work from home situation, but rather from the fact that we’re starting to have these two more mature support channels, with the ticket system being understood as a very asynchronous channel and the other one with an expectation of synchronous interaction. This has led to some of the challenges we were talking about – one of them is making sure we have the capacity to address these expected synchronous interactions because otherwise, it goes back to being detrimental.
Getting real about AI
Dee: Back in 2017, you wrote a really, really interesting article called “Are Automation and AI Bullshit?” I’d love if you would share your thoughts then as they were with the audience, and maybe, if and how they’ve evolved in the intervening period.
Francis: Absolutely. Yeah, I like to try and find some contrarian views, but I think things have changed. So back then, the main point I was trying to make is that there’s a very fundamental and critical distinction to make between what AI is today and what we would like it to be, and I used to call it the mythical AI versus the pragmatic AI. I see AI as a hammer looking for a nail, and people who don’t have a clear problem in mind are kind of hoping AI is going to help them solve anything.
For example, I ran into this product a few weeks ago called Copy.ai. It’s super interesting, it basically uses one of the libraries from the OpenAI project to generate text, and it essentially generates content for marketers. I mean, super fancy, it’s very cute. But ultimately, if you don’t spend the time understanding who your audience is and what they care about, why should they even spend time reading your content? What’s interesting about your content? No matter how smart, how fancy, and how good the words are, if you don’t deeply understand who the customer is, it’s not going to solve the problem.
And that’s what I see as the mythical AI, where people hope that there can be this AI that’s just going to write the perfect content for every single one of their customers, it’s going to adapt the wording based on who they are. There were all these things a few years ago where it would tell you the predicted psychology of your prospects: “Oh, for this prospect, you should be more assertive. And for this prospect, you should use more emotions.” Again, super cute. But ultimately, people are going to read your email because you’ve identified a core pain point and you’re explaining how you’re going to solve it. We all want to believe in the fact that we could work one hour a day and have AI solve everything for the rest of the day, but that’s not going to happen.
“I’m seeing people be more educated and have more realistic expectations around what AI does”
However, there’s a more pragmatic use for AI if we look at bidding strategies on ads or some of the things that MadKudu does of trying to do predictive routing, like which leads should go to your enterprise versus your commercial versus your self-serve funnel? These are things that have been identified as a challenge, and you have a clear initial process that is starting to reach its limits, and this is where you want to use AI to solve it.
To answer your original question, the thing that has changed is that, in general, the market has matured in its understanding of what AI is. I’m seeing people be more educated and have more realistic expectations around what AI does. A few folks have been burnt by AI vendors, and they bought these things hoping it would solve everything. MD Anderson invested millions in a contract with IBM Watson. Very early on, they started saying, “Oh yeah, we’re going to cure cancer with this IBM Watson AI because we’re able to scan through all the medical papers and all that.” I think it was like a $60 million contract or something like that. And they basically pulled out after a few years, realizing that it just didn’t work. I think these kinds of events had a massive impact on the market. We realized we have to tone it down a little bit, and we have to be realistic in how we approach AI.
To some extent, I see a parallel today with blockchain and DeFi, in the sense that everyone is throwing it everywhere without fully understanding what it does and what the current limitations are. It’s a very hot topic. And I mean, it’s great, right? We’re in that phase of mass exposure, people are excited about it, and that’s going to lead to more education. Unfortunately, it’s going to lead to a couple people burning their wings because they’re going to believe in it a bit too much before doing their research. But over time, we’re going to understand that system better. I think AI is in a much better place than it was back then. I still think there’s a lot of BS out there and I still think there’s a lot of data science teams building AI models without fully understanding why they’re building them, and that I think that’s a problem for organizations.
Making AI less creepy
Dee: The way you describe it almost makes it sound like a version of alchemy, this idea that you can turn data into gold, but that idea that AI could be more intuitive to human emotions than a human being sounds like an impossible proposition, or not necessarily something that we would want to see happen. Have you noticed any changes in how you think about it over that period of the pandemic?
