How to pivot a SaaS startup: three principles for doing it right

Executive overview

Most early-stage SaaS founders face a moment where iterating on the current path stops working. A pivot — a strategic change to product, customer, or market based on new data — is the third option between giving up and grinding on.

Done badly, a pivot is just starting a new company. Done well, it sharpens product-market fit and unlocks growth.

Three principles determine whether a pivot succeeds: choosing the right type, testing with an MVP before committing, and fully committing once it's working.

The founders who break through are those who lead with go-to-market, not product.

The three types of pivot

1. Zoom-in pivot — one feature resonates; hide everything else and double down on it. Instagram dropping all non-photo features is the canonical example.
2. Customer need pivot — same ICP, different problem. Outreach built email sequencing for recruiters, then realised salespeople had the same need more acutely and pivoted there.
3. Customer segment pivot — same product and problem, different part of the market with more urgency or budget.

Identifying which type you're doing matters. Changing everything at once is not a pivot — it's a new company.

Caribou's pivot: from marketplace to staff rewards

Alex (co-founder, Caribou) originally built a last-minute home care marketplace: connecting idle care workers to waitlisted patients via government-funded agencies.

• The model had early validation — care workers earning an extra $400/month, agencies adding $700–800 in revenue per staff
• The pandemic stalled government approval cycles, making scale impossible for a pre-seed company
• The two-sided marketplace required selling to government funders, private agencies, families, and workers simultaneously

During relationship-building calls with healthcare companies, a pattern emerged: every organisation ran staff incentive programs (referral bonuses, milestone rewards, employee of the month) entirely on spreadsheets and gut feel, with no data on what actually moved retention or recruitment.

This was a customer need pivot: same ICP (healthcare operators), new focal problem (staff incentives), and a far simpler sales motion — sell to one party, not four.

MVP first, not full rebuild

• Caribou's first pivot experiment focused on staff referrals only — micro-rewards for sharing job postings and advancing candidates through the hiring funnel
• Result: became the #1 or #2 hiring source for pilot clients at under $200 cost-per-hire vs. the industry average of $450
• Only after that validation did they expand the platform to broader incentive types

The temptation after deciding to pivot is to rebuild everything. That wastes months. The right approach: test the smallest version that can confirm or deny the hypothesis, then iterate.

Investor alignment during a pivot

• Caribou's angels had invested in the problem area and the team, not a fixed solution
• No investor resisted the pivot; all understood the market dynamics
• The healthy push from investors: "Don't lose sight of the demand-side marketplace — this rewards platform could become a Trojan horse to eventually own the supply-demand layer in healthcare"
• The original vision stayed intact; the pivot was a necessary intermediate phase, not an abandonment

Commit fully once it's working

Once the MVP iterations show traction, the third principle is: don't look back.

• Keeping one foot in the old direction dilutes execution and confuses the team
• Craft a forward-looking vision that builds on the pivot, not around it
• For Caribou, the staff rewards platform — with its deep database of healthcare workers — is the path back to the original vision of matching supply and demand across the healthcare system

Go-to-market leads product understanding

Alex credited the discipline of iterating on messaging and ICP (ideal customer profile) work as the mechanism that surfaced the pivot opportunity.

• Talking to customers consistently — even before the product was ready — built the relationships that revealed the staff incentives problem
• Getting comfortable with messaging that "doesn't resonate" is a prerequisite for finding messaging that does
• Go-to-market confidence preceded product conviction, not the other way around

Founders stuck in "one more feature" mode delay the conversations that would tell them what to build.

When to consider a pivot

Ask three questions:

• Is there a zoom-in, customer need, or customer segment opportunity visible from current customer data?
• Can you test that opportunity with an MVP in weeks, not months?
• If it works, are you willing to fully commit and stop protecting the old direction?

A pivot is not giving up. It's using new data to make a better strategic bet.