Monetizing SaaS: Episode 37 with Gary Bailey, Monetization & Pricing Expert

All right, Gary, it's so fun to have you on the podcast. I've recently met you, but I've been reading your work and your content for many months now, and it's always instructive. I always feel like I learn something. You're an incredible storyteller. But you know, you're a founder, you're an operator, you have a very deep background in banking and financial services products and monetization. Right now you're the founder of The Talent Capitalist, where you're teaching monetization as a discipline, which I want to dig into. Something that's so unique about you is the depth of your experience across product, financial products, AI, and pricing. Can you start with just how did we get to now? And then I want to go into like, what's the thread that unites all these interests?

Well, thank you. It's good to meet you again. Uh, well, I'm an accountant. I know people don't like admitting to that, so I could say I'm a recovering accountant, but I'm an accountant at heart. And accountants like certainty, no matter who you are, accountants like certainty. And so I went into audit and worked for Deloitte for 4 years immediately after uni. And I was lucky to be fast into financial services, just at the kind of like big bang, we call it in the UK. I think you call it Glass-Steagall or something where banking really went crazy stuff. And that was the early '90s, I think late '80s in the US, but early '90s in the UK. And that time, uh, almost like fintech in 2008 to '15, where an explosion of products just came to the fore, and all those products need to be valued. And so there wasn't a function called valuation or pricing, but that's what they needed. So they put whichever accountants who are slightly more interested in math, not like deep calculus, but just decent at math, they would put them into, into this department called Valuation and Product Control. And the product in this question was traded derivatives, which we know what, what happened in 2008 and many other times, 1998, 2001, 2004. It's not like it's a once a one-shot deal. And so the idea there is you had to value a myriad of products in which each time the product was created, it had, we might call it today, a feature flag or an entitlement. And each one was different. And depending on what bank you're at, whether UBS, Warlen, and Nomura, there were thousands of trades, hundreds of thousands of trades. And so it's pretty much the same as now. Imagine If a company had a salesman or woman, and every single time they went out and sold, it had a different feature, different clause, which actually happened. But imagine that's not what's supposed to happen. But you have to have a way of valuing each of those clauses, because otherwise, you're left with a mess, as we were in GFC. And so, the idea is that you would have this, you might call it feature flag or entitlement, which might be 100 fields. And each of those fields had a value, whether it be 3360, which is the day count, or who was setting it, the Australian bank, the UK bank, when did it roll over, all sorts of really, really immaterial things that would have seemed immaterial in the kind of paper and pen era of early '80s, but were actually important because the trader could get $10, $50, $100, $200 out of one of these small clauses. That he or she may change without obviously someone knowing. And so we had to develop a system which had hundreds of these features, like insurance, you could say, and each of those features had a value. And if the trader was to add or basically take one of those off, it would create a different valuation. And the idea is that we wanted to create a consistent valuation curve, but you wanted to make sure that some features were kind of, what's the word, discouraged, with evaluation penalty. And that was the genesis of saying, actually, okay, if we have to value 55,000 trades, the last book, the man, then you need a kind of, not an algo, but a tool that is able to bring in all the feature flags that you can review, and also to overlay daily pricing, we have to value it daily, sometimes more than once a day, and also have a reporting system that could explain it to someone else in simple terms, even though people in banking, there's not many people that really want to go down derivatives and option rabbit holes and board meetings. So you have to explain it like they're intelligent but also not that interested in like booleans.

So kind of this world of finance in the late '90s, early 2000s is not a world I'm very familiar with, but just to play back what I think I'm hearing you say is effectively you came from accounting, you valued control, you valued an appreciation for the value of a product. And at that time in finance, as it was beginning to electrify, um, you had effectively traders selling an enormous variety of products or SKUs without any type of real sort of compliance or operating controls. And so effectively, that's like, that's a real compliance risk, um, in the context of like you know, what are we actually selling? Is it valuable? Do customers actually get valuable value from this? And you were tasked with bringing controls into that world at the bank that you were working at at the time.

Yeah, that's perfect. Perfect, basically. And the idea is each of the traders thought they were selling a product. They would be called option product, derivative products. So they, long before this project manager and, and like product manager, they were seeing need that individual products that they were selling. And hence, we were just valuing those products, but we were valuing a massive database of products. So, that is very similar to what you would have now. Let's pick a company, HubSpot. I don't know how many clients they have, but say they have 50,000, probably more than that. Imagine if every single one of those clients could have a different feature. They kind of have really good segments, it's beautifully laid out. But imagine if there were no segments and every client was picking from the 100. And that's where I think we're going. And so through line from accountancy then worked for mostly banks in product control and then was fortunate or unfortunate to basically we had a problem with compliance, as you mentioned earlier. And the problem with compliance is that every time the trader wants to onboard as many clients as possible, One, bigger Rolodex or the equivalent, and two, you never know, they might buy something at some point. And so in the kind of post-GFC period, we had the regulators realizing that there were just too many clients doing too many random things, and so they, they clamped down. Uh, some of the Dodd-Frank, some of it was some of the UK versions of Dodd-Frank, some of it was MiFID, but all these, uh, kind of rules converged around Do you know your client? And you see it. I probably listened to a few of your podcasts where everybody talks about knowing your customer and knowing your client. What do they value? That sort of stuff. It's the same in banks. What on earth is Fynn Glover gonna buy from us and can he afford it? Really? So, sending you a million-dollar loan or whatever, or giving you leverage. As we've seen even recently, leverage goes pear-shaped. And so, Most companies had hundreds of thousands of, of Fynns and Garys, most of which they had never had any touchpoint with, never met after the first introduction. And so we had to cram that down to a manageable number. And so, approach that in a similar way, which is every single client is a feature, has features. And therefore, what features do we value and what do we want to price those features at? For you, it might be founders, how much money have you got, what family you're from, you live, etc. So that also had to be bought from a paper kind of insurance model, like 1800s insurance model. So let's try and actually value every single client. And they were like, you're crazy. No, no, no, every single client had a value potential, we call it present value in financial terms. And therefore, by putting all of their features, Fynn Glover's this tall, this height, this is where he's from, this is where he lives, this is his company, what he's done before, founded before, you can get value.