Francis: I’m not sure COVID has changed as much as I think the regulation has. One of the things that have been really interesting to see is how GDPR and CCPA, which is the kind of Californian version of GDPR, is bringing up a lot of really good questions. Because it’s starting to question how we think about models even philosophically. One of the questions is: If I train a predictive model with a dataset and one of the people that generated the data used in training the model decides that they want their data to be forgotten, do I consider that I need to retrain the model without their data? How disconnected is the model from that individual user’s data?
“You walk into a store and this sales rep you’ve never met in your life starts asking you about your trip with your kids five weeks ago. That’s super creepy”
Put more simply, it brings up a lot of questions around how we use AI at a personal level versus an account level. Basically, where does it get creepy? The typical example I give is it’s the same thing that a good salesperson would do. A good salesperson is going to do pattern recognition based on how you’re dressed, how you talk, how you stand… And based off of that, they’re going to determine what kind of a customer you are. The worst thing that could happen is you walk into a store and this sales rep you’ve never met in your life starts asking you about your trip with your kids five weeks ago. That’s super creepy.
And that is, to some extent, what people were getting close to thinking about doing with AI in marketing. They were starting to say, “Oh, we can go and scrape everyone’s Facebook profile. We can look at everything.” There were cases where there were some big issues with Hilton using AI and actually recommending your hotels based on stays you’d had. And of course, it has to lead to some issue where someone was there with someone other than their significant other, and Hilton sent recommendations that were seen by the actual significant other.
Anyway, there’s a lot of privacy infringement questions around how we use AI. And I think what’s interesting is that if you were to do this manually, you would ask yourself a lot more questions. If you were to write every single email by hand, the way you would approach how you communicate with that person would be very different than if you were thinking about it from a pure AI perspective. And that, I hope, is something that is going to continue. Privacy regulations are forcing us to think about what we would consider privacy infringement if we were writing that email ourselves, and it’s taking that level of scrutiny and applying it to anything you do with AI. Before, I think we were kind of thinking about that mythical AI, saying, “Well, let’s just dump all the data from the internet into this gigantic deep learning algorithm and it’s going to tell us exactly how to talk to every single potential customer in the world.”
“Ultimately, it goes back to identifying where AI will have the highest leverage in solving a problem”
Dee: It’s such a good point you make about the distinction between personal data and account-based data. And certainly, over the last five years, at the very least, I think people, even my own mother, have become more attuned to their own role within a data set and what that means. Maybe in Europe, as you said, that has been largely inspired by GDPR being in the news so much.
To jump back to the more support side of that, for companies that have yet to leverage automation or AI as part of their conversational support framework, what would you recommend as a good starting point so that you don’t end up going into some of those pitfalls?
Francis: I would say start with something that works but needs to scale and use AI and automation to increase the scale. The most important part is to spend time doing user discovery: meeting with your CS team, reading through the tickets, doing all of that discovery work to identify the common threads you’re seeing in there, and how you can make the process of addressing them more frictionless. Ultimately, it goes back to identifying where AI will have the highest leverage in solving a problem.
Typically, that’s where I see AI as an evolution of automation. You do things at a very low scale and then you realize, “I’ve been doing this over and over in the same way, so I’m going to start automating it.” And then you start adding a little bit more complexity. When automating, you need to have all these rules, and basically, AI is a sophisticated rule engine that can help your automation.
I think it always goes back to starting with the user discovery meeting with the CS teams, spending time and putting the work in to understand where you can have the highest impact. And then, rolling out small bits and pieces here and there of conversational support, and leveraging AI to increase the scale you’re at.
