It’s All About the Apostrophe

[Speaker 1]
Thank you. Indeed. I'm David.

He's Mark. Hello. A slight change to what was originally published, but we're going to talk about an apostrophe.

So let's get into that. Why the apostrophe matters. And this is a little bit of a potted history of where customer success has been and where we think it may be going.

And please get your Slido's out, lots of questions we've gotten, and hopefully we'll have a little bit of controversy and dispute. My moniker given to me by a friend was a CS heretic, so that's where I start from. Customer success, where did we start?

By the way, my background, I founded a SaaS company back in 2000. I appointed my first CSM in 2004. So I have a track record in this field.

Customer success. And this, by the way, is a generalisation, but a lot of people think of it as a team, and I'm not talking about just people in this room, but the people that you work with and work for. But it has a history of being focused on activity and sentiment metrics.

It has actually in the past been reluctant to embrace revenue. I can remember sitting in one of these meetings, probably nine or ten years ago, and I remember a very well-known CCO standing up and saying, I don't want my team anywhere near revenue because it will break the trust with our customers. And that's part of the heritage we've got.

And it's always been a little bit isolated, weak connections with the team. So let's introduce the apostrophe. Let's focus on customer success and not customer success of the department.

So where are we headed, or where are some of us already? It's where results drive revenue. By results, I mean results, measurable results for your chosen customers.

And that's not something that just happens post-sale. It's something that runs through the whole thread of the company and its go-to market. And the thing that joins it all up is that focus and a set of really, truly shared frameworks, shared ICP, shared view of the customer life cycle, shared view of the results that matter in the form of a library.

And I think if you look at what's happening, where we're headed is one, go-to-market organization. We've seen a lot of CROs taking over CS, and I think we'll see more and more of that consolidation. Now, I'm not saying that's necessarily one department, but I think we'll see much more joined up.

And one of the things that AI is very good at is identifying problems with silos. So that's my take. Mark?

[Speaker 2]
I can tell you what my experience has been over the last, I suppose, 10 years working with AI. So actually, my background before RevOps or CSOps was actually headless content and headless recommendation engines. And it's kind of interesting now to see customer success or post-sales motions catching up with where, I suppose, the B2B or the B2C market was 10 years ago.

So we actually built our first AI agents in Wigmore maybe two and a half years ago. And the first one was fun, learned a lot, probably learned more about what we shouldn't do than what we should be doing. But essentially, what happened was we tried to optimize the wrong thing.

And when we said the optimizing of the wrong thing, we tried to optimize the renewal journey, but only from the point of sending of an email. And of course, what happened was we missed the exceptions. We missed everything else.

And rather than actually improving the outcome, we actually made it worse. So that was kind of our first foray into this. Then for phase two, we actually had a look at the entire renewal process in terms of a PIA or a customer-led growth motion from day one to day N, which was the renewal date.

And by and large, that was successful until the organization decided to change its renewal process. And then, of course, the AI couldn't keep up, and then we were back to the drawing board. And what I was trying to do over the last year was trying to find a way of better defining the problem here.

And there's kind of this old, when I say old, I mean 1964 principle called Levitz-Diamond, which suggests when you're using technology, if one aspect of technology, people, processes of change, everything changes. And what I'm frequently seeing is kind of us getting back to the original principle, which is that AI is not just a capability. It's allowing us to break down those silos.

So rather than actually having a look at a single part of the customer journey, that if it's aligned the entire customer journey, then it will be successful. But the key aspect, as to Dave's point, is that change has always been constant. And as long as the organization is aligned along, I suppose, those value principles and you are measuring the right thing, you typically will get the right result.

But it's been an interesting journey from my perspective. But I think, Dave, from your perspective as well here, it's the organization design rather than just the workflow design that's key.

[Speaker 1]
Yes, it is. I love quotations. One of my favorites is, every organization is perfectly designed to achieve the results it gets.

And if you want to change the results you're getting, you have to change the way that the organization is designed. Now the other thing I think that a lot of people make is they confuse organization design and organization structure. Organization structure is only one part of it.

