This is a follow-on from a note on how technological innovations improve EBITDA margins.

-

In the 90s software was flexing its game-winning dominance. Encoding business logic into software was clearly a more powerful and efficient way of working. The best way to win had been defined, but it had one huge requirement: on-premises compute.

Those in need of software-driven solutions were required to have large installations in their buildings. A server room would house the main server, database, and backup system - and then each and every employee’s desktop terminal was connected to it. It was a huge, but necessary, cost.

A classical enterprise setup might look something like this:

The really insidious part wasn’t even the obvious costs. It was the hidden ones. Every major upgrade was essentially a mini implementation project. Companies would often skip upgrades because they were so painful, running on dangerously outdated versions until they absolutely had to upgrade.¹

Removing or reducing the costs associated with using software would be extremely valuable.

-

As the big players like Oracle and SAP spent the 90s building massive on-premises installations that cost millions and took years to deploy, the first wave of SaaS companies were emerging.

Salesforce led the vanguard, boldly declaring the “end of software” and a suite of products came with them.

So what happened to the companies that adopted them?

-

Adoption is always contextualised by risk. When deciding if you want to do something you create a mental balance scale - benefit vs. risk. The more the benefits outweigh the risks, the more likely you are to do it.

Companies that had invested deeply in on-premises systems saw huge risk with little benefit. They already had working solutions, were running critical processes, and absolutely didn’t want to invest in migrating it all.

However, if you were a small company with a small budget the idea of a million dollar on-premises deployment came with huge risk. Although SaaS products were nascent, small companies were now presented with almost all of the benefits of on-premises except now deployable for a fraction of the cost, in a fraction of the time.

There were feature-gaps, bugs, and outages, but compared to the alternative - it was transformative.

All of a sudden the rules of the game had changed. And it wasn’t just a single industry - it was completely horizontal. Across the board early SaaS adopters saw huge reductions in their IT spend whilst being able to maintain their ability to serve the market.

The advantage was leveraged and transmuted in different ways. Some turned it into lower pricing, some greater margins, and some created entirely new business models. One thing was clear: adopt the new order, or risk dying.

-

The casualties of slow adoption were everywhere.

Siebel, once the CRM with 45% market share and a $29B valuation, saw its value plummet as Salesforce grew. In 2006, Oracle acquired them for just $5.8B.

Sage, the accounting software giant, lost significant market share to NetSuite and Intacct in the small business segment. Even Microsoft had to scramble to transform its Exchange Server business as Google’s Gmail for Business gained traction with smaller companies.

By the late 2000s laggards were losing marketshare and knew they needed to change - but they struggled to figure out how.

They had invested millions in their on-premise systems and had IT departments that knew how to keep them running. Every SaaS adoption meant disrupting something that worked, albeit inefficiently.

So they compromised.

They kept their core systems on-premises but dabbled in SaaS around the edges. The result was exactly what you’d expect - they got the worst of both worlds. They had to maintain their old infrastructure while also dealing with the complexity of cloud integration. But they had no choice - the market was moving to SaaS, and they had to at least appear to be keeping up.

• “Our systems are too complex”
• “Security concerns”
• “Regulatory requirements”
• “Too much invested in current systems”
• “Our customers aren’t ready”

All of the above were used as reasons not to bite the bullet. By the 2010s the EBITDA margin advantage was clear enough that boards started forcing changes, even at resistant companies.

-

As with all EBITDA margin expansions, they inevitably return to their norm.

The vast majority of industry moved to the new ‘way to win’ and margins were erased through competition. At the same time, the pricing power of SaaS vendors increased, large scale deployments were stung by the linearity of SaaS’s pricing models, and having multiple vendors per function became the norm.

SaaS subscriptions now run into the millions for large enterprises, often costing more than previous on-premise solutions would have, yet form a necessary evil of running a technology company. Much like the deployments of the early 90s.

This pattern mirrors the 747s - a technological innovation created temporary margin expansion, but the advantage disappeared as the market adjusted.

The key difference lies in value capture: while Boeing had to compete with other manufacturers who quickly developed similar aircraft, SaaS vendors found ways to steadily expand their take of the pie through subscription pricing and ecosystem lock-in.

What started as a cost-saving revolution for customers became a margin-expanding transformation for vendors.

This raises an interesting question for the next wave of technological innovation: as AI and automation sweep through industries, will we see the same pattern of temporary advantage followed by vendor value capture?