why traditional SaaS is actually dead
I've been a little skeptical about the RIP SaaS claim for a while (not whether it's going to happen), but when and if existing software will have time to adapt.
after a few weeks using Claude code, i'll go on record saying traditional saas is ngmi.
i'm referring to traditional cloud-based software, where the primary business model is subscription-fee based - CRM's, productivity tools, collaboration platforms.
backstory. consulting with a fast-growing SaaS startup
back in the late 2000's, I consulted with one of the fastest growing marketing-automation saas startups at the time.
I was around as the company scaled from < 50 employees, to nearly 1,000. from thousands, to hundreds of thousands of customers.
i was assigned the bigger clients in asia pacific, helping them onboard - ensuring they were getting the most from the platform, etc.
in those early days, I could submit a feature request on behalf of a user, ping the dev team, and it would be implemented end of week. toward the end of my time consulting with the company, I would often be the hundredth upvote on a 12 month old feature request forum.
this is (was?) the harsh reality of scaling software to a hundred thousand+ users.
this will seem insane in a few years.
here's why.
1 users will auto-generate their own software
while we're still away from the fully-automated software engineer - given the right context, Claude Code's one-shot game is pretty fucking amazing.
if you're someone who uses the current coding co-pilot tools everyday, you are well aware of how rapidly and increasingly good these things are getting.
additionally, it seems a small amount of scaffolding can create a step-change in capability. the difference from VS Code, to Cursor agent, to Claude Code is palpable.
im a subpar engineer, and I was able to recreate Notion for myself in a weekend. never again in my life will I pay for a marketing, CRM or knowledge management tool.
vercel's CEO guillermo rauch said: "You don't have to be a designer or experienced engineer to build beautiful products—AI tools like v0 embed best practices and design excellence, allowing anyone to rapidly produce high-quality, production-ready digital experiences."
2 traditional, rigid SaaS platforms NGMI
i just can't see a world where single architecture software exists - a single tool or service with a pre-defined set of features serving thousands or tens of thousands of different businesses - this does not seem appealing in the near unfolding future.
there will be no such thing as a single software service used by two different businesses.
no mailchimp, no Salesforce, no Hubspot, and soon thereafter - no Shopify or Wordpress or WIX - at least as they exist today.
it might be one year, it might be three years, but it won't be five.
we're moving fast towards a world in which a business owner will provide vague instructions and contextual information, go for a coffee, and return to a fully-fledged and customised website and CRM suite.
in the interim, users are going to increasingly demand more heavily-augmented software. strapping 'ai-features' onto existing rigid software won't be enough to keep users, neither will lowering subscription fees and run rate.
yes, many of these services are 'integrating-ai', they are still growing, and they still hold significant market share, but whether they'll be able to adapt quickly enough to what is coming is another story.
3 frameworks will replace rigid software
simple base frameworks which 'wrap' increasingly improving underlying intelligence, will become more attractive - efficient and affordable.
with the right scaffolding and context, simple 'frameworks' wrapping increasingly improving underlying intelligence will become more affordable, efficient and attractive to users.
it will be the users raising their own bugs or 'feature requests'. those requests will be fulfilled not in weeks or months, but in minutes - any by ai, not humans. or at least initially by a single human overseeing many well refined agentic workflows.
regardless of what the CEO's are preaching, the reality is that a thousand+ employee software company is not well positioned to make this pivot.
cursor is currently raising another $900 million at a $9 billion valuation - they were a team of less than ten people (college grads), just late last year.
as the traditional saas empire falls, smaller, more nimble startups are better positioned to provide an increasingly attractive bridge for ex-saas-users to cross into the world of auto-generated software.