Yeah. I mean, it's so fascinating because like just the operational challenges to do what you're describing at that time with that type of scale and volume, I imagine were immense. I mean, you're dealing with thousands and thousands of customers, right? And you're trying to support product catalogs with huge SKU diversity. You're trying to support kind of profiles on customers, what they're using, what they're getting value from. I mean, it's just a big operational lift, I imagine, especially at the time when this was moving from pen and paper.

Yeah, you said, but actually Excel saved the world. And so Excel, essentially, this is before the million rows day, but you could at least have, I think it was 64,000 rows. You could download a bunch of clients for like a really big SaaS company now, you could download it all into an Excel database and then use some add-ins, a bit of VB, a bit of an Access database and bodge around. And that was the genesis of the model. Then someone would say, "Oh, that's really rubbish. We're going to do it in C# or whatever, and we're going to move it here." But they were always moving it from, "This actually works in Excel really, really well, but it's just really slow and sluggish and falls over." And they would take that almost like a model and just say, "Well, okay, I'm going to make it faster, better." Well, in this Excel file, like what I'm picturing is sort of like a modern customer file that you might see a SaaS company running their business off of where it's like name of customer, what plan they're on, what features they're entitled to, what their ARR is.

And if you're really sophisticated, you're moving into like usage as well and you're trying to aggregate finance data and product usage data. So is that kind of a good picture of what you were building?

It's not in reality, I suppose. In most clients, even in big SaaSes, in 200 columns, you can get their information. That's not efficient, but actually it's doable. And that's a good model that takes you to, actually, now I need to industrialize this. Do I buy it in? Do I do it myself? But it shows that it works.

So take us from the early 2000s, '08 to 2024, and, you know, what brought you to now with the Talent Capitalist?

So yeah, so During that time, we recruited hundreds of people. Basically, finance turned over people like nobody's business. So, kind of average tenure of like 3 years. So, you turn over a third a year. So, we were recruiting like crazy. And each time we recruited, you know, the problems of recruiting, basically, it's a bore, like, you have to interview people, this and that. We thought, how do we shortcut this process? And one of the issues we found is because UBS, it was a bit easier because they were like top tier, tier 1, they call them bank, whatever, top 10. But as we moved into tier 2, we worked at Nomura, and then I worked at a Chinese and basically South African bank. People would rather go and work at JPs or anybody with Morgan and men, JP Morgan, Morgan Stanley, And so, in terms of hiring someone really qualified, it was very difficult, not just the price, but also, they're like, "I don't care about the New York money, it's a second-tier bank. It's a good bank, makes all the money, but it's not the same as working at Goldman or JP Morgan." And so, then we reinvented the wheel and said, "Okay, what if we only take graduates?" And then the initial response, just like when we get to monetization, would be, "They know nothing." They're literally out of short pants. And so that's the truth. They literally have been to university doing what we would do at university, which is normally drinking in the UK and partying. So basically, they're only qualified to like not get drunk or no, they're sitting there, okay, how do you train these people up rapidly to run a book of 50,000 kind of clients and features, etc.? And to be able to be conversant in about probably less than 2 months. And so we recruit graduates who are numerous, not like PhDs, but numerous graduates in on batch, like 10 or 12, 50 at the same time. And then we would have people mentor them and train them up. But because of that, we had to make the actual valuation, press a button, ruthlessly simple. Almost like imagine if you and I worked at HubSpot and someone was saying, someone said to us, whatever, show me the above the line and below the line. Show me clients that are like hitting way above what we expect for this ICP and delivering, couldn't call it, revenue retention, etc., and revenue ARR, and ones that are way below what we expect. If that means someone would therefore have to set a benchmark for what ARR we do expect and which one we expect to expand, etc., and contract. And so, by jamming them in, in a short period of time, recruiting 10, 15, and similarly, basically saying, okay, we need to teach them very, very quickly, we had to make the process very simple. So that got me thinking, are there other disciplines where you have to teach someone something in such a rapid speed that you haven't got the time or the money or the prestige to recruit the best people? And if we bring it back to today, so imagine if you, Fynn, needed— someone gave you $100 million funding and said, recruit the best price in a month's duration people, from engineers, designers of your billing department, to the actual people who can do it, CSM, etc. You would spend a bunch of money doing that. And obviously, each time you'd attract the top person, they'd want lots of money, and you go down, and it just gets worse and worse. And by the 100th person, they're going to be good, but clearly not as good as the first person, if you can get the first person. But actually, some of the skills can be taught, because they've been at university for 4 years, they've been at school for nigh on 14 years plus. So the idea is to make it interesting, how can you codify it in such a way that makes it easy? To learn. And so in the last few jobs in finance, the KYC compliance job, and then the previous 25 years in finance, saying, okay, we need to— well, whether we like it or not, we are going to churn a third of these people because that's just natural attrition. Therefore, the people that come in need to have less than a kind of 3-month onboarding. And so therefore, we need to make the process very simple to explain. Componentize the process that you can obviously do like Model T factory job and also make it so they actually want to stay in that function, as in elevate them in a way that says, this is a discipline, this discipline is worth something. You have to value the discipline and work at a discipline and therefore, someone will value it and they have to also believe that. Because they were graduates, there had to be very little other than, are you providing food, water, drinks, and money? But as they got older, The ones that stayed did believe this is the discipline and this is going to become valuable.

