“Growth hacking” is often associated with short-term goals and shady practices. But what if you can use similar growth principles to attract customers in an honest way?
My guest today will tell you how to identify the right channels to generate long-term revenue.
This week we are joined by Guillaume Cabane, an expert growth marketing advisor who has worked with top tech startups such as Mention, Drift, and Segment.
In this episode, you’ll learn how to audit a marketing channel, craft buyer personas beyond traditional demographics, and generate creative growth ideas.
It's the antidote to marketing bullshit.
Receive a free, 8-lesson video course + a super practical, no-bullshit essay in your inbox every Tuesday.
"Louis is a genius. Sharp and super useful insights."
"You're literally the only marketer I can stomach."
"Anything with Louis always blows my mind."
Louis: Bonjour, bonjour, welcome to another episode of everyonehatesmarketers.com, the no-fluff actionable nothing podcast for marketers, marketing consultants, founders and tech people, who are just sick of shady, aggressive marketing. I'm your host Louis Grenier. In today's episode you'll learn how to generate result in the long term instead of focusing on short term revenue goals using growth marketing experiment. My guest today is a growth marketing expert and he's been so for the last 15 years, he is also French, which is quite interesting. I don't speak to many French people on this podcast. He's been working on projects covering B2B growth mostly, he has a special taste for that. His CV is, I would say, rather impressive. He's an advisor for the G2 Crowd or G2 now, Hull, and MonkeyLearn. Former VP of growth at Drift, at Segment, former head of growth marketing at Mention, I mean massive respect for your experience. Guillaume Cabane, welcome aboard.
Guillaume: Thank you Louis. Thank you. I guess because we're both French, we're going to spend one hour talking about cheese and croissant.
Louis: Yeah. Which is the cliche, but that's... I miss it. As you know I don't live in France, neither do you. So we both miss it, don't we?
Guillaume: Yeah, you miss the cheese man. Thank you so much for having me today. The thanks for the recap on my experience, I think I've been very lucky, honestly. Sometimes you just got to go with what you think you'll find luck.
Louis: Yeah. There's a bit of survivorship bias in tech. You see a lot of people extremely successful, then you look back at what they've done and you try to say, well, if I do the same, I'll succeed as well. But it's not really the case. So I get a chance to meet Pascal's podcast who talk to the survivors, the people successful. But there's probably plenty of people who try to do what you did, who struggle. I know a lot of French fellows who studied in in marketing as well, who had good business schools, studied in good business school like you did. He never had the CV that you have. Do you think it's all down to luck? Do you think there's a bit of skills?
Guillaume: I think there's a big chance... I often say that if I'm not a good marketer, at least I'm a good a logo picker. I've been pretty good at picking logos for companies that succeeded, which is this strange like halo effect of like you think that someone that works at a successful company is a successful person and a good person, which honestly is not true. That person joined the right company, sure that person got in, got through the interview process. But it's very much like a business school, getting in the business school is hard, but does it mean that you're a good person or you're smart, debatable. There's a strong halo effect, that's for sure.
Guillaume: I think I'm a good picker, I've been able to pick the right companies that will succeed and grow, that my friend unfortunately is mostly luck. Because I think that I don't have as much insights as investors VCs, and good VCs, bad VCs, you know what we say, they all have a 10% success rate. There's no such thing as choosing the companies which will succeed. Listen, if you have a way to know which companies will succeed, invest, you will be massively successful.
Louis: I mean that's brings a question then, it's, so you pick Mention... You work with Mention, you work with Drift, you work with others. So what was your thought process then when you kind of started to work with them? How did you know or how did you feel, actually that sounds like a nice rocket ship to be part of.
Guillaume: It's interesting, I've never talked about it. Because I get a lot of questions, which is like how was I able to join Segment, which was a massive rocketship and Drift, which is also a massive rocketship, at a stage where there were already fairly advanced. I joined Segment when there were 50 people, just after Series B, and I left 18 months later when there were like 200 something. I joined Drift, had 50 people in it. I'm not sure people realize it's really hard, those roles, those opportunities are almost impossible to get once the acceleration curve has started. All right. Because, every one wants in. My strategy that I've been working on for a long time is, I build long lasting connections and trust. Look Segment, Segment is a good example.
Guillaume: I joined early 2016, February 2016. I've been talking to the three founders, I found an email since late 2013 and we had back and forth for three years about product and market. Then when I was mentioned, I became a customer, then I started doing speaking, I did keynotes on how marketers should use Segment. And eventually in 2015, they said, Hey G, we think that we should work together. It was not a question of like, do we trust each other? Because, we did. Did we know each other? We do. They know exactly who I am, I know who they are, I know the product, I trust the product. It just a question of like, is this the right time for both of us? So all the rest is eliminated. Okay. Drift, exactly same story. I was the first enterprise customer, when they were less than 10 people. I had pushed them to integrate with Clearbit, with Segment and a lot of other things.
Guillaume: And eventually like same, trust, connections, I knew the product, I knew the team. All right. All of that stems from me starting. I wasn't entrepreneur a long time ago, in 2012, 2013 and a failed one, I have to say. When I stopped that I became... I did a part time mentoring for free in a French incubator called Numa and a couple of other places. I started to help people and that is what has produced the cure, the network building through mentorship, then through advisory. Building connections, helping... I realized how much the Americans think you paid forward and it comes back one day, it does, it really does. It takes time,, I think if you expect something like that's the wrong... That's a mental model.
Guillaume: But if you expect nothing, if you really help it, it pays back. I think one difference between me and many marketers, is that I usually accept quick demos, or quick coffees, or calls with entrepreneurs who are at the very beginning of their project. Just want to bounce a couple ideas, they want to show me what they're working on, or want me to try their tool and get some feedback. I usually take most of those if they're, at least in my space, which is B2B SAS. Where I think I can be helpful, then I keep in touch. Some of them succeed. So I quit my own survivor bias where of course those who survive, those who succeed over time, they have seen the value. If they don't see the value that's fine, they go somewhere else, but some do. That creates the resume you see today.