Different AI personalities
Francis: Yeah, I mean, there’s a couple of big plans on our side. We’re excited that we’re growing, and one of the big things we’re going to have to focus on, if we think of the B2B space, is that we have a typical kind of champion that’s using the MadKudu app, typically more of a marketing ops person, but we have a lot of either non-admin users or indirect users if you think of SDRs, sales ops, VP of marketing, VP of sales. And so, we’re going to have to spend a lot of time understanding the expectations of those customers and how they want to interact with MadKudu because it’s going to be very different from the admin user, and designing an experience that is relevant to them and provide them with the right level of answers.
A core part of that, which goes back to the whole conversation we had about AI, is we’re rethinking the UX of our AI and the AI personalities, which I think is going to be a massive topic for us. One of the typical things that we do with our AI is building lead scoring models. That’s determining if this lead should go to sales – yes or no, to put it simply. If you think of it from a marketing ops perspective, the role of that AI is really a gate. In the kind of standard AI personality terminologies, it’s considered a police personality. It’s literally saying this lead should go to sales, this lead should not – it’s making a decision and you’re supposed to act on it. And that’s, to some extent, what marketing ops is looking for.
“It’s something that very few companies have managed to solve, understanding how one AI can be presented in different ways to different users”
If you look at it from the perspective of an SDR, you don’t want to have an AI policing you and telling you, “Oh no, don’t go talk to this lead,” or “Go talk to that lead!” What SDRs typically want to see from an AI is more of a buddy-type personality, an AI that’s going to give them a competitive edge by telling them, “This is an interesting data point that you could leverage in your outreach,” or “This is a talking point that you should use,” or “By the way, this lead is probably a better leader for you to work with than this one.”
And so, the UX has to be very different because if you try to push a police-style AI onto SDRs, it’s going to be met with a lot of resistance, and it’s going to create a lot of friction versus actually getting them to adopt something that they perceive as something that’s enhancing their user experience. That’s going to be a really, really interesting challenge on our side, and it’s something that I think very few companies out there have managed to solve, understanding how one AI can actually be presented in different ways to different users. But I think it’s really critical for the adoption of AI across organizations. And if that succeeds, hopefully, we can continue the growth of the company. And that’s exciting to me because we have a bit of an internal bet at MadKudu that when we raise our next round and do our first user conference, we want to have Nicolas Cage come and be the keynote speaker. So that would be a lifelong dream that would happen.
What’s Nic Cage got to do with it?
Dee: That actually brings me neatly to my next question, Francis, which was one that we normally ask people in our other podcast, Inside Intercom, but I just thought it would be a fun one for you. You mentioned Nicolas Cage there. We usually like to ask people if there’s someone who inspires them or they aspire to. Tell us about Nicolas Cage and what role he plays in your life at MadKudu.
Francis: Yeah, absolutely. Nicolas Cage is an interesting one because this whole thing started as an inside joke, turned into a meme, and now is, to some extent, a core part of our culture at MadKudu. One of the questions every employee gets in their onboarding is, “What is your favorite Nicolas Cage movie?”
The part that I really like about Nicolas Cage is that he’s one of the most polarizing actors. I always find it fascinating to see how people have such strong opinions about him without necessarily doing that much research. And I find that it’s a good representation of our polarized society today without actually being a touchy subject, right? Nobody wants to talk about politics, nobody wants to talk about these difficult things. And I think he’s a good way to show that polarization on something that is not offensive.
What I also really like about him is that he takes a very contrarian view of acting. One of the main things that he says is that we have this conventional wisdom today that actors should be regarded as good representations of real life. Today, we deliver Academy Awards to people who managed to make you feel like the movie is actually reality. And I think that is something people don’t necessarily challenge. There’s no reason why that would be the only way of doing it.
“Nicolas Cage is to cinema what free jazz is to music, and free jazz does break a lot of the standards of pop music”
There’s been a couple of interviews that I thought were really interesting where he said that if you think of how actors at the opera actually play, we could think they overplay, but the reason they do that is that they’re on stage, and some people are very far in the back of the audience, and you want them to understand what is going on. And so you have to portray the emotions in a much more grandiose manner, and you have to make sure that even the person at the very back of the audience is going to be able to understand what emotion you’re trying to transmit with your movements. That’s very codified and it’s very acceptable, and there’s no reason that that wouldn’t be valid from an acting perspective in a movie.