And so just to make sure I'm fully understanding, effectively, you're running into an industry where there's enormous attrition. There's an enormous pressure on training young, young in career people, and these are traders, right? How to price and value the products that they're selling and trading. And the way you delivered that curriculum was sort of key to success, given the amount of attrition. And so what that sort of taught you is, how do I teach monetization as a discipline in an environment that is A, at scale, and B, requires rapid information transfer?

Yes, perfect. So probably I wasn't actually trading traders. Per se, other than the ones that were doing something crazy, basically, but mostly the finance people controlling those traders. As in, you can't do that, you're not going to do that, that's, we've just lost us money. And so really, babysitting traders, as it were, but every day they bring back trades and we were like, okay, reviewing trades and put them into portfolio. And so in reality, we had to make sexy. And if you take finance as an example, So in the kind of early noughties, what you're seeing, there's got quite a few podcasts. Most of the finance podcasts tend to be FP&A, finance planning and analysis podcasts. And that's because in the early noughties, there was this expansion of we need numbers and we need to do budgets and revenues and the GFC and LTCM and all these crashes. And every time we had a crash, the regulator would ask us for more, more budgets, more, more paper. And so the average accountant, if you take out the product controllers, me, who are just deep in the spreadsheets doing maths and arguing with traders, then there was another part of finance who the business said, okay, look, I need like pretty numbers on PowerPoint every two days. And the average finance person, well, I'm not doing that. I'll give you numbers once a month. That's not my job. And so you had this fracture of the sexy accountants who had the mathematical guys like me and Fynn Glover beavering away and arguing.

Quite aggro.

Then you had the rest of the finance, and then that fractured again between the "no," the computer said no, to be honest, once a month you get a number, to the guys and girls who are the sexy part of, "Yeah, I can get you a number every couple of days, and I'm going to make it look pretty on a dashboard, Power BI, whatever you need, I'm your man or woman." And they were FP&A. And so if you look back to the early noughties, there were no people in FP&A because there was no FP&A. Someone literally made up the title. And I first heard it in 2007. And it's the biggest growing part of finance and the number one, I think, yeah, I think about 40%, the number one kind of route to be CFO. So something happened in that period to make FP&A one, a route to be CFO, which it never was in the '90s. You had to be like a hardcore auditor. Uh, but now FP&A. And very similarly, something I read just in LinkedIn before, uh, we got on the call, and it's from a guy called Tim J. Smith, who's like a leader in pricing. I think he's part of the Pricing, uh, Society of some sort. And he was saying he was looking through Netflix and stuff, and he was saying Netflix have clearly done well. They literally, every time I look on my account, it's instantly price goes up. They never tell you other than, oh, we did send you an email, you know, you want ads, watching 30-second ads, what's that about? So they're constantly experimenting. And so he was saying, they're doing all this experiment, the revenue is looking good. But I can't find anybody who actually calls themselves head of pricing, pricing guru, pricing this, pricing engineer, billing engineer, any of the titles that you would think would be associated with, well, actually, I'm proud of what I'm doing. And so he was saying, either they're hiding it all in their closet, Head of Monetization and Pricing people or billing engineers spend time, or they're just not willing to take that title. And basically, it's just not something that they see as a route to senior management in Netflix. And that's the same in finance, but pre-FP&A, and it's the same now. We want people initially to say, I'm Head of Monetization and Pricing, or Head of XYZ Pricing and Billing, or I'm a billing designer. What's that? And so, in many ways, we have to inspire people to almost create categories, because that's what you need. You want someone to say, "Well, I'm a billing designer, and this is the way it should look beautiful." "I'm a billing engineer, and I make sure..." "I'm a billing PM," and not that much.