Louis: You did on growth hack your way into success, into finding those companies. It took you a long time and I think that's... Isn't it a great lesson in term of what we're going to talk about in the next few minutes, because you took your time, you build trust, you treated those people like people, build relationships. You didn't expect anything, you just gave. Then you receive in return, didn't you?
Guillaume: Yeah. The same is true for me being today on the podcast, I expect nothing. When I do keynotes, I do a very high number of speaking engagements per year, I do it to help people succeed. Because, I believe I can help them. It is true that sometimes people will reach you back years after saying, hey, I've watched this, I would like you to help me, I'd like to work with you. Sometimes the stars aligned. But I think one thing people need to realize that, for the stars, for the planets to align to become a great project, you need to have a ton of things going on, because the percentage is low. It's just a conversion rate on your website. For somebody to be in love of your product, your marketing message and say, yes, I want to buy, you need hundreds of people. The same is true for job opportunities, or advisories, or whatever it is the person like. You need to have a lot of those things going on so that you can select the best, the top of the top.
Louis: That's a great summary, what we just discussed there. In term of like your specialty, growth marketing, growth hacking, it definitely has a backdrop, right? I personally think-
Louis: ... Sometimes, all right. I will be in the side on the opinion that too many use growth hacking as an excuse to use shitty short term tactics, that serve them over customers. Overall it gives a bad rap to the methodology that is not new, but has been employed in those businesses. I'm curious for you, what's the limit between being persuasive in your marketing and being scammy, spammy, shitty in your marketing?
Guillaume: I mean it's interesting. I'm in Silicon Valley, I'm in the South. I have a higher concentration in my network and people I talked to you in the day to day of larger successful startups. I have a survivor bias, sorry. That's just a selection bias in my perception, my day to day. And of course, the growth people that I talk to, there's very little scammy, shitty, or I'd say intrusive marketing going on. If you look at Airbnb, or Dropbox, or other companies like that today. Just definitely none of that going on. Because that would not fly at all. We're talking more usually new ways to penetrate markets or channels, or how to do things differently to differentiate. It's a lot around creating a segmentation model, should be able to be very efficient at experimenting.
Guillaume: You find stuff that works better than competitors. Now the bad rap comes from like if you're pressured to find a leverage, have higher performance, you don't have ideas and you don't care about the brand, although long term. Then yeah, sure, things can go South. Yeah, for sure. But if we go back to my experience, it's funny, it's like if you walk back almost 10 years now, I worked in IT security for two years as a CMO for French it security consulting firm. I worked a lot in social engineering, pen testing, and things like that for two years. They're very interesting similarities between scammers and people who do social engineering for, let's say, malicious purposes and marketers. In both cases we are trying to convince a random stranger or anonymous person on the internet that what we say is true.
Guillaume: In both cases we're trying to have them pay, or give us their credit card or some money. Both cases we're trying to get a transaction, and both cases we usually start with an ad or an email. The only difference is that marketers in the end, there's a product. Now, I have a talk recently where I said, the issue is that marketers are incentivized to over promise. I think that's where it starts to slide a bit. I explain, I think the pond comes from marketers a bit, it also comes from how are marketers incentivized, and how of the company is structured to push marketers to do the right thing or the wrong thing. Marketers are usually incentivized for a fluffy KPIs, fluffy metrics, traffic, sign ups, leads, maybe social engagements, stuff like that. All that is BS.
Guillaume: It has absolutely no value, because we can cheat on any of those metrics to create a perception of that engagement at very low quality. I'm French, so I know that cheating is, I know about cheating. It's staggering that very few organizations have marketers incentivized for the bottom line, for revenue. Okay. The only thing that matters, and I don't mean just immediate revenue, we're talking of LTV, lifetime value of the customer. Now my argument for you is, if your marketing is incentivized for the longterm value and you've seen that... I usually assume that people paying for a long time or happy customers, that's a good thing for both of us. That's my assumption, I think we can all agree on that. People stay for long time, they're happy, good, it means they get value. So it's good. That's what I try to optimize for, that's what I want organizations to optimize for.
Louis: That's the answer, in a sense. It's like, as long as your goals are really geared towards the happiness of customers, then you're very unlikely to use shady manipulation tactic. Because, it's going to bite you in the ass. This person is never going to turn into happy customer.
Guillaume: I have a slide man, that I've used recently. Where I say, my wife is a salesperson, so at the diner table, we have both sides of the conversation. Salespeople, they have claw back. For those who don't understand what is Claw back, it means as a salesperson, if you oversell, if you push a customer to buy and that person shuns within a defined timeframe, usually three months, four months, something like that, that person shuns. The company will claw back, you can imagine a bear clawing the food back. So your company will claw back the commission, take it back. You will lose the commission from Bechtel, because it has assumed that you have oversold. We have sold someone who could not be happy. That is honestly pretty standard these days, and it's a good thing. It prevents salespeople from over-promising, overselling whatever.
Guillaume: There is no such thing for marketers, but there should. There 100% should, because it's even worst, example, marketers is pressed for a results like signups. If I'm a shitty marketer, let's put my shitty marketer hat on. And I want results, I don't have any negative incentives. Only I will take a bonus if I reach signup goal, what am I incentivized to do? I will over-promise on my ads by saying, this product will change your life. It will do it in only two clicks and it's free forever. None of that is true. That's not exactly what we do, but do we do some of it. We definitely do some of it. We all put some ad saying like, Oh, as easy as two clicks. I'm like, no man, this takes a lot more than two clicks even just for signing up.