So when you start to think about organization design, what you start to think about is, if we are truly focused on the customer and enabling them to achieve the results that they want, what does that mean for our organization? And I think what you do is you put that red thread of customer results right through the whole of the organization. So I think a lot of the mistakes that we see being made, and by the way, they're made at an executive level as much as they are at a sales level, in fact, even more so.

Once you figure out the measurable results that matter to key roles in your chosen customers, it's not as difficult to do as people think. We've just always shied away from it in the past. And that becomes what marketing markets, what it communicates.

It becomes what sales sell and the basis of their value proposition. It becomes what customer success and professional services delivers. And it becomes what the product is focused on.

And again, we make the mistake of building products and we add feature and feature and feature and feature, whereas what we should be thinking about products is, how do we enable key roles in our chosen customers to achieve the results that matter to them? How does our product help them set a goal? How does our product help them track that?

How does our product help them and enable them to achieve that? And that's the basis of what we should be thinking about products. And I'll come back to that in a minute.

So it's a complete life cycle approach, and one of the mistakes that a lot of people make is they think of this as a post-sale thing. It's not. It's a whole thing.

And I started a SaaS company way, way back. I mean, these gray eyes are not just a sign of age. August 2000, I started a SaaS company.

The one thing that changed the performance of my company back then was we built this entire life cycle about measurable results for chosen customers. We even scrapped the marketing process and the sales process and replaced it with defined stages that the buyer and the customer were going through, all defined in the way that they defined it for themselves. So we effectively built the organization outside in.

Now, that's not new today, but it was back in 2010. And it made a huge difference to our go-to-market performance. So what you have to do is think about, how do I get this organization aligned?

And the key to it is shared frameworks. Does the organization truly have a truly shared view of ICP, which is built around the measurable results that matter to key roles in your chosen customers? It's not about one metric.

Yes, there's often one metric that's labeled as being the reason for making the buying decision. If you look at buying now, buying is a committee. I think there's research recently from Gartner or Forrester which suggests that in an enterprise sale, there are typically 13 roles involved, of which five or six are the key.

So what you've got to do is figure out what matters to each of those, each of those key roles, and how do I deliver that? And if you've figured that out, then that becomes, as I say, the basis of the whole of your go-to-market. How many people have read the Blue Book?

You all know this one, the original customer success by Nick Mehta, Dan Steinman, and Lincoln Murphy. It's still worth a read. And in that book, there are 10 laws of customer success, and there are two I want to highlight which I still think are very, very relevant now.

One is customer success is a company-wide philosophy, but we've moved away from that. CEOs found it very convenient to say, customer success, that's your job. You take care of that, and everybody else ignored it.

It won't work. It never will work. One of the other rules of those 10 is product is the primary vehicle for delivering value, and we've lost sight of that.

In fact, quite often, customer success has been filling the gaps that the product has. Now I did some work. I published an e-book in 2016.

It was called Product-Led Customer Success, and it basically says, and it showed how to build the customer success process into your product itself. The problem then was twofold. One is we did not, thank you, we did not have the will to do it, but we also lacked the technology to do it.

We now have that, and Mark will come on to that in a minute.

[Speaker 2]
Well, the technical aspect is kind of the interesting thing from my perspective, because by and large, if you have a look at your customer journey, it probably dips in and dips out of 10 different platforms or 10 different systems. What I typically have seen in the past is that, and I'm going to use a very specific use case here, but it was a revenue impactful decision that an organization made simply because of a system constraint where we had a CSP on one side and we had a CRM on the other side. The CRM didn't have access to all the data points to be able to identify leads, and we didn't have actual permission to be able to create leads in the customer success platform.

They had to be done in Salesforce, but by and large, that was this typical system constraint that we were facing at the time. The typical other type of constraint you'll see, which is kind of related to the system design thing, is budgetary constraints. So marketing is responsible for lead generation or lead qualification, or marketing is responsible for communications, and what you'll end up there is rather than kind of a system silo, you'll end up with a functional silo, and then lo and behold, MCP comes along, and suddenly the silos aren't technical.

