As the son of an English Lit teacher, I love reading. But I’ve always had a tough time with business books — especially the sales ones. They all have something interesting to say, buried under hundreds of filler pages. A lot of work for a little insight.

All that to say, my bar for sales books is really high. And my favorite of all time is The Sales Acceleration Formula by Mark Roberge. 

Before it, I was captivated by the art of sales — the heroics and creativity that elite reps deployed to get deals done. But then Mark’s work introduced me to the science of sales, and made it more exciting than the heroics.

Mark has a new book coming out. The last one framed sales in the SaaS era; this one is here to help us understand the science of scaling in an AI world. I got early access, and I can tell you that it’s worth your time.

What follows is a conversation between me and Mark about the lessons from this book. We talk about what he thinks is the biggest imbalance in startup decision-making, the place for art in sales, how AI will reshape selling, and his worries about today’s startup culture.

Asad Zaman: I really enjoyed reading your new book, The Science of Scaling. In it you frame two fundamental questions we need to ask ourselves: Are we ready to scale? How fast can we scale? And you then share a scientific approach to working through them. What were you seeing in the last few years that convinced you that this work needed to be published?

Mark Roberge: If you think about how important the decision is around when to scale and how fast, and compare that to the rigor thatʼs put behind that decision, it has possibly the biggest imbalance of the entire startupʼs decision-making. That’s something worth trying to fix.

Asad Zaman: Why is this imbalance so prominent and persistent?

Mark Roberge: First off, I think there’s a simple blueprint that people follow: Start a business, find product-market fit [PMF], and convince someone to give me a lot of money at as high a valuation as possible. To get a lot of money, I have to convince investors that I have a big market. And I also have to convince them that Iʼm going to go fast. They buy the plan and then I do the math. And it means I have to go from one salesperson to 20 in a month.

The other problem is the contradictory incentives and misalignments of founders and VCs. As a founder, your company is your bet this decade, whereas as a VC, your company is one of 20 bets in this two-year vintage. 

So, a VC would almost prefer you to unhealthily scale this thing to see if it works. Go big, or go home.

Thatʼs what causes some of it.

Asad Zaman: Between these two questions — are we ready to scale? and how fast should we scale? — which one is the bigger problem right now?

Mark Roberge: “When should we scale?” is a little easier for folks because fortunately over the last 30 years in entrepreneurship, weʼve gone from building startups in 1995 where we actually sold multimillion-dollar contracts before we even had a product — it was called vaporware — to having the work of people like Eric Ries and The Lean Startup on PMF.

These days, we build products hand-in-hand with customers. We have design partners, and the investor community is on board with many low-revenue quarters while we figure out PMF.

Then we go from that to hiring 10 reps a month. That’s the biggest problem. 

Asad Zaman: And in that problem, what’s the most shocking part?

Mark Roberge: It’s really weird that we set this annual plan and live by it like itʼs a religious doctrine. When youʼre <$10M and you miss Q1, often theyʼre telling you to hire more reps.

I think weʼve got to have a flexible mindset where, on the one hand, itʼs okay to set up the annual plan and try to go 4X, but when we check in again in four months’ time, and we might see some things in the business that tell us to go faster. Or, on the other hand, we might see things that tell us that the plan is right, that we should stay at the same pace. Or, we may see some things to tell us that we have to stop, fix, and then get back on pace. 

But people just don’t think that way.

It’s almost like VCs think that if they get off that 30% IRR track in one quarter, it’s over. But you’re not going to sell this business next year. You’re probably in this for five or 10 years. And God forbid you sell it seven years from now, no one’s going to ask what you did, what your growth from Q1 to Q2 was seven years ago.

Asad Zaman: Yeah, it doesn’t have to be this straight, perfect line.

Mark Roberge: It never is.

Asad Zaman: Okay, I’m bought in! I want to be scientific. How should I think about GTM now?