Guillaume: We minimize the perception of friction, of effort, and we maximize a perception of value. When actually we're not honest, we're dishonest in some ways. What we do when we're doing that, is we're pushing the friction downwards in the funnel, on the sales, on the CS people, on the company. The disillusion will happen, it is going to happen, just not within your org. It's going to happen somewhere else. That is what we need to solve. So the marketing tactic, the growth marketing only matters if you create longterm value, then it's fine.
Louis: Let's take the scenario of you starting to work for B2B SAS company, I mean that's our specialty. We'll take that as an example. As I mentioned in the intro, you're an advisor to multiple, so you know exactly this scenario. Let's say you start working with a new one, you might want to take a real example or fictional one depending on your NDAs and whatnot. It's up to you. But let's say they come to you with a problem, like usually they want to grow more. Can they say that's kind of the problem? Like they [crosstalk 00:16:39].
Guillaume: Usually they want to grow more and they don't know how to.
Louis: All right. So they come with this problem, the goal is, they want to grow more on, you know that LTV is a North star. They need to generate happy customer at the end. What is the first step that you take to help them? Or at least, what do you tell them to do as first step?
Guillaume: Using my process, the first thing is that before accepting a customer, I will start looking to the data. Because there's a lot of people who I will not work for, because it's a lost cause. I'm good at, I didn't think that. This is the first thing. That's okay through that, the first thing is I'll try to look at the funnel and to understand, do we have enough high quality interested prospects somewhere in the funnel? But we need to improve the conversion throughout the funnel. Or are we lacking quality of people, or we just hacking volume. It's one of those three. You can simplify the world like that. If you have enough high quality prospects in your funnel, then you need to improve your sales efficiency, maybe you need to touch the people a bit more, you educate them a bit more.
Guillaume: This is a couple of things you can go in like that. And you look at where funnel contracts where you are losing potential revenue. If you have a lot of folks in your funnel, especially at the top, but they don't convert, we probably have a quality issue. Usually I'll go into strategies around that. I'll do enrichment attractions, who do we have or not converting, how do we bring them? Basically that's usually a disalignment of marketing or branding and the product. It's like we have a blog that talks about fantastic things, which brings lots of entrepreneurs, but we actually sell a product that it is completely different. I'm like, yeah, that's not going to work. Okay. The third is that you just don't have enough volume, then we've got to wonder, have we tried stuff or have we not? Do we even have a market and audience?
Louis: Enough quality prospects in the funnel that are not converting, prospect but maybe not quality, and then no prospect.
Guillaume: Correct, yeah. The people I won't take usually is people who don't know, because they are not tracking and I'm like, yeah, that's not going to work folks. If you don't know what's happening, that's a lost cause usually. And so that's that. Once I go through sort of the first three and identify that, I try to understand... So if we look at usually what happens, people who come to me, they have some volume. Growth marketing is usually helpful only post product market fit. If we take the words, growth marketing, to grow something that exists. It's not marketing education, it's not finding product market fit. All of those are great things, It is not what I do. I usually am not helpful if you're creating something brand new that has no competition, that has no existing market and you need to create the market. That is a completely different branch of marketing, like traditional branding and marketing education, which I don't do it all.
Guillaume: I respect that, but I don't do. Those are also excluded. Usually people come to me, they have... No, some of them they already have a product, which is usually pretty good. Honestly, usually entrepreneurs they have pretty good product. The product is rarely the core problem. I mean you can see the problem is like this high train of thought, it's rarely the core problem. Usually they have a pretty good product. When I look at companies that succeed and fail, I see companies of great products that fail. I also see companies with average products that succeed. Really success is not tied that much to product quality, it can help, it's not that much. Usually what we're going to see is that, one, lack of marketing, just like they're not testing channels.
Guillaume: They don't have the habit of testing channels, that is like low volume. That could be because they just don't have the habit of experimenting, so I'm going to help them come with a framework, okay, let's test messages on channels. Let's understand who the audience should be for this product, where are those people, then what messages makes them resonate. Then I'm going to understand like, I look on a very granular, channel by channel, message by message, all those people converting. I'm trying to find what fits, what makes people tick. So I have a tip for you, I'd say nine out of 10 companies, who I get access to the data. I see something really funny. One, they pay a lot of money to acquire customers, huge acquisition budgets. Two, they really don't track, they don't keep the information of the acquisition in their CRM. Example, you buy stuff on AdWords. You're going to buy some clicks and some traffic from AdWords.
Guillaume: If you're good at AdWords, you know that you should have very granular campaigns, ad groups, and ad messages. All of that, those insights you can pass in the query string, then pass into the lead. If there's a lead form, you can store it in the lead forme and that's on the CRM, which can help salesperson understand what made this specific person tick, what made them interested in my product, what was the core message? Nine out of 10 companies just don't capture the data or throw it away and salespeople need to do a discovery call to understand, Oh, why are you coming to us? What made you interested? That is the good example of what I sort, you see. Because I work across all of the silos of the company, I see where we are losing interesting insights that we have paid for, that are helpful to help us create a better experience. That is the core thing I would do. I tried to create a better experience through the entire life cycle of the customer to make them happier and more likely to pay for my product.
Louis: Okay, so let's cover those two aspects. One is the auditing. I want to know more about how you do that. And the second part is the insight and how you make sure that you don't lose insight throughout the funnel. The first one is audit. I think that would be interesting to people listening right now, who probably have some of the tools that you're looking into. I suspect you use Google Analytics, Google AdWords, anything that you can get your hand on. How do you actually audit those channels? You said you look at the channels, then you look message by message, ad by ad, or whatever else. What is your method there?