So interesting. So, one of the things that you've written so much that I just find, A, incredibly instructive, and B, kind of inspiring, especially in this vein of we need to sort of, as an industry, grow up around how we organize and operationalize pricing as a discipline. One thing you've written is the lack of pricing education coming out of business schools is totally inadequate, certainly in a, in a pre-internet, we-sell widgets world. Yeah, but it's, it's just criminal in an internet age where there are hundreds of new businesses with lots of new monetization models, the rise of AI. And so I want to kind of try to dig into this idea of ownership within an organization. Monetization, pricing, packaging, it touches every function in a company. Often executives have different incentives around pricing and packaging. And because it's so cross-cutting, it so often leads to such ambiguous ownership, which is a real kind of reason it's often not that well run as a discipline. I'm curious if you have a point of view on who should own monetization in an organization. And maybe no one should, it sort of is contextual, but I'm curious if you have, if you've developed a really strong opinion on that front.

Good one. Yeah, I have a very strong opinion, which actually most people hate.

So we've got a hot take here.

I think, first of all, if we win it back and say, okay, what's the only thing that we need 12 people to agree? And essentially, most people are saying pricing should be done by committee. But if you're cynical and you say the only thing that we need in this environment with this society, civilized society, 12 people to agree. So, so then we say, okay, who are we trying to convict? And the problem is, without getting on that Buffett and Munger issue, basically the incentive stuff, someone has to be incentivized to make this the best it can be. They can make mistakes, they can get sacked, they can get employed, but someone has to be ahead of this and say, this is what I'm going to do. And that's kind of the first, and that already exists in a lot of firms. There are loads of pricing people in different firms, etc. That exists, but it's not, as you say, widespread. And that goes back to the— it's not in academia. It is in a few firms, I don't know what percentage, but it's less than 10% who actually say I'm head of pricing. And it does need to move forward and every company needs it. And so if you look across the space of all the functions that we have, we have product managers, they could do it. Definitely capable, definitely mathematical, but they really have a full-time job, and that's really basically making the product as functional as possible and allowing people to onboard themselves, etc. So that would be a part-time job for them, even though they're more than qualified to do it. For sales, no one likes sales, fun and pricing. So they could do it though, but no one likes it because it's back to banking in the early, with 50,000 different variations. Marketing could do it. Yeah, they could do it, but they actually have a real job, and their job is getting clicks. That's a real job, and doesn't matter how much we pay them, they will never take this part-time job, whether they take the job of getting clicks and getting awards. It's not possible. So who are we left with? Customer success. They could do it, but they're on the phone trying to get clients not to churn. So then who's left? Basically, Engineers, engineers aren't going to do it. They can do the mathematical part, but I'm not going to say engineers lack empathy, but they may lack a bit of empathy for some of the pricing nuances. And actually, they're busy doing a full-time job building the product. There's only one or two functions left. HR, HR aren't doing it. They're not going to get their hands dirty doing it. Compliance, they're definitely not doing it. They're not allowed to. So then you've only got finance left. Finance, no numbers, people like it or not. There are sexy parts of finance in sexy clothes, FP&A, and there's parts who can churn out numbers, do Power BI, do crazy analysis, go down looker rabbit holes, all sorts of stuff. And is it something that's valued? Well, not valued, but is it valuable? Every company has it, even Netflix. Netflix is making a bunch of money doing pricing changes. And to your point, Fynn, they may be happening by committee, osmosis, or They just, maybe knee-jerk, let's just increase the price to $10.99 and reduce it to $4.99 with an ad tier. Now, you and I could do that in high school, which, but it makes money if you're a big company, but not every company is a big company, and they can't operate like a big company. And so, part of making it a discipline means we have to, one, center it in a location where The finance in my mind are mentally capable of doing it. And of all the functions, and I'll probably get canceled for this, they are the least mentally challenged out of all the, all the functions. And because they've been to university, they've probably got master's in maths and MBAs coming out, CFAs. But actually, the job of finance isn't challenging for that operationally, massively challenging. But from a kind of intellectual point of view, not as much. Whereas if you look at the marketer, the market has got a massively wicked environment, very, very quick feedback loops and rapidly changing environment. It changes pretty much every month. There's a new hack, a new technique. There are no new hacks in finance every month.— there is very low feedback loop in finance. Similarly with customer service, high feedback, wicked environment, things are radically changing, number of tools they've got to deal with. So on all those vectors, number of tools they've got to do it, how wicked is the environment, how quick is the feedback, sales, marketing, product, they're all very quick feedback loops. So those guys and girls are mentally and intellectually challenged. I would say people in finance are physically challenged because there's so much work to do, volume, systems that don't talk to each other, no APIs and stuff. But intellectually, compared to where they are, they're all university grads, MBAs and stuff. So I think finance, in my mind, is a good place. One, because everybody says AI is going to kill finance. It won't. But certain jobs in finance, yeah. Could you get paid more doing product monetization? Maybe. And I look at some of the characters in the kind of field. Yeah. He was happy to put product and monetization on his title in 2010, 2011 at Facebook. Head of product monetization.