They're process and they're outcome based. So rather than actually trying to interface into multiple of these systems, that red thread, rather than actually disappearing between the systems of record, is now splicing straight down from it, and it was actually quite interesting even on the keynote today to hear that Gainsight's top adopted feature is now MCP, but it's interesting to see that we're still talking about the feature rather than the outcome that MCP can deliver.

What it typically does deliver is ability to access the right information to make a decision, but sometimes where it is missing is what is the right decision to make, and that's where we've kind of taken a different view to this, is that unless you can provide kind of the organizational context centrally, rather than actually living in a system, then your AI or even your individuals will still come up with the wrong answer, and it was actually a cheeky question that I posed to Dave earlier, is what's the system of record for strategy?

How does anyone know where we're going? By and large, it doesn't exist, so if we can't codify the destination, how do we not only ensure that AI is going to act right now, but how do we make sure that it's going to act in the best interest of the company and also their customers, and actually the Karpathy quote is probably the interesting one here, and it's about what we prioritize when everything is easy to prioritize, is that we see quite a lot of agents that are still designed around functional lines, such as, say, a retention agent or an adoption agent, but where we found very much as a point of interest is that if we put in a customer outcome agent on top of it that also had a kind of an identification in terms of the cost to serve, then suddenly the AI, irregardless of the model, was actually giving better results, but it was interesting because from my perspective, we've never codified strategy, and it typically lives in different objectives and different systems, and in certain cases what we found is it's not aligned, so say your support team would have different objectives to your customer success team, which will have different objectives to your product team, and then what happens when all these things compete, but it's an interesting concept that I think we're kind of missing.

There's a lot of people trying to optimize something, but sometimes I think the best approach here is maybe to step back and figure out what we're optimizing for, so the time-save optimization is typically kind of not what we should be thinking about, because endlessly what happens is you'll end up spending more time doing something else, but are we optimizing for customer results? Not only retention, but also how do we tie the customer results to the financial number as well? Because what will happen is, and what we've seen happen in the market at the moment, is that to an extent you'll see a remarkable kind of short-term increase in retention, and then a major drop-off pretty much off the back of it, because we're still not optimizing for the long-term success of the customer.

[Speaker 1]
If you think about it from the perspective I always looked at, which is a CEO, what I'm interested in is optimizing the whole piece, and I think one of the biggest mistakes that we're making with AI at the moment is we are optimizing in silos.

[Speaker 2]
Indeed.

[Speaker 1]
In fact, I've just put a post up on LinkedIn this morning which is, is AI a friend or a foe to alignment of our organizations? And if you're not very careful, and by the way, one of the things that AI is good at doing is looking at a broader set of data. So AI, I think, has got huge power to help us rethink the way that our organizations work, and there's quite a lot of work being done at the moment in certain fields around what does the new organization look like?

And it's probably increasingly going to be not departmental focused, it's probably going to be project and capabilities focused. Much more dynamic than we've got at the moment. But the problem we're stuck with is we're all applying AI to our own little silos.

Well, if those silos are not aligned in some way, and I believe that customer results, measurable customer results, is the strongest basis for that alignment, if marketing is optimizing for leads and sales is optimizing for close rates and products optimizing for adoption and customer success is optimizing for outcomes, what effectively you've got is four or five different teams all pulling in a slightly different direction.

So AI has the power to actually make the problem worse. And if you don't get, if you don't solve that problem before you apply AI in a substantive way, it's just going to make bad things happen quicker and faster. Sorry, quicker and bigger.

So what we've got to do is, and that's what that second line is all about, are we using AI to rethink the system or just optimize the old one? And everything I see, and please tell me if you think I'm wrong, is that marketing is using AI, sales is using AI, product is using AI, CS is using AI, but nobody's really looking at the whole piece. We've got, I mean, if you look, follow SAS and there's a guy called Jason Lemking runs a group called SASTRA, go and read his stuff about AI SDRs and what they've done there.

Great. But the question we've got to ask is, is SDR the right model anymore? Not just automate what we do.