Guillaume: I had one recently, I'm not going to name them, but I had one recently. They have, I would say, reasonably good traffic. I'm going to call that any... No. I'd say, let's call it like a 1,000 to 3,000 unique visitors a day. That's reasonably good, not huge, but reasonably good, of which half is organic, half is like paid or activity through like marketing channels. I start creating funnels and it's like, okay, what is the conversion rate of my organic channel, at all steps? How many go through my homepage, my signup and my product, eventually pay and stay for, let's call it, 90 days. Then I roll back the revenue from each visitor. I have a value per visitor in case I'm going to say, okay, well an organic visitor on average is worth, let's call it, $10, okay.
Guillaume: Based on like lifetime value. Like LTV is $10, if the math works out. Then I compare like with other channels, what's the LTV, what's my LTV of paid, maybe of social, and other things. That gives you an average. Now most people stop here, which is a mistake. What I usually do, I start to break down within each channel. Because, the average number is very misleading. You actually have a huge distribution, because if you look within organic, that average $10 is actually spread on like 3% who pay and 97% who don't pay. So making an average on such a big distribution is statistically a very big mistake. Now you need to understand who are the people who pay, what do they have in common? They signed up, which means you know who they are.
Guillaume: I do B2B. You know the company, you can can do some analysis. I usually pull data from Clearbit, I use Clearbit Reveal provider, to enrich those companies. To know more about their size, their industry, a couple of some graphics. I also work with the company called MadKudu, which I do advise, which does lead scoring models and gives me insights into like, what do they have in common? The trick I have, one of my core tricks recently is I will use an API from Clearbit, which is called Clearbit Reveal. Which helps me understand from a set of IP addresses, which companies are in that stat, so which companies are visiting my site anonymously but not converting. Score them look which ones have paid, understand within one channel, let's call it organic, let's call it a hundred people. How many are actually qualified?
Guillaume: Is it 50, is it 10? How many have signed up within those? Now if I have for example like the average I see is like 30% quality, pretty good quality. If I have like 10 people up to 30 that qualified out of a 100 from that channel, which means my actual conversion rate is not 10 out of a hundred, it's 10 of the 30. Because the others are not qualified, they're never going to pay. Now 10 out of the 30 is a pretty decent conversion rate. It's a third, 33%. Okay. If my conversion rate within that segment is actually five, now I know I have an issue. If it's two, I have a huge issue. And I try to understand the other people who have not converted, but are qualified. Why have they not converted? Is it a different message that they have clicked on?
Guillaume: Could be that message resonates for people who will never convert. Example, you have in your AdWords ads around like free and free trial, stuff like that. You have other ads which are more on the features and the product. If you treat AdWords as a global group, you're going to miss that detail of who are the people who are coming, even if they don't sign up by default. Maybe you're destroying value by pushing qualified people through a free message, then they see your pricing page, they have, I'd say, disillusionment. You're destroying value. You should actually remove those free messages, those free ads, because anyways they're not bringing people who are going to pay. I try to be very specific, I try to understand like... Then I do an average on the small percentage of people who are qualified, and I should be able to convert. I do a lot of data enrichment. Clearbit is one example, I also use Datanyze. for example, I was at Drift in the past. Drift is a live chat tool.
Guillaume: Look at the traffic. Who are we selling to? We're selling to businesses that have salespeople. Okay, that's good, I'd say, smaller sample. Now if people are using our competitor Intercom, that's even better. I know they're qualified. They're already for the tool. If somebody who uses Intercom comes to Drift, that means they're unhappy about that tool. If they don't convert it means we are unable to convince them that we have a better solution, which means that marketing needs to change. Here, this is positive marketing. I know enough to know that I'm doing a bad job for that specific segment. That's my other face. I pull the examples, I show to my salespeople and to my CEO. I show examples saying, hey, like those 20, 50, 200 accounts came, viewed five pages. I have six different cookies coming, they have Intercom, they have this thing, have the right size. They're in the US, they did not sign up. We have an issue, we need to change the message. That is clear as water, because I can tell them how much value we have lost.
Louis: Let me try to repeat it in my own words, please correct me if I'm explaining it wrong. Traditionally marketers have a very tough job to try to understand the traffic coming to their site. It's traffic is anonymized, you don't know who they are until they sign up. What you do first is you look at their behavior, you look at signup customers. You try to identify what are the common attributes that make them convert, whether they're for Drift, they're salespeople, they used to use Intercom. Maybe that's the best attribute. Once you do there is then you try to do the same for traffic. That is, I haven't converted yet, try to guess kind of what is the quality of this traffic based on information that can be online from them.
Guillaume: Because I'm in B2B, and that's specific to B2B people. I can go from the IP address to a company name.
Louis: Yeah, sometimes. It's not perfect, I mean that's [crosstalk 00:29:55].
Guillaume: Totally, it's not meant to say, here's all the people, it's like 30% coverage. But for statistical analysis it's more than I need.
Louis: Yeah, and that's important to say. So we don't expect those type of tools, as you said Datanyze or Clearbit arrangement to do actually tell you every single visitor. To tell them who they are, their name, and whatever. That's not going to happen.
Guillaume: That's not going to happen. The point only like statistical analysis, because the distribution of coverage is the same from one channel to the other, more or less. If you have much lower quality on one channel than the other, then he tells you something. One quick example, when I was at Segment, Quora was being fantastic traffic, great conversions, great quality, fantastic signups. We said we want more of that. At the same time Quora launched Quora ads and so I thought, hey, this is a great channel to experiment on, let's buy more of that traffic. By doing this very quick iteration on traffic quality, I was the most in 48 hours that the paid traffic from Quora was of a much lower quality than the organic traffic. And so we stopped.
Louis: Within what, 48 hours you said?