So it's just such an interesting take. And I think what's— it's just so interesting because Obviously FP&A has the analytical skillset to be really skilled and strong around monetization, but you're making an observation that they are the function in the business that has the most latent, maybe latency in their job to go pursue that discipline with the right amount of focus, the right amount of attention, maybe even the right amount of capacity to influence all the stakeholders in the business. I think that's really interesting. You know, I think in SaaS, at least my perception is there are sort of people that believe monetization and pricing should live in product, and there are people that believe it should live in finance. And clearly you're on the finance side. What's the best argument you've heard for it should live in product?

Uh, so I suppose you ideally want to build pricing as a product. And I think that's why I read the article by Jeff Dean. It's very good. If you see pricing and billing and monetization as a product in itself, then the number one thing that says to do is to, one, make it look beautiful, two, to onboard people onto it kind of easily, and then to constantly be experimenting and iterating, which finance isn't the native DNA to finance to constantly, let's do 100 A/B experiments this week. That's not the native finance thing. They can do it with push, but that's actually from a product perspective. Yeah, that's native. Let the LS run as many experiments as we can. And if you do, whatever they say about AI, we've had like machine learning for years. AI does help you run experiments. It shouldn't give you answers, but it should say, well, this is an experiment. In the same way we ran Monte Carlo simulations in finance to say, what do we think the value is? Well, That's— there's nothing wrong with Monte Carlo on that being done in product on pricing. And so product would be very good if they had the time to really own it, but that would be the only job. You are head of product for billing or monetization, whatever the title is, and monetization is your goal. Make it look beautiful, delivering it, easy to onboard clients, easy to understand, easy for them to change stuff. Almost make it so the client or the customer is configuring their own pricing based on— you call it entitlement. Very good.

I want to learn more about The Talent Capitalist, which is your business. Um, what types of customers are you serving, and how do you sort of describe your offerings and, and the services that you deliver to those clients?

Yeah, so it's If they haven't hung up the phone immediately, they start to listen. And the idea is, I suppose, to describe it as, we're recruiters for jobs that don't exist. And that sounds ridiculous because, well, I don't want a recruiter, one, I hate recruiters, and the job doesn't exist anyway. But the idea is, in a similar way to what we're saying now, we believe that the job of, let's just call it, "Munitionation Architect," make up a term. I've used that term before. No one ever really caught on. Should exist. And you're like, people are like, what is it? Well, an architect puts buildings together, does diagrams, does this, make sure the pipes go in the right place. It buys billing systems that need to, entitlements, each flag, it puts things together and said, this is the outline of your house, your quantization house. You need billing here, entitlements there, blah, blah, blah, et cetera. And you're like, oh yeah, that's exactly what I want. It's beautiful. So, so I basically will say, this job doesn't exist, but I know you need it and I'm going to help you recruit for it. I'm either going to recruit for it or I'm going to find someone who won't admit to having spare time in your, in your business. And I'm going to literally not sit on them, but basically train them up in this new discipline. And so it's normally people say, I don't want to recruit anybody, but actually I think Tommy over there has got some spare time, even though he won't admit to it, or actually he's just very keen, he wants to learn a bit more, or Jenny, or whatever, and they say, "Okay, maybe this can help them." At worst case, it's training. At best case, it's value-add. So, I think it's value-add, but at worst case, like, whatever it is per month, they're like, "Well, yeah." The person's happy because they're being taught something. And so, if you look at the kind of top-down training that everybody gets in a big firm. Some of it's compliance stuff, we get that. And some of it isn't compliance, it's actually, what would you like to learn? You're like, well, I want to be a speaker, I want to be a circus artist, or whatever. This is in a similar way, someone said, well, I don't care about economics or pricing that much, but if you ask me to do it, and you're paying for it, yeah, I'll learn it. And you find that people take interest in that. So, the two frames are, one, we're recruiting for jobs, and this, but also, Very good.

And so let's explore the discipline. So if, if for listeners who are tuning into this, what's the mental model for understanding monetization as a discipline through the lens of The Talent Capitalist, through your lens?

