From Legacy to Lift-Off: Go-to-Market Lessons from SaaS CMO Steve Martin

Steve Martin started his technology career at Lotus, a year after IBM paid $3 billion for it, and has spent fifteen years running marketing inside B2B SaaS companies since. That span gives him a long view of a problem most operators meet only once: what happens to growth when your own installed base becomes the constraint. He is blunt about why annual planning breaks teams, skeptical of "fail fast" for a practical reason, and clear that AI has already broken the search playbook the industry spent a decade building.

Show Notes

Steve Martin has spent fifteen years running marketing inside B2B SaaS companies, and started his technology career at Lotus — arriving a year after IBM paid $3 billion for it. That span gives him an unusually long view of a problem most operators meet only once: what happens to growth when your own installed base becomes the constraint.

The constraint nobody models: your own customers

Steve draws a direct line from Lotus in the late nineties to the CRM businesses he's led recently. Both were decades-old companies with entrenched customers sitting happily on legacy platforms, and both needed to chase markets those platforms couldn't reach.

"Your ability to move is limited by your legacy base in many cases."

The trap is sharper than the usual innovator's-dilemma framing. Build the new platform and you face a choice nobody wants to own: spend months engineering data bridges that perhaps five hundred customers will ever use, or skip them and ship faster. Skip them, and the moment you launch, your best customers ask why they can't have the new thing.

"You're maybe sending the message to the install base that we're leaving you behind. We have this other thing that we think is the future, and you're not in it."

He's watched it play out both ways — long migration paths nobody walks, or no path and thousands of hours spent manually moving the customers who insist.

Green growth versus selling to yourself

Steve splits revenue growth into two tracks and is blunt that both are hard. Selling more into the existing base is the easier motion — those customers take the call — but it has to carry real value rather than a price rise, or it breeds resentment.

New customer acquisition is harder, and it's the one that counts:

"That's real growth. That's green growth. And if you get into an exit situation where an acquirer is doing diligence on you, they're going to look at revenue for these kind of qualities. Is it green growth, or are you selling to yourself?"

His DaySmart experience shows an add-on done well. Payments looked novel fifteen years ago: heavy up-front work on the gateway, in exchange for one percent of every transaction at roughly 95% margin thereafter.

"You're getting paid every time your customers get paid, which is a little different than just cashing that subscription check at the end of the month."

The sting came later — the acquirer viewed the business as a payments play, and the go-to-market reoriented around it. In Steve's words, "if not a 180, at least a 135."

Why annual planning breaks teams

The most useful stretch of the conversation is on modeling.

"The CFO wants to know how many units we're going to sell in Malaysia next August."

Take last year and add five percent and you're accused of taking the cheap way out. Build it bottom-up instead and you spend hundreds of hours on a forecast that can't be produced accurately. Then Q1 closes behind plan, escape velocity is beyond reach, and compensation is already pinned to the number.

"I think modeling in general does a bigger disservice to anything in software operations that I've seen."

The asymmetry is real: beat the number and "you must have been sandbagging."

His answer is to fight during planning rather than after. If the model assumes a twenty percent jump in your slowest month, say so — "I can't support this. I won't sign off on it." Not as a Debbie Downer, but sticking to the data.

And when you're locked into a number you can't reach, he argues for fundamentals: lead quality, qualification discipline, onboarding, in-app guidance, free trials, referral offers. If six months in the bonus is gone, it's gone.

On "fail fast"

Steve is skeptical, for a practical reason: failing fast usually means killing something a month before it would have worked. Digital campaigns need volume for the platform to optimize. Pull out early and you've paid the learning cost without collecting the return.

His fix is a conversation held *before* the spend — with the CFO specifically, not just the CEO and sales leader:

"We're not going to launch and light up the world. You're not Apple. We can't get a million dollars on day one."

AI has already broken SEO

Asked what he did successfully three years ago that would fail today, Steve doesn't hesitate: search. The industry went from keyword stuffing to high-value content, and much of that content has now been obviated — buyers ask ChatGPT, and Google intercepts its own results with AI overviews.

"Every company blog — no one ever read those things, and with good reason. But they were designed so that page crawlers would read them... Well, now that's an utter waste of time."

Nobody has worked out how to optimize for AI overviews yet, and he's candid that the industry is waiting for a leader to emerge. He's more optimistic about the other side: building in days what took months, and agentic systems that assemble analysis faster than a Salesforce report ever could.

"We can do things we could never do, and we can do things we've always done ten times faster."