Mark Roberge: In the book I suggest thinking about GTM just as you do an economic system or a financial system. At the highest level of the system, thereʼs a scientific frame. As you gain scale, you can calculate your ideal customer profile [ICP] with pretty good accuracy. From the ICP, you get your buyer journey, and then, your sales methodology. The inputs to the system are your hiring profiles, and then how you generate demand. The system runs every day, with the way you pay and coach your reps. And it spits out activities that lead to a forecast that lead to revenue. 

Now, AI has the potential to elevate each component. For instance, if you were world-class last year, your Rev Ops team may have refreshed the ICP every quarter, and they probably evaluated three to five dimensions — company size, geography, high industry. AI allows us to do it across an infinite number of attributes every day. 

Asad Zaman: Got it! Now, where does the art or creativity fit into this? 

Mark Roberge: The creativity comes in the modifications: At some point, a rep will come up with a new way to present a feature, and maybe Product adds their own spin on it. Thatʼs a creative breakthrough. 

Or let’s look at hiring: At scale, we can attempt to attribute the knowledge, skills, and attributes [KSAs] of our hires. And we can attempt to correlate KSAs observed in the interview with performance in the funnel. Thatʼs science! But the creativity comes in when you work to see a new pattern or ask an interview question or assess behavior in a different way. 

Eventually, when something new starts working, then that creativity is justified by the science. 

Asad Zaman: I find people struggle with this balance a lot. Some over-index into no allowance for creativity and art, and others might over-index the other way. How does one find the right balance themselves?

Mark Roberge: I think it depends on the context. On one side, I want to extract the art out of it as much as I can, because that suggests a lack of prediction. But on the other hand, there are a lot of contexts where you never want to go that far, because there isn’t always one path to success. You need to choose a leadership context where you can understand your true self and let that come through, so you can thrive as a leader. 

There’s some merit to that as a seller, too. Some people are extraordinarily natural relationship sellers that can build rapport in seconds. We don’t want to stifle that.

Asad Zaman: Do you think deal size is an important input when trying to figure out the balance? 

Mark Roberge: Deal size is a massive input into how much art should be there. It could be the strongest, but I don’t think it’s the only one. 

I can give another largely correlated one: how much evangelism is necessary? If you’re selling pencils right now, not so much. But, selling AI agents, very much so. Even if the deal size is the same.

Asad Zaman: I’ve been thinking a lot about the transition from SaaS to AI, in the context of GTM. The last time we went through a transition like this was on-prem to cloud, and GTM changed completely. With AI, we’re probably going to have profound change again. 

Curious, what do you think are GTM’s biggest strengths going into this transition?

Mark Roberge: Great question. We’ve made amazing progress in the last 20 years at bringing more science to sales. In a lot of cases in the 1980s and ʼ90s, there wasn’t a lot of methodology, and there was next to no data in the CRM. It was more like, “Here’s a territory and a phone book. Come back with revenue or you’re fired.” Then, we entered the cloud era and Inside Sales brought data, which brought a lot more science to the process. 

Moving forward into AI, that’s going to be a strength, because that science translates into first principles around model design and optimization.

Asad Zaman: Cloud had a really interesting impact on sales through deal sizes. They came down for similar products because of SaaS pricing, and that drove a number of changes in how we sold, including enabling the assembly line that we see now. What do you think happens to this assembly line in the AI era?

Mark Roberge: I have a lot of conviction that we’ll reverse the specialization curve that you mentioned. Now, there are pros and cons to specialization: The advantage is you get to specialize a role by both skillset and cost.

If you look at the entire spectrum of skills from “get the meeting” to “sell the meeting” to “renew the customer” to “expand the customer,” many people argue that “sell the meeting” is the most challenging one to find. Hence, AEs make the most money. And you donʼt want to waste that skill on getting the meeting and renewing the contract.

However, specialization comes with a cost: If you comp an SDR on the number of meetings, you get a bunch of shitty meetings. In contrast, if an AE is in charge of setting their own meetings?

Asad Zaman: You don’t get shitty meetings! 

Mark Roberge: Exactly! With AI, the first benefit we’re seeing — and will continue to see — is an increase in actual selling time [i.e., the percentage of time a rep is in front of a customer or prospect]. Best in class has been 25% to 30%. That’ll rise to 80% — it already has for some companies — because AI eliminates a lot of administrative work, and it also provides an amazing crutch for all aspects of the process. 