Yeah, so I suppose it starts with kind of trying to avoid the hand wave. So if you say Really, there's creating value, then there's measuring value that you've created, then there's negotiating how much of that value you should share in, then there's basically systems that you put in place like yours to actually capture that systematically, and then there's the price point, which is stand the monetization. And so, to take an example, I love Google, basically, the price of monetization. And if you take away all the monopoly stuff and the other stuff with issue nor by design, they've said we need to basically create value. But even more important than that, we need to measure the value. And it sometimes surprises me how many firms have not really understood Google Analytics and the Facebook Ad Manager, I think it's called. This is the gold. Yes, they've got great ads and their monopoly in the search and all that stuff, but the gold is the analytics. Because the analytics say, Fynn, if you put an ad on or give us £100, we're not going to guarantee, but if it does make £300, we're going to show you that in a dashboard. And then you're going to get excited because it's going to be a dopamine 'Oh gee, if it shows you've only made 90, then you're gonna go and beat someone up in marketing and say, "That was terrible creative."' But immediately you've got something to act on. You've got something to iterate. And Facebook, I believe, even show you how other people are doing with their ads and what ads are doing. So literally, you couldn't get a more perfect market for information symmetry than, say, Facebook ads. And so if we take from the top, create value, but more importantly, demonstrate the value you create, then negotiate. And we come back to Google, how much they've negotiated. Then monetization is some of the mechanisms, the kind of industrial mechanisms to capture it and then price point. So we take Google, for example. Google will say, on average, let's just take, you're a decent marketer, you're going to get 3x ROAS.. So you put £100 in, you're going to get $300 back, or £300 back. That means you kind of make £200. So, so the question is, okay, so you've made £200, you probably had some ad agency who were part of that. So you made £200, that means Google's made £100. So in reality, Google's taken 50% value capture because you made £200 and they take £100, so that's 50%. Your agency's taking another 10% of the £100, or 20%, so they're taking 20%. So all the way along that kind of ecosystem, the digital ad ecosystem, I think is the best form of monetization because it shows value is created, and you can say that was a good ad, that was a terrible ad. It shows the value creation measured in a dashboard. And then you can see, you can agree, well, actually, yes, the agency can take 20% of the $100, that's $20, 10% of the overall created. Google could take $100 because they've created me $200, 50%. And so all the way along the stack, you can see, and then all the architecture around that, basically is the monetization. And so to treat that architecture, whether it be billing and entitlements, say, actually, how do I put the architecture in place? You have to first believe that it is an architecture, and it isn't just a, yeah, we just have a few clicks and we collect the money. And so all the way along that stack, I see the discipline is demonstrating the value that you create, then measuring it. And I suppose people say, but what about where you haven't got this kind of straight-through process? Well, I suppose whether it's mathematical, as in Google demonstration of value capture, or psychological, you have created value. The client says, okay, there's no math, but psychology, selling, or psychological selling, I believe you created 200. Again, then you just go down the stack. And so those 5 steps of value creation, demonstrating the value creation in the dashboard for a numerical number, then how much of that can I have, negotiate your value capture, and then monetization is the architecture around capturing that value, and then price point is, is just a heuristic that makes people understand it and be able to explain it either to themselves.

Every time I read your writing, and now every time I talk to you, I feel like I come away with sort of a, a renewed or different or unique lens on things. And that's very true of this idea of, of monetization as a discipline. You know, before I asked you that question, I was asking myself, how do I think about monetization as a discipline? And I've often articulated it internally, or to put people I'm talking to is there is a price, which is just a number, and then there is pricing architecture, which is an incredibly complex system typically if the business is complex to support the value capture from that price. And I think what you're describing is so powerful because it's so accessible. It allows people to say, I need to understand what the value is. I need to be able to defend the value so that I can negotiate not to churn or to increase ACV. I need to have the systems in place that will allow me to monetize across different products, reduce friction for the customer, empower them with transparency. And price point is kind of like, it comes out of all of that. But those first 4 are maybe the most important. Edit anything about that? Is that a fair reflection? Would you change anything about it?

Yeah, if you approach it from that way, what Google and Facebook have done is they've made you agree to the price. If you're the client, you're agreeing because you can put on the creative, you can click what, what in terms do you want it to be? CRO, like optimization, or do you want, do you want someone to see your ad or do you want someone to go to your website? So you're literally configuring your own price. And I think this is where pricing is going if AI happens. And the benefit is, imagine if every single HubSpot customer could configure their own entitlements. And each of those entitlements come with, well, yeah, that's going to be your price because it would one, be unique that actually no one could go on and compare their price on G2 or whatever, because they would have to have the exact same entitlements at the exact same time. With the vaccine on the staff, it would be almost impossible. So this is where— but to do that, you have to have an ability to say to individual clients, I need to demonstrate that value to you. Google and Facebook do it easily because they say this is blah blah. We would have to be able to do that with most SaaSes and say, if I can configure HubSpot and I want this many kind of seats and this and this and usage, this is how much I'm going to pay, $8,900 £25.20.

Someone said to me yesterday that one of the things that they're most interested in as it relates to pricing and packaging over the next 5 years is that they believe we will see more SaaS companies actually mature and get very good at value-based pricing. And the argument was, hey, we're coming from a world where companies largely are not that good at understanding the value that an individual customer gets from the product. As we become more sophisticated with aggregating financial data and usage data via AI, we should improve as an industry at understanding per-customer value. That's different, I think, than we're going to see a world of performance-based pricing. But maybe react to that. What do you think about that person's argument? And what's the difference between value-based pricing and performance-based pricing in SaaS?