Guillaume: 48 hours, yeah. The benefit here is that if you look at Segment's sales cycle, so trials, sales whatnot, is months. So usually marketers need to pay for traffic or for that channel four months before they realize that there's no LTV. In my case, because I run predictions real time based on past conventional, past data, I'm able to know as soon as I have statistical significance within couple of days, whether it's worth continuing or not. Which in this case it wasn't.
Louis: You basically get an understanding of the credits of the traffic, you said like if it's what? If the quality is 30% is usually quite high. If 30% of the traffic matches the type of people you want to try, it's quite high. You said if it's below five, you said it's really bad, right?
Guillaume: Yeah, for sure. 30% of people in your ISBU add your customer profile, it's a very high percentage. Very high. But like 5% is also very be low. You can do the math yourself on a napkin, you just calculate your cut to LTV, you see that going from a third of the traffic that you pay for... Imagine that your cost per click on AdWords is going to be in the couple of dollars, and you have one third of those that are qualified. You know it means that your qualified visitors can be, is cut like 10 bucks, three times three, 10 bucks right there. You have a 10% promotion to a proper sign up, which means now your sign up cost is $100. If you have 10% of those who pay, that means your pay is like $1,000. Can you recover $1,000 with a SAS product? Yeah, usually in a couple of months you can do that. Now if it's 5%, now it's five times more expensive. Your pay is like now 5,000, 6,000 bucks. That's hard to recover. All right. That's the way I think about it.
Louis: That's your audit, basically the end results of audit is a funnel with conversion rate after conversion rate. From visits to first touch, second touch, et cetera, up to revenue. As you mentioned at the start. Up until the business metric that actually matters, not the fluffy, shitty message that you can add in the middle of it. At the end of this, once you have disinformation, you also said, a few minutes ago, you said, nine out of 10 companies don't carry across [inaudible 00:33:44]. They kind of miss all of that. How do you advise folks to do it better, without necessarily spending thousands and thousands per month on the tool? But how do you ask people to do it so that they actually get data and not waste money on and losing it in the [inaudible 00:34:02].
Guillaume: I mean, so once the opposite is done, usually what I see is, on a channel by channel basis, which channels have high quality, which channels convert. I see opportunities, it's there, it's clear as water. You're going to see channels with good traffic, good quality of traffic, low conversions. I didn't say, well let's start, let's drive that traffic to a specific landing page, and we're going to try iterate multiple messages. It's the first thing. All right. Let's try to engage those people differently with different channels. Maybe we should have a chat on that page, maybe they need more help, depending on the channel and the type of company you're bringing in. Of the channels, you're going to cut them, reduce costs, because the quality is not there. Channels, you say, well, we have good traffic, good quality, which need more volume, let's invest more there. Those are all very obvious outcomes to be honest.
Louis: Let's talk about the scenarios. They may be obvious for you, but I think it's good to repeat them and to see the type of scenarios. So you have the first one being good traffic, low conversions. That means, okay, we're really good people, let's optimize the page on the funnel. Then you could have shitty traffic, therefore shitty conversion. Let's scrap that.
Guillaume: Yeah. Or let's change completely how we're putting the channel. It's the wrong message, it's the wrong audience, wrong targeting.
Louis: Let's see, completely start from zero again.
Guillaume: Yeah, it's great.
Louis: Then you might have good traffic, good conversions, but not enough volume. It's like it's converting two leads a day. It's not enough. Let's find ways to improve the channel. Are those the three scenarios or do you have more?
Guillaume: Yeah, it's. Basically those are the three main scenarios that we have.
Louis: What are the most common, what are the ones that you see the most out of these three to started?
Guillaume: Sure. Good channel, low conversions, low volume, sorry. So low volume is a typical problem, that happens and that leads to... For entrepreneurs who are building their business plan, it's a typical mistake is that they starts doing some stuff from social, like LinkedIn, stuff like that and they have a couple of clicks a day, visits a day that convert pretty well. Then they built a whole business around that, that's a misunderstanding. Because the channel, the market will react... The early adopters will find you, will react positive to your message because that's who they are. They're early adopters. It does not mean that there's an infinite supply of those people. Usually, as you scale costs go up, and conversion goes down, it gets harder and harder. Just think there's a tree and what we are eating is just low hanging fruits.
Guillaume: You pick those fruits, then you need to climb the street. So people forget that the after the law in which you needed to climb the tree. Usually there is just not more with those people and you just need to accept that. You've mined the easy conversions of that channel, now you just see if the rest make sense. Sometimes you can say, okay, I'm going to completely exploit that channel. Example, At Drift, we found in the early days that posting on product launch was good. Really, it was free and brought really good traffic that converted well. The problem is you post once and it's done. So how do you maximize that? What we did is, we said, well, we're going to post every month a new feature called the product. We're going to have a nice video.
Guillaume: We are going to have influencers repost, up-vote it and whatnot, we're going to win. We are going to be the number one every time we post. So we built a machine that we knew how to crank, and we launch every month. We posted for 12 months straight. If you look on Product Hunt for Drift, you're going to see like 12 products, month after month. We mined that channel weekly, everyone on that channel was going to learn about Drift. Then we moved on to the next channel. I think there's a stronger benefit to maximizing your presence on a channel by being all in, then moving on. Rather than being very disperse, and having very limited returns. There is a take all strategy.
Louis: That's interesting, because I do remember seeing Drift every single fucking month. I've been like, Jesus Christ yet another product. But I mean, yeah, people remember that. As you said, yes. Let's take this scenario. You're a new entrepreneur, you've noticed that the first product, and the product can't work really well. You write your business plan on it saying, hey, the next five years our growth is going to be 100x, we're going to do product hunt every month, the next five years. It's not going to work, right?
Guillaume: It's not going to work.
Louis: Because one, you're on a plateau, you're going to reach some people. And two, as you said, diminishing returns. It's not because you have a 10% conversion rate today on the channel that if you multiply the volume by 10, that you're going to have 10% after [crosstalk 00:39:13].
