Legacy SaaS Is Worth Its Data

Over the last six months, I’ve spoken to more than a dozen executives at mature, established software companies. Every single one of them is facing the exact same existential threat. “Legacy” SaaS is now a dirty word.

And it gets worse. Customers at these legacy SaaS companies are looking at their platforms and asking a frustrating question:

“Why can’t I just build this myself?”

The hard truth? They can.

Replicating an enterprise software platform used to require dozens of engineers, millions in capital, and years of development. Today, the combination of generative engineering and AI-native infrastructure has driven the cost to replicate workflows, user interfaces, and standard legacy SaaS features down to near zero.

Because the cost to replace systems is plummeting, legacy platforms are depreciating assets. Traditional SaaS revenue is being heavily discounted by the market. A company that used to command a 15x ARR multiple is suddenly trading at a 5x or 8x metric.

But these legacy players have a massive, unfair advantage. They just haven’t realized it yet.

The End of the Code Moat

If you are running a $20M ARR legacy SaaS company, your valuation isn’t tied to your code anymore. The market has completely re-priced standard software functionality.

As Sequoia Capital notes in their landmark industry report, Generative AI: Act Two, the tech landscape has moved past the honeymoon phase of generic software wrappers. Enterprise value is shifting entirely away from standard code layers to deep, vertical context.

Furthermore, venture firm Andreessen Horowitz highlighted this structural shift in their analysis of The New Business of AI, proving that traditional software applications are suffering from severe margin compression because code alone no longer provides long-term defensibility.

Startups can steal your feature list, but they cannot steal your history.

The Great Value Divergence

While the value of the software layer is crashing, the value of legacy data is skyrocketing. Jerry Chen of Greylock Partners mapped out this exact architecture in his definitive thesis, The New Moats: Building Systems of Intelligence.

Chen argues that standard applications are commodities; true enterprise value belongs to companies that wrap proprietary data around customer workflows to create a “System of Intelligence.”

Your unfair advantage is the decades of deeply contextual, historical customer data sitting in your databases.

In the venture firm NFX’s foundational playbook on Data Network Effects, they demonstrate that while a competitor can easily clone a user interface, they can never capture or clone the compounding, defensive loop of a legacy player’s historical data assets. Your history is an un-clonable asset.

The Pivot: From Software to Intelligence

Legacy SaaS companies have never had to monetize their data before because selling the software platform itself was profitable enough. That era is over.

As Marco Iansiti and Karim R. Lakhani write in the Harvard Business Review essay, Competing in the Age of AI, software platforms that fail to transition into data-driven prediction engines face rapid, systemic obsolescence. To protect your multiple, you must pivot from a software-first narrative to a data-first narrative.

This does not mean selling your customers’ private information. You cannot let your users think you are weaponizing or selling their specific data.

Instead, the modern playbook requires three precise moves:

Anonymize and Secure: Clean, isolate, and aggregate your historical datasets at scale, keeping privacy airtight.

Build Proprietary Models: Use that aggregated history to train internal, specialized models that deliver predictive insights no generic AI can match.

Deliver Unfair Insights: Embed those insights back into your application, turning your software from a passive tool that records data into an active engine that dictates strategy.

The Answer to the Question

The next time a major customer looks at your platform and says, “Why shouldn’t we just build this ourselves?” you look them in the eye and give them the only answer that matters:

A simple example

Picture a company that has sold scheduling software to dental offices for fifteen years. Its screens are easy to copy. Its records of millions of appointments, cancellations and no-shows are not.

That history can predict which patients are likely to cancel, which days run short, and which offices are about to grow. A new competitor with better screens can't offer that on day one. It would take them years of customers to catch up.

Where to start

1. Inventory the data you already hold and how far back it goes.
2. Get clear permission and privacy rules in place before anything else.
3. Pick one prediction your customers would pay for, and build that first.
4. Show the result inside the product, where customers already work.
5. Tell investors the data story, with numbers, at your next board meeting.

What buyers look for

When private equity firms or strategic buyers look at software companies now, the questions are shifting toward data. How unique is it? How clean is it? Can it train something useful? I wrote about what buyers' models miss in what private equity spreadsheets leave out.

If you're curious what your company might be worth today, the valuation tool gives a range in a few minutes.

Stop fighting the commodity software war. Your code is a liability. Your data is the moat. Turn the brain on.

Related reading: What the AI-Native Executive Team Looks Like · The Rise of Hollow ARR in AI Companies · AI-Native Companies Move Faster in Every Department