AI moves faster than your budget cycle
Hi Everyone,
The Three Horizons model from McKinsey tells you to split your projects three ways: today's core business, emerging opportunities, and a few long-term bets that sit years away.
The model still works, but AI is changing how quickly projects can move between those horizons. A five-year bet can now become a viable product in months.
Today, we'll walk you through four questions that show whether your planning has caught up with that speed.
When a future bet becomes today's priority
At the end of 2022, Intercom's growth in new revenue had fallen for five quarters in a row. Six weeks after ChatGPT launched, the company had a working prototype of Fin, an AI agent that answers customer support questions.
CEO Eoghan McCabe says Intercom then put close to 80% of its R&D budget on Fin while it was still a tiny share of revenue. Fin grew from $1M to over $100M in yearly revenue, and this June, Salesforce agreed to buy the company.
In a matter of weeks, something that would have once sat in Intercom's "future bets" bucket became a priority worth putting serious resources behind. Moving that quickly was a deliberate choice. Leadership made the call, shifted the budget and resources, and accepted that other projects would have to be dropped.
Could your company make the same kind of strategic shift, at speed, rather than scrambling to catch up? The questions below will show you.
Four questions for your next planning meeting
Where did last quarter's experiments end up? If they were all cut or moved into the main product plan, the budget you set aside for long-term bets may now be funding today's work. Make sure that was a deliberate choice.
Has a supplier already shipped a feature you thought was years away? If OpenAI, Microsoft, or your main software supplier now includes it for free, decide within 30 days: build a version that goes deeper, use data they can't access, or close the project and move the team.
Where are the people who ran your last big bet? If they went back to routine work or left the business, your next bet starts without anyone who's done it before. Keep one or two of them involved in whatever comes next.
How often do you review the split between core work and experiments? Once a year is too slow when a future bet can now become a product between two of your reviews. Revisit the split every quarter, even when nothing changes.
The projects that get cut first
After a big bet becomes part of the core business, it's tempting to cut every other experiment because innovation feels covered.
But your Horizon 2 projects — the ones building your next source of growth — need the most protection. A new market or a second product still takes time, even if AI can speed up parts of the process. Choose one Horizon 2 project, give it a fixed target (like 500 users by December), and fund it until that date.
Go deeper
👉 Steve Blank: The Fatal Flaw of the Three Horizons Model — read this before your next planning meeting, and scroll to the comments for the reply from the model's co-author.
👉 Mostly Metrics: How Intercom reaccelerated growth with outcome-based pricing — use this to see the 80% decision from inside the company, with the numbers behind it.
👉 Madrona: How Founders Can Apply Zapier's AI Playbook — read this for a second example: Wade Foster paused all of Zapier for a week after GPT-4 launched to rewrite the company's plans.
Coming up tomorrow
Your first profitable quarter is a milestone, but it can also lead to bad decisions. In tomorrow's issue, you'll learn what to keep steady before you spend, hire, or expand.