Guillaume: The reason I use mine, you got to think a channel like a mine. I explained this recently in a keynote, I said costs on Facebook are going up. CPCs are going up on average. The reason for that is that inventory is stable, but there are more advertisers. Think about this, Facebook is doing a lot of interesting and shady stuff. One thing they're not doing, creating more Americans. So the number of people you can advertise to on Facebook is stable. That's what it is. If your market [crosstalk 00:39:52].
Louis: That's why they're launching their dating product, I think.
Guillaume: Sure. Long term strategy, longterm, very longterm. For the moment as a marketer, if you're marketing to some professional, or some type of role in businesses on Facebook, that is a finite market. It is as finite as if you were mining gold in a mine. So you're trying to do it the most efficiently possible, you're trying to understand that this has a limit. You're going to get to the end of it, which is different for B2C. B2C the volume is big enough you [inaudible 00:40:25] almost close to infants. In B2B, if we look at Drift, that is a good example. Companies that have like over 10 sales people in the US, there's a finite number of that. Now if you look they have a website, they sell through internet, We talk over a couple 100,000 in the US. And that's it.
Guillaume: Okay. So if we go back to the quality of marketing, how we invest for the longterm, I always think, okay, I could burn that market to the ground in two days if I wanted to spam everyone. It's easy, I could definitely spam all those companies and I would do a great quarter. Then it'd be fired. Now the question is how do I maximize my revenue over the smallest number of people, so I can be the most efficient possible. I need to find, I need to know who is ready to buy now, who's the right person? What is the right message? Those are the things I'm trying to figure it out and which channel should I use? I don't want to spam them. I want you to convert with as little effort as possible. It's margin maximization, CAC to LTV.
Guillaume: So I tried and that's why I'm buying a lot of data. We talked about Datanyze, about Clearbit, I do a lot of other things to understand who's in the buying cycle, which tool are they using? Am I a competitive situation? Are they hiring salespeople? What's happening that I can know, I can leverage to help my sales process, to help my conversion. If they don't match that, it's not the right time. I know that I'm going to need to push too strong, and we talked about claw back. It means I will face claw back. It says, if I'm pushing an expensive live chat to people who don't need it now, they're going to shun.
Louis: I mean, that's a great lesson. There's something we talked about before in the podcast about sales actually, the sales principle of, don't bother knocking on doors of people who don't have a bleeding neck problem. They have a bleeding neck. Don't bother talking to people who don't have money. Don't bother talking to people who don't have the permission to say yes. Those five qualifiers from sales, that's basically what you described. You identify through your way of your data analysis, the people who are the most likely to convert. Because they're actually looking for a tool, you don't need to convince them or anything. They're fucking looking, you just need to be there for them.
Guillaume: Don't lie to yourself. I think an issue for marketers is we tend to lie to ourselves, because we need those numbers. I think white papers and webinars is a good example of how we lie to us to ourselves. We need to give leads to salespeople. So we produce a white paper, we do a webinar on a topic that is really just tangential to our core product. Then we have like a thousand people who download the white paper, we say to ourselves, hey, here's a 1,000 leads. Honestly are there any leads, and we have... There's one intern from IBM who downloaded the white paper, is IBM a lead? I see that all the time, I see those big brands [Competing against big brands on a low budget? Read this.]. I once had like a thing, there was one guy from Delta Airlines who downloaded something. When I was at Segment, I'm like, yeah guys there's 10,000 people in the IT department at Delta. This means nothing. This means absolutely nothing. Now if we have 20 people from Delta that read something, then yeah, then I have a signal.
Guillaume: But it's a bell-curve logic. One person from a black logo means nothing.
Louis: You mentioned data analysis to do this. Let's talk about how you managed to build the typical profile of someone ready to buy, because your tools that you mentioned like Datanyze, Clearbit. They are great at giving you firmographics, demographics, roles and [inaudible 00:44:18] shit like that. But do you go beyond that, look at psychographics? Do you look at behavioral stuff, like triggers? Just let me give you an example quickly. It's fairly clear that people buy Pampers are planning to, are going to buy Pampers the next few weeks, are going to have a baby. They have the baby, then that's going to happen. It's a life event. If you only look at the person professionally speaking or whatever, you never going to fucking know she is pregnant. That's one example. In B2B there are other scenarios, that's the only one I could come up with right now on the fly. So long question, but do you have ways to have any find stuff beyond traditional demographics' information?
Guillaume: Yes. I've built a specialty acquiring, I'd say what we call, a signal data, a business signal data. Before I get into that, I want to say something. I have found no benefits in looking to personal stuff. The life stuff, like you said, knowing what they do on their own personal life, when I'm selling the business software. One, it's creepy. Two, it really just doesn't help conversion. So stay down. And third it's legally complex, whereas business stuff is very legal and okay to capture. I'll go through a couple of examples. We talked about Datanyze, which technographics, what are companies using as technologies. Yeah, that's a good one. I've used stuff from PredictLeads, which is a company that tells me what roles are companies hiring? Are they currently hiring for salespeople, CS people, VPs of sales, VPs of product, whatever.
Guillaume: This is public data that's on their job board. So the company scrapes all the job boards of all companies in the US, and structures that. So it's legal, it's out there. It's just structured in a way you can you use it, because for me, a company hiring a bunch of salespeople is a good signal if I'm selling a cell store. That's one example. I buy the data from G2, I was a customer initially. Because they have instant data telling you which companies are currently in a buying process, looking at multiple tools in your category. Also great signal. I look at [Bamboo 00:00:46:40], which tells me what companies are reading on the web. I have like tons and tons of singles like that to try and understand what's happening in the life of that company. I use it both to understand is it the right time, also to craft the best message.
