Four engineers in Dublin built a side tool just to talk to their own customers.
That tool became Intercom. Sixteen years later, it sold to Salesforce for $3.6 billion.
I’ve spent my career watching companies protect their revenue at all costs. Intercom did the opposite. Its self-cannibalization strategy is the real reason it had a business left to sell.
▶️ Listen to this episode
[ Apple Podcasts ] [ YouTube ]
A Chat Widget With No Business Plan Behind It
In 2010, McCabe, Traynor, Lee, and Barrett were running a crash-monitoring tool called Exceptional.
They had paying customers and almost no way to reach them. No CRM. No message history. Just a clunky export they had to wrangle by hand.
So they built themselves a way to message users directly. That side project outgrew the main product.
They sold Exceptional, moved to San Francisco, and rebuilt the company around one idea: make internet business personal.
That’s not a slogan I’d greenlight lightly. Most founders say something like it. Few structure a decade of decisions around it.
Selling Small When Every Investor Said Don’t
Silicon Valley orthodoxy in the early 2010s said you couldn’t build a real software company on small businesses alone.
Intercom ignored the room.
Traynor hand-wrote outreach to former customers, mocking up what Intercom would actually look like on each recipient’s website. No automation. No drip sequences.
That’s brutal work. I’ve run teams that resisted exactly this kind of manual selling because it doesn’t scale.
But it does something automation can’t: it puts every rejection directly in front of the person making tomorrow’s decisions.
Paired with a simple “We run on Intercom” badge on every chat widget, the company turned its own customers into a sales channel. Revenue hit roughly $50 million in annual recurring revenue within a few years.
Marketing to the Job, Not the Persona
Growth stalled once the “one platform for everything” pitch got muddy.
Nobody wakes up wanting a communication platform. They wake up needing to capture a lead, onboard a new user, or clear a support queue.
Intercom turned to Clayton Christensen’s Jobs to Be Done framework and re-interviewed its own customers about what they were actually hiring the software to do.
This is the part most marketing case studies skip: the insight didn’t just change the messaging. It changed the product.
Intercom unbundled itself into distinct products mapped to real jobs, so the pitch matched what customers were actually buying.
It also built a content engine — books, a blog, podcasts — instead of gated whitepapers chasing form fills. Owning the conversation is what let Intercom charge like a premium vendor instead of competing on price.
The Drift, the Cuts, and the Return
Scale dulled the edge, as it usually does.
McCabe stepped back in 2020. Karen Peacock roughly doubled revenue to $200 million. But headcount swelled past a thousand, and the company started resembling the bloated enterprise vendors it was founded to replace.
The board brought McCabe back in October 2022. He cut about 13% of the workforce and rewrote company values into operating rules instead of wall art.
Engineering output and morale rebounded within a year, even with a smaller team. I’ve seen the opposite happen more often — cuts that gut a company’s capacity along with its payroll. The difference here was pairing the cuts with a clear operating rulebook, not just a smaller org chart.
The Self-Cannibalization Strategy That Rebuilt the Company
Then ChatGPT launched, and Intercom’s leadership read the moment correctly, fast.
Support economics had always come down to one formula: inbound volume times handling time, divided by headcount. AI could push handling time toward zero.
Most incumbents in this position protect the legacy business. Intercom didn’t. It launched Fin, an AI agent, and openly declared war on Zendesk with a product that threatened its own seat-based revenue.
That’s the uncomfortable core of a real self-cannibalization strategy. It’s not a slide in a deck. It’s shipping something that makes your own sales team’s job harder, on purpose, before a competitor forces the issue.
Pricing for Outcomes, Not Seats
Intercom didn’t stop at building Fin. It rebuilt pricing around it.
Instead of charging per seat, it charged a flat fee — starting at 99 cents — only when Fin actually resolved a query.
Customers stopped buying software access. They started buying finished work.
I’ve sat through enough pricing debates to know how rare this is. Sales teams fight to protect the seat model because it’s predictable. Intercom tore it up anyway, and Fin scaled to more than 30,000 business customers with a resolution rate cited around 76%.
The Rebrand, the Sale, and the Real Lesson
In May 2026, Intercom retired its own name and became Fin — the product had outgrown the company that built it.
Weeks later, Salesforce agreed to acquire Fin for roughly $3.6 billion, folding it into Agentforce. It’s the largest deal ever involving an Irish-founded tech company.
A billion-dollar valuation in 2018 turned into a $3.6 billion exit eight years later, and the bridge between those two numbers was a willingness to attack its own revenue model twice — once with pricing, once with the product itself.
What This Should Actually Cost You Some Sleep
Every executive I know says they’d disrupt themselves before a competitor does it for them.
Almost none of them do it. Protecting this quarter’s number always feels safer than protecting next decade’s relevance.
Intercom’s board and founders made the harder call twice. That’s not a growth hack. That’s the entire story.