Goal setting is one of the most widely studied ideas in psychology and performance research with hundreds of studies conducted in numerous countries and contexts. Two organizational psychologists, Edwin Locke and Gray Latham spent 35 years building the body of evidence behind it and their central finding is not what most goal-setting advices on the internet tell you.
The Finding: specific, difficult goals outperform easy and vague once. The two variables doing the heavy lifting areSpecificity and Difficulty
What "specific" and "difficult" actually means here:
Difficult goals that are specifically defined lead to increased achievement compared to "do-your-best" goals that are ill-defined. "Grow the business" is a do-your-best goal wearing the costume of a real one, It feels like direction but it gives you brain nothing to aim at. "Get to $5k MRR by adding 40 users in Q3" is specific and hard for most early-stage solopreneurs. This combination, not the ambition alone is what the research ties to performance.
This isn't some one off study. A 1981 review of laboratory and field studies by Locke and colleagues found that specific, challenging goals outperformed easy or vague goals in roughly 90% of the studies reviewed. A formal meta-analysis by researchers Anthony J. Mento, Robert P. Steel, and Ronald J. Karren reached the same conclusion, finding strong support for the difficulty and specificity.
The conditions that make difficulty actually work
Locke & Latham's own 2002 review identified few more conditions that make hard & specific goals improve performance:
- Feedback on progress
- Commitment to the goal
- The ability and knowledge to actually perform the tasks
A hard, specific goal with no feedback loop and no real capability behind it isn't just a stretched goal but a setup for failure. This is the part most goal-setting content skips, because it's less quotable than "make it hard." But it's the difference between a goal that produces performance and the one that produces a slow, quiet failure.
Where the popular version quietly broke from the research
If you've set a goal in the last 25 years, you've most probably used SMART - Specific, Measurable, Assignable, Realistic, Time-related. Most people assume SMART is the goal-setting research, but it isn't. George T. Doran introduced the SMART acronym in a 1981 article in Management Review as a way to write management objectives. It was a practical framework, not an empirical study of what makes goals effective.
Here is the part that matters: Doran's original version used Assignable, Realistic but over time, acronyms drifted to Achievable and Relevant, which is what most of us are familiar with. The modern “Achievable” version can also create a tension with Goal-Setting Theory: if “achievable” is interpreted as “don't set a goal you might fail to reach,” it can push people toward goals that aren't difficult enough to produce the performance benefits identified by Locke and Latham.
That's the contradiction. The research says difficult is one of the two load-bearing variable. The most popular operationalized version of "good goal setting" has an entire letter dedicated to making sure the goal doesn't feel too hard.
Where difficult goals actually backfire and why a warning matters
None of this means "make everything as hard as possible," and the research community has had a real, sharp disagreement about exactly where the line is.
In 2009, Ordóñez, Schweitzer, Galinsky, and Bazerman argued that the benefits of goal setting had been overstated and that its systematic harms, including a narrow focus that neglects non-goal areas, distorted risk preferences, increased unethical behavior, and reduced natural motivation had been largely ignored.
Ordóñez and colleagues built their case on real companies, not hypotheticals: Sears once imposed a $147-per-hour sales quota on auto repair staff, mechanics responded by recommending and billing for repairs that weren't needed, a scandal Sears' own chairman later attributed directly to the goal. Ford's push to ship a car under 2,000 pounds and under $2,000 by 1970 hit its target, the Pinto, with a fuel tank design flaw that made rear-end collisions dangerous, because safety wasn't the goal, speed and weight were.
Locke and Latham pushed back hard in their own reply that same year, rejecting the idea that the benefits of goal setting had been overstated. But the mechanism Ordóñez and colleagues identified is worth taking seriously on its own terms: A goal that's specific enough to hit will get hit, even when pursuing it creates consequences you never intended.
The goal needs a boundary, not just a target
Here is the precise version of that design. The specific failure mode Ordóñez and colleagues identified, narrow focus that neglect non-goal areas has a specific fix for a solo operator: Before setting a goal, name 3 -5 recurring things that keep the business alive but aren't part of the goal like replying to support requests, invoicing, security patches, whatever applies and protect a fixed weekly block for them that the goal is not allowed to consume regardless of how the week goes. Setting the time block frequency to daily could actually hurt because then you would be on war with context-switching
This isn't a motivational nice-to-have. it's the structural answer to the documented risk of difficult goals: they pull all the attention towards themselves at the expense of everything not written down under them. A hard goal should earn your focus, but it shouldn’t be able to claim the rest of the business as collateral.
What this means structurally
You need goals specific enough to aim at, difficult enough to matter, feedback frequent enough to correct its course and protected boundaries around the other parts of the business that goals have not covered. The answer isn't easier goals. It's better-designed difficult goals.