Louis: Let's say you used the hiring data. So you know people in the industry that you want to target are hiring a lot of salespeople. Now, you know the business name, you're able to extract that as [crosstalk 00:47:11].
Louis: What do you do with this information then, what do you typically do with it when you have that information?
Guillaume: Typically, that will go up to, in my case, which is I'd say a bit of advance, it authentically flows into enrichment, the contact discovery, then goes into outbound emails and advertising. In a more simple use case, you could just add it to your CRM and give it to a salesperson to reach out to. In my case because of the volume, it's completely automated. If you want to go there, I built a system which is kind of a bid management system. Where we, based on the business behaviors, we create a virtual bid in dollars of how valuable is the company, how likely they are to close, multiply by the future predicted LTV. We know how much of the LTB we're allowed to spend on the CAC, which means we have a bid and based on the bid we have augmented outcomes in terms of marketing. Especially going up to like giving them a gift, sending them a gift through Sendoso. Because, it is a profitable bid.
Louis: Simple, right. Simple stuff.
Guillaume: I have the examples I like to use to explain how that works. Two years ago if you went to Segment's website anonymously from a large enough company, it would have a live chat bot that popped up. Just if you were in that category, the live chat bot instead of saying how can we help you, which is like not very differentiated. The live chat bot would say, how did you like your coffee? That's a weird thing to ask for a live chat bot. Because people are like, what? What Do you mean? What is this thing? I'm on a business websites, so they would click interact. I already win them, they are now interacting with my marketing. The bot would ask, do you like like tea or coffee? Do you want milk? Do you want sugar? What it's doing, it's building an order.
Guillaume: I know the IP address, which means I know the company, which means they know the HQ. I only need to ask for the first name, to put like Louis on the cup. What was happening is that we were building the order, sending it through Sendoso, through Postmates, and delivering the desired hot drink in most us cities within 20 minutes. It's about 15 to $17 for us, which for the right company is an amazing CAC, because it's a memorable experience. I'm delivering something of value. If we go into psychology, I'm creating reciprocity. People are now on my chat, on my site, what happens is the salesperson jumps in and says, hey, while you're waiting for the coffee, do you want a demo? It's hard to say no. Now I have a $17 demo with a large company. This is an amazing cost. The example of like bidding on the right experience.
Louis: How did you come up with this idea? I'm not necessarily interested in this idea in particular, but when you know the opportunities, as we mentioned the three scenarios. When you know there's an opportunity, how would you like to experiment, I think come up with ideas like that? Where does this come from?
Guillaume: As a marketer, I try to be very close to salespeople. I just want to see how of the best salespeople sale. The reason why is that my job is to automate the best human actions. I'm not doing much more. Look at the best salespeople, the very best ones, they learn a lot about their future customer. They will ask lots of great questions. They're going to invite the people to great experience, they will send them gifts, they will send them really well crafted emails. They're going to be helpful. They could do all those things, I really wonder, can I do it without interacting manually? Can I do this? Can I do that? So for example, like me buying third party data is the same as a salesperson going to a call and asking the questions. Me sending gifts, is the same as salespeople sending gifts. My emails, outbound emails, cold emails have value in them.
Guillaume: They are helpful, because I have realized that being helpful is one of the best ones. I have another example for you. A helpful outbound. I worked for this company called Gorgias, which is a... It helped us for Shopify stores, people who sell on Shopify eCommerce and stuff. We realized that those brands, digital native brands, they'll focus on Instagram. So they've posted any product on Instagram, if they have a good following, people react and will post messages. Most of them positive, some of them negative. As a brand manager, as a CEO of such a brand, you shouldn't react to those negative messages, and try to calm things down a bit, if it ever happens. And some people forget. What we did, we looked at all the events in the US, we script the Instagram profiles, we then script the posts regularly.
Guillaume: We look at the post with high engagement. We didn't look at the responses to those posts, we script the content. We use MonkeyLearn from natural language processing, NLP, and sentiment analysis. We look for negative messages. If there's negative message, we didn't look back to see, is there response from the company. If within 48 hours there's no response from the company we screenshots the interaction, inject it into the email to the founder and say, hey, you should take care of this. That's it. We do not talk of a product? We do not talk about ourselves? We just say, there's something, you should take care of it, which is valuable. Really, it is. You should do something and creates trust over time. We do that one, two times, three times, eventually the guy thinks, these guys are pretty good. I should look into what they're doing. We have built trust now. I have created positive good marketing here.
Louis: It sounds like the way you come up with ideas really is about scaling what humans do in a sense. You're scaling what salespeople do, scaling what you would normally do. Let's say you would talk to this founder next to you, go through an Instagram and say, you didn't answer to this guy. You basically scale that using technology, so that it feels like it's good marketing, right?
Guillaume: Yeah. I'm just displacing humans, by doing what they're doing, 100%. Yeah, don't forget my wife is a salesperson, I learn a lot from what really good sales people do.
Louis: That's an unfair advantage on you.
Guillaume: I guess, yeah. You take what you have.
Louis: Guillaume thanks so much for going through this. A lot of interesting stuff, I think people can take a lot from what you said in terms of the method you use, and the thinking behind it. Not the ideas themselves, because every company is different. But the methods, the way you're thinking about it is extremely interesting. There is one thing I've been dying to ask you actually, I'm looking at your CV, on your experience. Maybe other people have asked you, but I'm curious. You stayed less than two years at Drift, less than two years of Segment, less than two years at Mention, right?
Louis: Why is that?