Well, I suppose you could argue that value-based pricing sometimes gets mixed up with usage-based in the sense that I may, you and I may have Databricks and you've got a more engineering background than me and therefore you're going to get more value out of it, even though I may waste time using as many units as you because I've just configured it badly or We both have the same usage, but we've got different value. And so if we take that then to performance, we're saying performance is also individually different. You may be happy. And let's just take, let's just take an example. Venture capital. Venture capital has moved into kind of, in my simplistic world, two worlds. You've got asset managers, got a bunch of money. And that's a different game. It still works. Basically, I've got billions and therefore there is no way they can get 10x on billions. It's just not possible or hasn't been possible to do that. And then you've got, got like a $100 million fund or whatever, can probably get 10x. And so they have two different performance metrics. So if I'm going to sell a venture capital product, then one guy and girl's happy with 2x and the other one wants 10x. So they might be using the same resource, but they've got very different perceptions of what performance is. 2x is cool, up top, that's what we're aiming for at the bottom. And that's the same for every business or whatever. And so I think they're right, SaaS is getting better. But if we look at the number of questions, so if you look at the old days, there was a salesman or woman, and they would go and ask you a bunch of questions. And they would give it a go and they would do a good job. But that's not scalable. Then we had the kind of 3 or 4-tier pricing table, and that is really asking you, Fynn Glover, the customer, questions. Gold, silver, platinum, whatever. They're asking you questions just indirectly. You're going to know the value better than me because your value is actually different. Your value of your seat on a plane is different from mine, even if we're sitting next to each other. You maybe don't see the best friend that you haven't seen for 20 years, and I'm just, just pouncing about on a plane trying to get something. So, so the idea, the only person that really knows the value is you, the customer, me. And so I think the goal is to get you to pick the price, as Google does. You pick the price, and, and then you're able to tailor it. Oh no, I don't want this, and I don't want the vegan lunch and stuff. And then you're able to change it in the same way you could have 3 people in your organization picking HubSpot, and you may configure it differently, each of you, because you may value things differently.

So the next 5 years, we see the continuation of a B2B SaaS customer who has become very empowered and demands from vendors flexibility in the way they purchase and how they are billed and in the products that they configure. And the features that they have access to.

Yeah. So if you look at the kind of RFP process, say you were doing self-source now, and we did self-source, and you had this procurement person who was pretty good at this sort of stuff, then the RFP comes into Procurement Gargoyle and says, I need these features, blah, blah, blah. He basically prices it, him or herself or whatever, via your pricing tool. And he comes back and says to the business, well, This is what you want. I don't need to go to a salesperson because I've got this pricing algo from that company, from Salesforce, whoever. This is the price. Now they've got two options. They can go and negotiate some backdoor deal with Salesforce, but imagine Salesforce has moved to a, like Google, where automated, or they can just change the features, or they can just say, no, that's the price. And so there's three options really mean that actually you should be empowered in the firm. It might be the procurement guy or girl or someone, but someone in the firm is able to accept.

So to this end, I mean, there's a lot of, there's a lot that happens from a systems perspective and from an infrastructure perspective to support that type of flexibility for the customer. As you teach monetization as a discipline, I imagine you're having to teach kind of analysis, negotiation, sort of all the kind of key skill sets associated with understanding how to price and package the strategic side of this. How do you focus on teaching those types of people, especially when they're coming from FP&A, how to understand the system's complexity, how to influence the org across systems and I ask that question maybe with one other sort of anecdote, and that's that having now spent a lot of time with many different pricing and packaging consultants, as well as pricing and packaging leaders at larger companies, one thing I often hear is, I've figured out the kind of the recommendation to the business, but I can't implement it, or implementing it will take many, many quarters. And so I'm curious how you've kind of sought to teach people with that constraint in mind.

Yeah, so it's a perfect point there, Fynn, because essentially, I suppose, from finance, you've always dealt with systems that actually away with it, don't talk to each other, and that are more complicated than they need to be. Finance is like accruals and cash, that's it. So, if you think about it in that way, yes, we are still way behind on the systems curve. But by having a cross-section and saying, this is what I want to achieve, then they know most of it necessarily won't be as achievable, but at least they've set out the stall. But what that then does is open a conversation to finance, say, when you are looking at changes to your system or from CRM, these are the kind of integrations that we need. We need this data. We, because we have this data model of what billing and pricing should be. Therefore, we know we can't get it on day one. But we try to, every time you're doing a shift or turn of the front system or the CRM or the front office system, we would like these two extra features. But at the same time, the person has to be thinking, whether they make it up in Excel or they do it in a database or they do some Python thing, that they have to be having this own data model of their own. This is my monetization system, and I optimally can't get it downstream, doesn't mean I still can't record it and manage it. Yes, it's manual, and I've got to do journals, or I've got to tell people to do this, or I've got to overlay and update finance, but it's still doable. You see what I mean? So the only part should be, I cannot do it because I have no more hours. That's fine. And then—

So it's sort of at a heuristic level, if you're looking to excel as a monetization specialist, obviously you need to have the analytical toolkit to run the analysis, but you also need to be able to start from a point of these are the inputs that we will need, this is the data that we will need. And if you can at least start there, yeah, you're well ahead of the curve at understanding how to work with engineering, work with product to evolve the infrastructure in such a way that it allows you to deliver the flexibility that pricing model, et cetera, that you hope to be able to offer?