And by the way, this is not new. If you look at history, every new significant technology advancement has started by optimizing and improving the way that we've done things always in the past. Henry Ford once famously said, you know, how did you know to build a car?

Did you ask customers? He said, no, because if I asked customers, all they would say was I want a faster horse. But he had to, he rethought the whole value proposition, he rethought the whole way of thinking about travel and transport.

And that's where we've got to get to. We've got to sit above each of these silos and say, what's right for the whole piece? And I can go back and look at what we did at my company, Click Tools, years and years ago, and it was, if I had not as CEO gone in there and said, we're going to look at the whole chain, and there's a way of doing it by those shared frameworks that says, all right, I can build that architecture and allow the individual teams, the individual experts to put their expertise into that, but everything has to comply with that. And key to that is metrics.

So can I ask another question? Your companies, how many of your companies are reporting at exec team level, board level, the measurable results that their chosen customers get? Put your hand up if you are.

There's one half hand up. It's not right. I have always said, how can we claim to be in customer success if we aren't measuring and tracking the success of our customers?

It's a bit of a fallacy. So I would encourage you all to go back and think about how do I work with my colleagues in marketing and sales and products and services to get us to build? And, by the way, I think CS is in a great place to do that.

I'm working with a company in Europe at the moment, and I'm working with the CS team, and what we're doing is we're building this library of results that matter to customers, but as we build it, we're going to sales and saying, is this what you hear with customers talking about? We're going to marketing and saying, what response do you get for these sorts of messages? So we're almost using CS to drive this change up, and the phrase I use with the guy is, let's drive it through stealth.

We're not going to say, and you need to do this, this, this, and this. We're going to say, this is what we need to do. We'd like your expertise and input to help us do that.

And you find, oh, that's useful. So then you start to break down these barriers. You start to say, you know, there was a session earlier about handovers.

The best handovers are the ones that don't exist, because they're not needed. So another thing we do with organizations is, and if you take this view about customer success being a company-wide philosophy, you bring customer success, i.e., measurable results for your chosen customers, and that becomes a sales process deliverable. A sales isn't closed until the customer's agreed that says, this is our first milestone, this is how we set it based on our maturity and the results we know that matter, and this is the plan to get us to that stage very quickly.

I'm not interested in what we're going to be in two years' time, though that forms part of the bigger picture. How do I get them there very, very quickly? And by the way, research suggests that in a lot of cases, if you've not got the adoption that's delivering those results within 10 to 20 days, you're probably already in the possible churn bracket.

So we need to completely think about how we optimize the whole, and as I say, I think we're in a really good situation.

[Speaker 2]
Actually just a quick question for the audience, and it's more of a hands question. Is everyone finding that their AI capabilities or budget are still kind of delineated along budgetary or departmental line? So hands up if that's the answer.

Marketing has their own AI capability, sales has their own AI capability, customer success has their own AI capability. Okay, that's interesting. Now, how many times have they disagreed with each other, or are they even talking to each other?

So you can see there's a great saying that I, it was actually a Gainside employee that told me to it, I don't know if they authored it, but an organization should never see your own, or sorry, a customer should never see your org chart. So are we now in a situation where AI is the org chart to a certain extent? But for every time that we've kind of had a look at the AI capabilities, not only of a platform, but even of a team, they've never actually agreed on one thing.

So to a certain extent, if you ever want to have a bit of fun, watch three AIs with different outcomes bicker about what they should do next, because it is absolutely hilarious. Because they can never reach consensus in terms of what happens next, because they're all optimized to solve the specific problem that they were designed to optimize, which is typically departmental or budgetary. Now try and figure out what that's going to be in terms of the perspective of everyone who's touching those AIs.

So marketing is getting the AI perspective from the marketing team. Custom success is getting their AI perspective. Finance is telling, AI is telling them that you need to chase more invoices.

So how do we solve for it? And to a certain extent, there's this thing called one truth, many lenses, which is if every single AI is working off the exact same context with the exact same definitions and working towards a shared outcome, then the decision can be made. But I'm desperately kind of afraid that we're just about to repeat the same mistakes for the last 20 years.

