Stretch Goals Are Lying to You

Every OKR guide tells you to set stretch goals. Aim for 70% achievement. If you're hitting all your targets, you're not ambitious enough.

This advice is well-intentioned but often destructive. At 25-75 person companies, stretch goals frequently help directionally but hurt operationally. The core mistake is using a single goal to do two incompatible jobs: enforce execution and pull ambition upward.

Those should be separated.

The core tension

There are two competing truths about goal-setting:

• If you always hit goals cleanly, you're likely sandbagging and leaving potential on the table.
• If goals are framed as aspirational, teams learn that missing them is acceptable, eroding accountability.

Both are correct. The failure mode is collapsing these into one number.

Why stretch goals break down at your stage

Companies of 25-75 people are especially vulnerable to stretch goal dysfunction.

Execution now matters more than inspiration. Most value at this stage is unlocked through focus, sequencing, dependency management, and operational hygiene. Stretch goals introduce ambiguity right when precision is required.

Goals interact with performance and identity. At this size, goals touch performance reviews, promotions, credibility, and resourcing decisions. Ambiguity around "was this meant to be hit?" becomes corrosive.

Second-order effects dominate. Missed stretch goals lead to defensive planning, buffer-building, risk avoidance, and optics over outcomes. Teams adapt rationally, just not in the way you want.

The real problem

You're asking goals to do two different jobs:

1. Set a hard execution bar
2. Express ambition and upside

When combined, both fail. The execution bar becomes soft because "it was a stretch anyway." The ambition becomes theater because no one really expects to hit it.

A better model: Commitment + Explicit Upside

Instead of stretch goals, separate committed targets from upside potential.

Committed Goals (Required)

These are:

• Expected to be hit with competent execution
• The basis for planning, staffing, and sequencing
• Binary and taken seriously

Missing this triggers a real retrospective. These targets aren't conservative; they're defensible.

Upside / Acceleration Band (Optional)

This is:

• Explicitly not required for success
• Not used for performance evaluation
• A guide for opportunistic decisions

Key rule: Not hitting upside ≠ failure. Hitting upside ≠ redefining success after the fact.

Ambition is preserved without poisoning accountability.

"But we'll leave room on the table"

This fear is usually overstated.

If you set tight committed goals, review progress frequently, reward outperformance, and reinvest wins aggressively, you still capture upside.

The real risk is optimizing for theoretical maximums instead of actual throughput.

High-performing teams push harder when momentum is visible, constraints are removed, and success compounds into trust and autonomy. Not because numbers were inflated.

When stretch goals do make sense

Discovery-heavy work. R&D, new product bets, and zero-to-one initiatives should use goals framed as learning targets, output ranges, or scenario envelopes. Not pass/fail commitments.

Founder or exec-only contexts. Founders can stretch themselves because incentives are fully aligned, failure cost is internalized, and no one else's performance is implicitly judged. Stretch goals degrade once they cascade.

The anchoring philosophy

Goals exist to create clarity and alignment, not to maximize theoretical effort. Ambition should live in scope, sequencing, and follow-through, not inflated targets.

Or more bluntly:

• If a goal matters, it should be hit.
• If it's optional, don't pretend it's a goal.
• If you want more output, improve leverage, not numbers.

A useful pressure-test

Ask yourself:

If the honest answer is "Nothing, it was a stretch anyway," then it wasn't a goal. It was theater.

For more on building an operating system that creates real accountability, see how confidence scoring offers an alternative to percentage progress, or read about the problem with self-reported OKRs and why truth matters more than intent.