Yeah, totally. So if you look at AI, it's kind of getting a good beating for the last few weeks with stock market and stuff. But if you look at what we've had so far, we've had this kind of Copilot concept of Fynn Glover. If you say you're in FP&A or in product, whatever, Fynn Glover, I can help you do something. I will help you do better pricing optimization. If it's in pricing monetization, I will help you because you can search my big web database. So, sub-top. Okay, that's one view. And then obviously there's you, there's the AI helping you, and then there's this pricing table, then the client interacting with it. And so the next form of that could be, well, actually, what if the AI said, Fynn Glover, can you help me? It sounds ridiculous, but imagine the AI and you turned into the AI's helper. So you're the copilot and it's actually Fynn Glover, the one driving it. It would say, give me all the data you have. First question would be from AI, all internal data you have, all the RAG, etc. Give me access to some sort of LLM. Good. Also sign up for BVD or some other kind of LexisNexis so I can download clients and stuff and stuff. Sign up to Bloomberg, I want that as well. And it would ask all these questions of you, and it would ask questions, okay, give me this database information, you have to pay 10 grand for it, but this person had this really good— sign me up for ZoomInfo. It would do all this stuff, and it would ask you questions, and you'd be like, I can't afford all that. But imagine you could, and I was asking you now for a bunch of data, relevant data, industry data, G2, this, that, And it will come back and say, here's the pricing table. Do you want to— And so if we turn it around and you have AI as the person and we are the copilot, AI would say, just like if we're trading bonds and AI is trading the bond rather than us asking AI to help us trade bonds, AI would say, give me all the bond data from 1906 and I'll run some algos on it, as some firms do. Give me all the changes. Give me all the downloads. Give me everything Jay Powell said in the last 18 years or whatever. Give me blah, blah, blah. Give me this. Give me the presidential— give me Polly Market. It would do all sorts of crazy shit. Because the AI would say, give me all this information and I'll come up with things that you can't actually predictably rationally agree to. But trust me, they work. And then you test a few or whatever. And so I kind of, I like economics, I study economics as well. Economics are, we are predictably rational. This, if this happens, Fynn Glover will do that. And then that didn't really work for everybody, but it worked for many years. And then someone said, okay, we are predictably irrational. That Kahneman stuff, actually, you would rather not lose any money, 2 times 1 or whatever. So he doesn't want to, he'll lose 100, he doesn't want to lose 50 quid for every 100 quid he can gain, all that sort of irrational stuff. So, so behavioral science and economics said, okay, let's know now we can do predictably irrational. But with AI, we almost need to go, no, this is intelligence chemistry and the compound it comes up with and the actual drug. We can't really work out, we can't work out where this vaccine came from., but we know it works after we've tested it. And I think that's the leap of faith people take is, and some, I saw this in the city play out really well. You had, in 2000, we kind of electronified a little bit. We had loads of budget, we spent a fortune in finance and stuff, and in banks on technology. JPMorgan now spending billions. But there was a shift, and it was, the shift wasn't how much you spent, It was speed. Who can get to the exchange the fastest? If you can get there faster than someone else, no matter how good, you're going to basically see that trade earlier and you're going to market. And so the Citadels and the Jane Streets and stuff, mostly Citadel, are just speed merchants. Very good. They're not better traders. Some may say they are, but they can get to the exchange quicker than some of the other firms, JP Morgan. But that happened in the early '90s, '90s. And you can still see it now because they've got this budget that they spent, and they've got these wired, etc., that goes to the exchange faster. Then the second phase of neural nets was, what if I can use a computer to help me trade better? So a lot of computers— so a lot of people were saying, well, can I use this machine learning to help me trade better? And it worked for a lot of them. I said, What if the machine just trained and I helped the machine? And then they just pile in a few limits, pile in some rules, pile in like an insane amount of data from all over the place. And this guy Ben Stancil wrote about this at the weekend. And I read it and I thought, oh my God, this guy's explained our market. And so with you, Fynn, helping the computer, the computer's going to ask you like, can I have the movement of the stars? Can I have this? Can I have the weather? Can I have what's going on? Can I have every semantic? Kind of information from Twitter.

Um, so much that's interesting there. I think we should do another episode where we really go deep on the future of monetization in the world of AI. And I'm, um, also a huge fan of Ben Stancil's writing. I think he's one of the most— he's just a clarity creator, and he's just entertaining as hell as well on all kinds of interesting subjects. Um, a closing question I like to always ask is Is there a moment in your career that was most formative for you professionally, a specific moment?

Yeah, near the end of my career in banking when I was trying to do something that was massively optimal in my mind. I proved it out. This would save you a bunch of money. And if you have a cohort partner saying, yeah, this is great, blah, blah, blah. And presented it to a senior Xcode committee. I was almost laughed out the room. Yeah, but I just completely got the incentive alignment wrong. And so it kind of taught me that actually big ideas mean nothing, as long as if they're basically given to the person that has a completely different idea of what they want. And so really, it taught me, know what people want, not what you think is good for them, or actually may be good for them. Know what they want. And so past that, I thought, actually, I need to go deeper into this whole— I got the behavioral science and incentive and like signaling theory and all the other stuff, which I don't touch. I didn't touch an accountant in school or college. And so it made me kind of try to understand the particularly irrational and the sense.

I love that. Stem interact.

Cool. Okay, Fynn Glover, thanks for inviting me. It's good to see you, and we'll speak again.