[Speaker 1]
It is back to that bottom line there. We are using AI at the moment to just rethink and retool what we currently do. The real breakthroughs will come when we put that aside and say, how do I optimize this?

And I think the question for people in this audience is, how do I make that happen? Because you should be the people that know customer results that matter to your chosen customers better than anybody else. How can I use that knowledge to spread the word, to actually get the change by stealth, get the change, get everybody thinking around the same things?

Because if everyone's focused on customer success and not just customer success focused on it, then it's a very different organization. And that's where you get the breakthroughs. You get the breakthroughs through completely annihilating silos.

So that leads us to our final point. Here's a provocative thing. I don't believe that any organization will ever be successful if it's not focused on the success of its customers.

But I question, in the future, maybe even today, do we need a customer success department to do that? And if not, what takes its place? Because I think the people that crack that particular conundrum will take a lead.

[Speaker 2]
Well here's an ethereal one. We're talking about organizational structure here, but the way that I've seen most successful customer, I suppose, initiatives working is when you do have these cross-functional tiger teams. And I think a lot of organizations have this model already.

It's just scaling it, I always thought was the challenge. Because by and large, if I'm looking to inject a capability into an existing organization, there has always been the presence of this tiger team, this cross-functional team with remit and agency that can actually solve a problem without actually having to go to a line manager of department B. But just as another probable question for the audience is that, do you see any organizational redesign where customer success is, I suppose, de-compartmentalized into an organizational-wide function rather than the department itself?

Is there a show of hands? Still no. Interesting.

How many of you guys are bickering with sales and marketing? We kind of see the problem.

[Speaker 1]
There's a really interesting point, and that is, if you want to drive change, typically what you do is you get a cross-functional team together. You get a cross-functional team together, and they think about the design, they do the selling of it internally, do the implementation of it, and then that team disappears. Well, what if that is actually the model that we need for the organization of the future?

It's made up all the time of outcome-specific, that could be for customers and or for the organization, of outcome-specific teams. Probably the people that know this best are, and the lessons I always take, is the engineering and development team at Spotify. People don't have specific jobs.

They have roles and capabilities. So I'm an expert in this field, and all Spotify does all the time is it brings its engineering people, its teams together in different formations to deliver very specific things. Why can't that be the model that we use?

Part of the problem is that we are inculcated with organizational silos and hierarchy. It goes back to the Industrial Revolution in the 17th, 18th century. Hierarchy and determinism.

Nature's got a lot of models that we can follow. For example, which termite is the project manager for building a termite mound? Who designs it?

It isn't. It's an emergent process. Which leader is the leader of a shoal of fish?

No one. These are much more dynamic forms of organization, and I think what we're going to get is more of that, because AI will take a lot of that coordination capability and make that possible.

[Speaker 2]
It will, and you still have the same challenge, which is the financial responsibility inertia. To be honest with you, it's a frustration for me as a person who just wants to build cool things, is that these are always the constraints that we typically face when we're accelerating. But how do we help it, David?

[Speaker 1]
I don't know. Well, I think we have to let go of some of the things. I think we have to let go of the fact that we are a CS person.

I remember when I was chief exec, and one of the things I said to my leadership team, when you come to this table, you are a business person first and a departmental team leader second. When you come to this table, you have to think about what's right for the business as a whole, even if it's not right for your team at that point in time, because it's that greater purpose that we need. So I think we'll end up putting much more emphasis on purpose and culture and knowledge than we do at the moment.

We use hierarchy to solve that, and it's not going to work. It's going to break.

[Speaker 2]
And I think as systems become, I suppose, less bounded, the traditional constraints of complexity as well are also being diminished. So when you have access to the right information at the right time and centrally, then certainly one of the constraints that we've typically seen from an organisational design is becoming a little bit less pertinent.

[Speaker 1]
We've spoken. Any questions? I think we might have stunned them into silence.

[Speaker 3]
Thank you both for such a great session. We've got some questions rolling in, so we can jump straight in.

[Speaker 1]
Please go. Stand at this point so we can dodge the bullets.

