The OKR Trap: Why Hitting Your Numbers Can Still Mean You’re Failing

The OKR Trap: Why Hitting Your Numbers Can Still Mean You’re Failing

Measure what matters, but remember WHY it matters

Photo by Luke Chesser on Unsplash

Last quarter, an AI team celebrated hitting their leaderboard ranking target.

Tokens consumed: 500% above budget.

User sentiment: tanking.

Their OKR was a bull’s-eye straight through the WRONG target.

This is the OKR trap and unfortunately, it’s everywhere.

We’ve built a system where measurable outputs masquerade as meaningful outcomes.

Hit the number, celebrate the win, move to the next quarter.

But somewhere between the goal and the result, we lose sight of what actually matters. And why. So we end checking every box and still trailing when it matters.

The Token Trap Example: Leaderboards That Reward the Wrong Behavior

Consider the AI token example I referenced above. The teams had a leaderboard ranking system- ostensibly to encourage AI usage, accelerate AI augmented productivity gains, enable faster development, benefit from time and cost savings and pour it all into higher order activities.

Lofty objective- so, of course we need measurement to ensure these goals are implemented.

The metric this monitoring hinged on?

Token consumption.

Clean, quantifiable, easy to track.

Any team that wants to move up the rankings, they use more AI, incorporate AI assistance, ship more, including in more segments. Or so says the ideal path.

What happens in actual reality is unsurprising and predictable.

Teams generate longer outputs. They add redundant reasoning steps. They query the API multiple times for the same problem, caching the entire context when a simpler approach would work. Competitive teams even go as far as to run their prompts through an expansion layer before submission, padding the token count without adding intelligence.

We all know how that story is trending.

But here’s what the metric didn’t capture: the actual utility of those tokens. Were they solving harder problems, or just solving the same problems verbosely? Were users getting better answers, or just longer answers? Was the team shipping features, or just manufacturing numbers?

The OKR looked like a win. The business outcome was a loss and an expensive lesson to everyone.

Metrics can be gamed without creating value. Focus on any metric without context and eyes on how it ladders up to the results in the metric becoming the goal. And the actual goal gets buried in the monthly review.

Photo by Brett Jordan on Unsplash

The Vanity Metric Spiral: Followers, Views, and the Hollow Audience

Now let’s shift to content creation. Platform metrics look almost identical across channels: Medium, Twitter, LinkedIn, Instagram.

Let’s take a basic example.

Followers: 50K. Looks impressive.

Monthly views: 100K. Solid traffic.

Engagement rate: 2%. Quiet alarm bell.

But the follower number stays as the headline. It anchors everything else.

A creator chases it because it’s visible, it’s comparative, and it feels like success. So they chase virality, not depth. They game the algorithm. They post hot takes instead of thoughtful analysis. They optimize for clicks, not understanding.

And it works. The followers come.

But they don’t convert. They don’t read. They don’t apply the ideas. They don’t return. The audience is an ocean a mile wide and an inch deep . Outcome is the creator is drowning in the metrics that are supposed to prove they’re winning.

Medium metrics reveal this especially clearly. A post hits the home page feed. View count spikes: 5K, 10K, 15K. The creator watches it climb, thinking about reach and visibility. But the average time on page is 45 seconds. Most readers skimmed the headline, maybe scrolled past the first paragraph. The clap count barely hits hundreds. The reading list saves in single digits.

The OKR was 10K monthly views. Technically hit. The actual outcome: thousands of people passing by, and tiny fraction who thought it was worth saving.

This pattern repeats across every platform. LinkedIn posts optimized for the engagement algorithm get hollow reactions from distant networks. Instagram carousels chasing followers land in feeds of people who will never engage with the creator’s product or thinking. Facebook followers who merely glance or like in the group but never show in events or otherwise.

The metric by itself is not a vanity metric. It’s a useful yardstick. But unless viewed with the lens of the funnel, you optimize for only the numbers and create the leaky hole in the process. Then comes the realization that you can win the metrics game and lose the business game simultaneously.

Photo by Annie Spratt on Unsplash

The Reverse Engineering Problem

The trap exists because we optimize for the wrong variables or worse, only the metric not the outcome.

In the former, the process is we need a metric, so we picked one that was easy to measure. We needed a target, so we picked one that was easy to hit. We needed a win, so we looked at a scorecard instead of at actual outcomes.

But some questions keep the process grounded-

Did this AI system actually solve meaningful problems?

Did this creator build a real audience?

Is this engagement genuine?

In the second, we picked the right metric- but began to reverse-engineer the behavior that produces the number, regardless of whether that behavior produces the outcome.

The system incentivizes gaming. And we’re all subject to the incentives.

Either way, the metric becomes the focus.

This gives way to what is known as Goodhart’s Law coined in the 70s!

When a measure becomes a target,

it ceases to be a good measure.

Photo by Antonio Janeski on Unsplash

What Actually Works

The fix isn’t to abandon metrics. It’s to measure for value, with context.

For AI systems: Stop counting tokens. Count solved problems. Measure user satisfaction, actual cost per solved query, whether the output was actually used or deployed. Measure whether the system solved a harder class of problem this quarter compared to last.

For creators: Stop counting followers. Count engaged readers. Measure reading time, return readers, whether people actually took action on the ideas. Measure conversion — not to a product necessarily, but to engagement: Are these readers engaging in the next piece? Are they sharing? Are they thinking differently?

For any system: Ask the question behind the OKR, not just whether you hit the OKR.

Why did we set this target?

If the answer is because it was easy to measure, ditch it. That’s not a strategy. That’s tokenism. (sorry, pun was too hard to pass up!)

Most organizations and individuals know this, but knowing and doing are not the same. Celebrate the number but don’t simply shift to the next quarter’s OKRs. Pause and ask whether the win was real.

Start asking: Did we solve the problem? Did we move the needle on what actually matters? Or did we just get good at playing a game where hitting the number means you’re losing?

The difference between those two things is the difference between strategy and mirage.

[embed]Setting SMART goals is Only Half the Battle Are you setting up the system and support to be successful?medium.com