This means that AI decreases the specialization advantage, and I think we’ll see a move to full-cycle reps who get the meetings, close the meetings, renew the companies, and expand the customers.

Asad Zaman: Assuming AI keeps getting better, where does this go then?

Mark Roberge: When you go into Star Trek world, I think it could transcend traditional departmental and functional boundaries. For example, take a B2B software company: You have Engineering, Product, Sales, Marketing, HR, Finance — because of human brain limitation. There’s not a lot of PhDs in finance that do code and vice versa.

But, because AI elevates our abilities, we’ll have the opportunity to blur these functional lines: Finance would love to be closer to what’s happening in Sales; Product would love to be closer to whatʼs happening in Customer Support; etc. 

Asad Zaman: We see the advantages of this in the B2C version of the Growth team that Chamath [Palihapitiya, founder of Social Capital] came up with. It’s a team that uses engineers, designers, and data scientists to find viral growth, and has obviously worked really well on that side of the market. 

Mark Roberge: I’m so glad you brought this up because I hadn’t made this connection. 

The challenge is directly in line with the functional divide: In B2C, Growth teams were successful when they were under Product. In B2B, that was extremely rare; they were in Sales, and it stifled the issue because the whole premise of a Growth team is that it’s full-funnel and cross-functional. Meaning, they owned everything from awareness of the product, to activation, to setup, to value extraction. That required marketers and engineers and designers.

The problem with B2B implementation is when you put Growth in Marketing or Sales, you can’t attract engineers. You’re limited up to the activation. After that, you can’t do any experimentation. 

So, you’re absolutely right: Those well-implemented B2C Growth teams could be used as a guide for us to reimagine the organization as these functional boundaries blur.

Asad Zaman: You’ve dedicated the proceeds of your new book to McLean Hospital, a global leader in mental health research and care, which I really respect and appreciate. Stepping away from GTM for a minute, what are your thoughts on the shift in the culture of work in tech? Does it worry you?

Mark Roberge: If we look at what’s going on in the tech world right now: It’s a gold rush. I don’t know if we saw a lot of weekend work in the ’80s and ’90s. Maybe there were some super-passionate people, but it was probably unusual. Then, the rise of the startup career in the early 2000s meant you had to sacrifice in order to gain the upside — working your ass off for little money and high-risk reward. First, it was expected of the founding team. Now, it’s expected from everyone through ’til the IPO.

It doesn’t sit well with me instinctively, but I’m a little worried that I’m too oriented toward my 49-year-old self. If I was entering this as a 22 year old I think I’d be okay with it.

What I’m more concerned about is if folks coming in perceive this as a lifetime career. I’m hoping it’s a temporary gold rush, which the moment deserves temporarily. Because there’s a long-term impact of this lifestyle.

Asad Zaman: Health-wise?

Mark Roberge: I remember when I was in business school 20 years ago, they brought in a lot of alumni from 20 years prior who were like, “Look, I was in your seat 20 years ago and now I’m CEO of XYZ Company. I’m about to retire. I have three homes. And I’m miserable and depressed. I never lived a life. I have no wife. My kids don’t talk to me. My health is degrading and I have no friends.”

It makes you think.

Asad Zaman: My last question: Do you have any advice or thoughts for if and when we hit the bottom of the Gartner Hype Cycle, or the bubble bursts? If it happens, I worry it will be tough for GTM people.

Mark Roberge: For those running organizations: Donʼt give up. Thatʼs when organizations are made and you probably have to pivot quite a bit.

And, for the talent out there: Donʼt get stuck in a walking zombie of a company that’s not pivoting. Oftentimes in these tech disruptions, the most successful companies are started in the “Trough of Disillusionment.” So, don’t go back to consulting in banking, as they did in the cloud era. AI is the future, and just because we get to an era where itʼs a black cloud and poo-pooed on, that’s the time to triple-down.

Become a Topline insider by joining our Slack channel.

We want to hear your feedback! Let us know your thoughts.