[Speaker 3]
So how can leaders actually measure CSM success when moving to outcome-led solutions?

[Speaker 2]
There's a latency issue here. So people work off the assumption that as soon as a change is made, the outcome is immediate. And what typically happens here is that people will, I suppose, over-index towards the immediately measurable outcomes.

So while you typically will see, and actually this was kind of a point of debate between myself and David earlier on, is that we have a tendency to try and measure quickly what we can measure. But to a certain extent, a lot of it is optics. So the way that I would traditionally try and do this is back to kind of the whole tiger team approach, was measure something that has not been measured in the past, which has a commercial impact.

And we're seeing more of this with the PLG motion lately, is that a typical good use case for being able to deploy AI is to deliver something that has not been delivered before. So say, for example, the launch of a new AI feature. If you want to build a capability around, I suppose, the adoption and the attribution of revenue to say a PLG motion towards that, that's a really easy one to measure.

And it's a really easy one to deliver value for, because it's incorporating product, it's incorporating sales, and it's incorporating marketing. And it's something that new. So you're not trying to retrofit it on top of an existing process.

You're creating a net new process and something that is traditionally you haven't managed before, at pace, and with a proven outcome, which essentially is additional sales, with a cost to serve it as well. But it's a great question. It's not easy, is the bottom line.

But try and find something with a commercial impact that has not been measured before, build a process, an end-to-end process around it, to David's point, and see how it runs.

[Speaker 1]
I think there are two parts of the answer that I would give. One is, I wouldn't focus on the CSM success. I would focus on customer success.

So my first thing would be to go to the how we've arranged our portfolios and say, show me a measurable improvement in the results that matter to your chosen customers, as they would report it. And the reason I'd start there is because it gets us looking at what really counts. The second thing I would do, I would recognize that the CSM is not the person that delivers success.

It's the organization. So one of the things I did back at my company again was, and I never quite got to being able to scrap sales commission completely, although that was going to be the next step, one single company-wide dashboard. And there were really two basic things on it.

One was successful customers. The second one was our net ARR growth. And that applied to everybody in the organization.

It was bonus-based, it was paid quarterly, and we tried to shift away from. So the problem with measures is, as soon as you put a measure on somebody, that becomes the goal. And it's not.

That's not the goal. The goal is, it doesn't sit in here, it sits in our customers. So I'd start there, because it forces us to look outward, not inwards.

[Speaker 3]
Great. How would you balance the pressure leadership puts on having AR tools and features with the need to have proper and defined, verifiable outcomes and metrics in place early and necessary step?

[Speaker 2]
This is a frustration for me, is that I'm always concerned of a leader who just says, we need to do more AI. It's like we need to be more customer-centric, or we need to be more strategic. It's undefinable.

There's no right or wrong answer to this, that there is a certain period of exploration. And we're seeing that quite a lot at the moment, where people are exploring with MCP. And even when you're looking at the metrics we've had today, we saw with Gainsight, we've used MCP as the highest adopted feature.

Great. Have we developed an outcome off the top of it yet? Maybe not.

It's the whole concept here of what is our AI strategy, versus what is our strategy while leveraging AI to be able to deliver something meaningful. I always get concerned from the pressure that's being put on people just to adopt AI for AI adoption sake, because to a certain extent, we're reintroducing the same old problem, which is adoption does not necessarily lead to value. I'd actually love to have a chat to the person who put this question, because it's not one of mine.

But it's not AI adoption, it's the outcomes that we're delivering off the back of AI. And I've seen it even internally, ironically enough, is that if you get people to use AI for everything, you're actually changing what made you special as a person and as an organization, which is delivering results for your customers.

[Speaker 1]
I think I would turn the order of the thing around. I'd go to my leaders and say, these are the outcomes that really matter. Do you agree?

And by the way, those should be outcomes for the customer and the company. Don't try and trade one off against the other. I always look for what I call the and answer.

What is it that's going to do this and that? And then once I've got that agreement, then I go back and say, well, here's how AI can help us. Here's where we should stay away from AI completely, because it's got no role to play.

So you reverse it.