And sure, it goes against the standards today, but why are those the standards? And should they be? The good part is that you don’t have to agree and you don’t have to enjoy the art, but I think it does bring up a good question that there are a lot of things that we take for granted and that we accept as being the right way of doing it. And sometimes, it’s just interesting to ask that question. I think a good reference is that Nicolas Cage is to cinema what free jazz is to music, and free jazz does break a lot of the standards of pop music or how we listen to scales and how we expect improvisation to work on top of the 2-1-5 scales, or whatever.
There is something interesting to it, and I think it opens up our perspective a little bit. I find him really interesting for that. It’s very deliberate, and the more you read into it, the more interesting it becomes. So I think it’s a good way to get people to become more curious and to question what they take for granted.
Dee: I love that. And maybe it also makes people a little bit more comfortable with agreeing to disagree, which is something I think is really important for work and life in general. I’m curious, though, are people allowed to use Fast Times at Ridgemont High? Because that’s the only film I know of where he’s credited as Nicolas Coppola instead of Nicolas Cage.
Francis: Yes, they can. Absolutely.
Dee: Good to know. It is a classic.
Dee: This series is all about hearing how companies scale. I would love to know if there was a key event in your career that helped you scale professionally.
Francis: I think one of the things that helped us a ton at MadKudu and helped me professionally and something I would recommend to everyone is just put in the work with customers.
In the early days, we were very fortunate to sign Segment as one of our first customers. They became a customer fairly early after their series A, and they were already a bit of a rising star in Silicon Valley, a lot of the other B2B SaaS companies were looking up to them because they were showing stellar growth.
And so, every week, I would actually go to their office – I had set up a little standing desk next to the coffee machine – and I would work there for one entire afternoon, sometimes even a bit longer. I would meet up with Guillaume Cabane, who was their Head of Growth at the time, and we would have conversations that were very specific to the projects that we were working on, but also a ton of conversations that were more related to questions he had about how their business was doing, things that he wasn’t sure about.
That allowed me to get a much deeper view into all the things that surrounded MadKudu as a vendor for Segment. It gave me exposure to their sales team because their reps would come over and we’d have conversations over coffee. We would just chit-chat here and there and I would understand where they had concerns, the things they liked about the product, the things that they didn’t understand. And that was really instrumental in understanding what the levers of adoption and levers of retentions at Segment were going to be.
“Investing time with your customers is always going to pay back tenfold”
Today, it’s a little bit harder to go onsite to customers, but I would say the main thing is creating space in your relationship with the customer to make sure that you can talk about things that don’t necessarily directly relate to your product, and going a little bit beyond what your product does to really understand what the other questions they have that somewhat relate to your space. First off, to understand how important you might be in the grand scheme of things, but also to build empathy and build the relationship. A lot of good things come from it.
That was instrumental for us because it really built a strong relationship. It built very, very strong advocacy on the Segment side, and that is one of the core things that led us to strong acquisitions afterward, where a bunch of great companies were coming over to MadKudu because they were saying, “Hey, we heard great things from Segment and we’d love to do the same with you.” So, investing time with your customers is always going to pay back tenfold.
Dee: Very good advice. And it certainly sounds like it was a game-changing moment for you. Lastly, then, Francis, where can our listeners go to keep up with you and your work?
Francis: I try to post regularly on our blog, so that would be on MadKudu.com. Then, they should feel free to follow me on Twitter and on LinkedIn. On Twitter, I’ll make more comments about Nicolas Cage, about music, and interesting topics in general. And LinkedIn is more focused on, I guess, less polarizing topics and more B2B SaaS.
Dee: As it should be. All that’s left is to say thank you very much for joining us today. I really enjoyed chatting with you.
Francis: Yeah, likewise. Thank you for having me.