Guillaume: I am very efficient at a specific scale. I'm really good at taking companies that have almost no Martech stack, almost nothing in place. And creating the right infrastructure, the team, the processes to automate all of that. Eventually companies grow, they become rocketship. If you look at Drift, joined when they were 50. There was like marketing wise, very little from a marketing automation and growth standpoint. I left, it was a bit beyond 300. At 300 my job as VP of growth, in charge of a team of like 10, 12 people is really a couple of things. Hiring, firing, one-on-ones, exec meetings, none of which are actually coming up or producing experiments. I have enough ego to believe I'm a pretty decent growth marketer. I'm definitely not a 10x like mage. That's what the job is eventually. So I then rinse and repeat. I say, okay, that's good. Now I can go somewhere else and start again. I know if I can do that a couple of times, it's a pretty good story.
Louis: Got you. This is why you switched to adviser, which is a bit easier to do-
Louis: ... that what's your best, Right?
Guillaume: It's easier to get out. It's easier to get out indeed and help them. They come to me, because they say, hey, we just raised a CBC with ACBSP and we're like 30, 40 people. We really need to... Now we have the budgets, we need to set up all of those things, and I say, okay, I know how to do that.
Louis: What do you think marketers should learn today that will help them the next 10 years, 20 years, 50 years?
Guillaume: I think sales jobs and marketing jobs are slowly getting displaced by automation. If you think the example of people who could do paid marketing, that job is disappearing. It's being replaced by ML engineers. Which makes kind of sense like changing your bids based on yesterday's data, really we are replaced by algorithms. The same is true for a lot of other things in marketing, which is deciding who should get what piece of content. Couple of years from now that is going to get completely automated, which means we have a couple of things that will resist. One is coming up with the copy and the creative, that's going to resist longer. It will eventually get automated, but much further downwards. So coming up very creative, a very good brand and great copy person, that has a lot of value.
Guillaume: The second is somebody who builds the systems, the automation. So being a technical market, which is of course like the path I have taken. Being able to like stack the bricks or connect the pipes, if you want, between existing tools, being extremely good at knowing what's the next wave. One thing I want to say here is, if you work for a company and you're using Marketo, Marketo is a good tool. Sure. Fantastic. But how do you differentiate? Do you think you can have like 10x better copy than anyone else in your market? Honestly difficult. You have the same technology, which means we're going to send the same kind of emails with the same tool, you don't have a competitive edge. My strategy, if it's not copy, my strategy is to use the bleeding edge of tools.
Guillaume: Be always on lookout for, how can I get a competitive edge through technology? What's the new and upcoming tool to send emails, how is it better? Same for doing, doing social engagement or doing anything else, scraping, and anything else. I always am on the lookout for that. Always. I tried to use things that people are not using. So marketing is extremely competitive, especially if you go to Silicon Valley with the vast amounts of money that are here. It is extremely competitive. If you're looking at what I'm doing, we talk about crazy stuff. I do stuff like, yeah, they have coffees on websites. But I also do things like content personalization and email personalization at down to the user level. Yeah, it's expensive. It's complicated. But I have to do it, because I want to create a better experience than my competitors.
Louis: It's about better user experience as you mentioned. And LTV, happy customers. Let's not forget that. Thanks for this answer. What are the top three resources you'd recommend listeners today? It could be anything from podcasts, to books, to tools, to whatever.
Guillaume: The more expenture you become, hopefully the better network we have. I think a good chunk of the value I get comes from being in closed communities. To be honest, the best advice, tips and tricks isn't publicly shared. I am in many Slack groups and WhatsApp groups of small communities that will share privately the learning, especially the failures. It's essential to know what hasn't worked for others. Essential. That's one thing. Second, I read many books on psychology. If you haven't read Robert Cialdini's book on influence, read it. And if you have read it, go read his latest book on Pre-Suasion, which about what I'm doing. Psychology books, a lot of that of course. The third thing I'd say is, try to understand what are others doing, spend time looking to competitors, try to understand what are they doing, what is their edge or can I find something? I spent a lot of time doing that. Note, marketing is competitive, you don't live in the bubble. Most of the people compete with someone else.
Louis: Yeah, I concur with Cialdini's book, I mean it's been mentioned so many times on the podcast. It's actually the most mentioned book on the podcast, most mentioned resource. But every time I force myself to re-read it every six, 12 months. Every time you're like fuck I forgot this, I forgot this actual thing and let me try something on this. Humans are really good at other things, but we are so bad at empathy and understanding others. Because it's so counter intuitive. So you need to force feed yourself in term of the biases that we have, then you can use that in your marketing.
Guillaume: I could also recommend Freakonomics, really good book on understanding human behavior. There's a couple of those also. There's the one from Dan Ariely also. There's couple of good books there, that I would recommend to understand, how are humans irrational? I think the book from Dan Ariely is Predictably Irrational.
Guillaume: Which is really good book also. I use a lot of that in my day to day.
Louis: I read The Choice Factory recently, I don't know if you read it-
Guillaume: I have not.
Louis: .... It's really good. It's also psychology studies and stuff. It's 50 studies, it's really well written, and it's quite simple to digest.
Guillaume: A quick example, I read a bit of studies from, HBS, Harvard Business Review. One of those I read a couple of years ago was how hot drinks increased likelihood of accepting a business offer, versus cold drinks. Which is the basis for my hot coffee experiments.
Louis: There you go.
Guillaume: I read that, and I said, damn, this is interesting. It's Recommended if you're buying a house or a car, always take cold water.
Louis: Guillaume you have been a pleasure to talk to. Thanks for sharing some of your secrets, for the real secrets. Probably I'm going to have to join your WhatsApp group, but where can listeners connect with you, learn from you and ask you questions if they need?
Guillaume: LinkedIn is a good source of choice. There's a bunch of videos of me on YouTube. Also, if you want to check out the different keynotes I've done, they're all on YouTube. There's like probably 30 or 40 of those.
Louis: Nice. Once again, thank you so much.
Guillaume: Thanks Louis. Thanks everyone.