[Speaker 3]
You mentioned that a deal shouldn't be considered closed until the first measurable customer outcome is agreed. How do you make that work when sales are still measured on bookings and closing deals? Terrifying question.

[Speaker 1]
I'll tell you why I tried to go about it, and it's back to that one company metric. And I've been thinking quite a lot about this quite recently. If I was to introduce a company metric now, it would be the ratio between CAC plus cost to serve and lifetime value.

And I would make that part of everybody's scorecard. I would make that the top item in everybody's scorecard. Because, I mean, you know every day, how much churn do you get because sales don't sell properly?

One of the advantages of being a CEO is you can start to influence those things. So you've got to be able to pull data together that says, look, these customers stay a long time, and their lifetime value is this. And by the way, the longer they stay, the CAC plus cost to serve to our lifetime value tends to fall.

I'm not saying it's always the case. So if you get more of these and less of those, we're going to improve our lifetime value, we're going to improve our growth rates, we're going to improve our net revenue retention rate. You've got to model it and use wherever possible the data you've got.

And just say no, no, no. I once worked with a customer, I could never get the CEO to accept this, but we were going to, we did a lot of work on ICP. And I suggested to this guy, we should bonus sales, commission sales based on ICP fit.

So if you bring in a $100,000 deal, and it's target ICP, right in the centre, we'll give you 105% of commission. But if you come and it's a, we've got a $100,000 deal, but it's nowhere near ICP fit, I'm only going to give you 80%. That'll cure it.

Charlie Munger, who worked for Berkshire Hathaway, the, who's the other guy? Warren Buffett. Well known investment company.

He always said, if you show me the comp plan, I'll show you the outcome. And by the way, that again, is something else that's modelable, particularly with AI.

[Speaker 3]
So next up, what would an executive need to be shown for them to move from focusing on sales and close rates to the existing customer success?

[Speaker 2]
I actually did a model around this a long time ago, but it was essentially measuring the number of customers who went through a renewal cycle versus the ones that didn't. Then we measured the cost to serve just from a point to sell perspective, and also an onboarding perspective from a hosting perspective, from a support perspective. But actually, we were able to measure a very specific delta between the financial benefit of a company with a customer that was retained for renewal cycle versus the financial leakage of a customer who didn't actually make it through the first renewal cycle.

And what happened was there was more of a proxy metric. But the interesting thing was the reputational damage that the customers who left after one year did because they went to a competitor. And then the competitor also used them in the marketing material to say that your product is actually very, very difficult to implement and you don't receive value.

So one of them was actually proven with numbers, and the other one was actually proven with the experiential data that was used against them by their competitor. But it was actually an interesting thing that it actually changed the way they sell because of it. It didn't happen immediately, but they were a little bit more cautious in saying yes to everything in the sales cycle.

But a lot of it with this time is feeding back the right points of information to the right people at the right time so they can make the right decisions. But it is not an easy point to sell.

[Speaker 1]
I don't want an exec that's going to move from focusing on sales and close rates to existing customer success. I don't want it. I want an exec who's looking at both.

Because if you get the sale right, you get the... And again, probably a lot of you have got the data to be able to model that, to show that.

[Speaker 2]
It is actually modelable. It's making sure that you get your ammunition in early before you fire your gun. Because typically people want proof and always kind of keep the proof in your back pocket and learn to use it at the right moment so that people make the right decision.

But there's more than just the cost of sale. Engineering effort as well for bad fit customers is usually insurmountable.

[Speaker 1]
If that sounds unreasonable, I make no apology for that. Because all progress is based on the unreasonable person. George Bernard Shaw, not me.

[Speaker 3]
How do you see the CSM role getting measured with current evolution? For example, just AR targets or growth targets?

[Speaker 1]
Both. So in the work I do with companies I've talked about this earlier, there's a concept of a two-column dashboard. Column one are the traditional SAS metrics, unit economics.

ARR, NRR, growth rate, CAC ratio, that sort of thing. Column two, results for customers. What's the average that they're achieving?

What's the top? What percentage of our customer base are actually achieving results within a certain time frame? And it's both.

I keep going back to the same thing. Always be careful about looking at trade-off answers. Put that little bit of extra thought in and saying where is the solution that answers both this and that?

[Speaker 3]
Next up, how to facilitate the cross-departmental alignment on a common AI strategy?

[Speaker 2]
Very, very carefully. So the important thing here is experimentation. So no one actually has the full answer in terms of how to use AI, but there are common themes.

And there also are themes from history. So firstly, cross-functional AI. Secondly, don't be afraid to experiment.

Experimentation is needed because you will learn more from your mistakes than you will from your success. Don't put all your eggs in a single basket or a single vendor. So there is the build versus buy argument, or the build or buy.

Do both, but do it in a controlled manner with costs at the center as well. But the interesting one actually, what we found was successful, is measure results long-term. And don't optimize it for, say, time-save.

But if you can actually attribute an AI initiative to an actual measurable result, and even in isolation, by and large, what will happen then is it will start to blend into your other organizational initiatives when it comes to AI. What actually we found out was probably the most, and it may sound remarkably boring, but it is not actually AI. It's creating common definitions that the actual AI can understand, as well as the systems.

So a lot of the time that we've played with LLMs and any sort of AI capabilities, if it actually doesn't understand your definitions, you're still going to get the wrong result. And if you actually put the definition of customer value in terms of also the measurement capabilities in somewhere centrally, by and large, the AI, even with a free model or a lower-cost model, will deliver better results than kind of your frontier models, as well, because it's got better context. And especially with the fact that the models are changing so quickly, the one thing that don't change is organizational knowledge.

So find a better way to encode it and to be able to deploy it across multiple AIs, because otherwise what you'll end up with is what I've seen quite a lot of, is that, firstly, the AI will hallucinate. Secondly, the AI will give you an answer that is not grounded in the reality of the organization. And thirdly, the AI, if it's not told what to optimize for, it will optimize for the wrong thing.

[Speaker 1]
I've heard this from CEOs a few times, and they always give me the same answer. Before you can have a common AI strategy, you need a common purpose and you need a common view of how you're going to serve your customers. That's the spine on which you can then build your AI strategy.

[Speaker 2]
I'm going to get controversial here. Well, not controversial, but is there an AI strategy, to your point? I mean, we're not just creating a strategy related to AI, rather than a strategy related to outcomes.

[Speaker 1]
Yes. We jump into technology too soon.

[Speaker 2]
Indeed.

[Speaker 1]
Or too often. We create new categories, we create new products, which quite often drive us back into our stovepipes. What you need is, I think, what my friend Mr. Lyons calls a business operating model. A common business operating model. How does this business operate as a whole? And it's not a book this thick, but really good ones are two or three pages.

Here's the customers we're serving. Here's the lifecycle we're taking them through. Here's the results that matter to them.

Here's how we're going to measure success for them and for us. That's the core of what you need. Once you've got that, then you can say, right, AI can help us here.

We need people there. AI can help us here.

[Speaker 2]
Yeah, I mean, I think my bugbear with AI strategy is kind of the same term. Firstly, strategy has always been undefined in terms of its inputs and outputs. The same thing that customer centricity has been always undefined in its inputs and outputs.

It's all very good saying, what's our AI strategy? But the point that you made earlier is, what the hell are we trying to achieve here first? And then that will dictate how you leverage AI as one capability in your organization, rather than the outcome.

[Speaker 1]
Can I quickly, because we've got 20 seconds left. I'm not sure it does. I think the future is, and I said this at a conference a couple of years ago, CS, customer success, is a distributed activity across the whole of the go-to-market organization.

Now, does that mean there's no role? No, because someone has to, I did some work with a company and they said, oh, let's appoint a CCO. And we ended up saying, no, don't do that.

What you need is a chief customer value officer. One person, three person team, did it all. So, I think we've got to think outside.

If we go with what we've currently got, then the danger is we'll always get what we've always had.

[Speaker 3]
There's some great questions left, but unfortunately we are out of time. I want to thank Mark and David for a great presentation again, and let's give them one round of applause.