Goal-setting research has one finding that has survived more replication than almost anything else in the field, and it says more than "set goals." In a 2002 review of thirty five years of their own work, Edwin Locke and Gary Latham reported that specific, difficult goals consistently produced higher performance than telling people to do their best. The finding holds across well over 100 different tasks, more than 40,000 participants, and at least eight countries, with time spans from 1 minute to 25 years. The same body of work also hands you the mechanism by which goals work, the conditions under which they stop working, and two cautions of its own: a warning about the side effects of narrow numeric targets, and a reason to set a learning goal when the work is new. Thirteen weeks is long enough to see a number move and short enough to correct it. That is the part a quarter is for.
The finding the rest of the field sits on
The headline effect sizes run from .42 to .80 across the meta-analyses the review pooled. A second result is that the effect grows with difficulty. In the same review, Locke and Latham reported a positive, linear relationship between goal difficulty and performance, with effect sizes from .52 to .82, levelling off "only when the limits of ability were reached or when commitment to a highly difficult goal lapsed." A goal set beyond the ability to hit it stops working.
Exhibit 1
Specific, difficult goals beat do-your-best goals in every meta-analysis Locke and Latham pooled.
The group-level picture matches. In 2011, Ad Kleingeld, Heleen van Mierlo and Lidia Arends pooled 49 effect sizes from 739 groups and found that specific, difficult group goals outperformed non-specific goals at d = 0.80, against d = 0.23 for specific, easy goals. The same paper carries a second result that matters for a firm of three to twenty people. Individual goals aimed at maximising a person's own performance had a strongly negative effect on the group's performance, while goals aimed at the person's contribution to the group had a positive one. Where work is shared, which is most of it in a small firm, that is the framing the data warns against. Set the quarter's goal as each person's contribution to the shared outcome.
How a goal actually moves performance
Locke and Latham described four ways a goal works:
- It directs attention toward goal-relevant activity.
- It raises effort.
- It prolongs persistence.
- It prompts people to find and use task-relevant strategies.
For an owner who does everything, the first one is the one that changes the week. A goal decides where attention goes, and what switching costs an owner who does everything is why that device matters more in a small firm than in a middle office.
A goal only works under certain conditions. Locke and Latham named four that decide whether a goal works at all: commitment to the goal, feedback on progress toward it, the complexity of the task, and the person's confidence that they can hit it. Feedback is the condition a small firm can actually engineer, and their sentence on it is why a weekly review exists. "The combination of goals plus feedback is more effective than goals alone," they wrote, because without progress information people cannot adjust their effort or their strategy. The thirty minute weekly review is that sentence put on a calendar: the review reads the number, and the owner decides whether to keep effort or change strategy.
The conditions also mark where the effect thins out. The goal-setting effect is roughly half as strong on complex work: goal difficulty effect sizes of .48 on the most complex tasks against .67 on the least complex, and .41 against .77 for specific, difficult goals versus do-your-best. The practical reading is that hard work needs a different kind of goal.
Exhibit 2
A goal reaches performance through four mechanisms, and feedback and commitment keep them running.
The cautions in the same literature
In 2009 Ordóñez, Schweitzer, Galinsky and Bazerman argued that goal setting is over-prescribed, and listed its systematic side effects: narrow focus that neglects non-goal areas, unethical behaviour, distorted risk preferences, corroded organisational culture, and reduced intrinsic motivation. The last one belongs on the list as a contested claim. They assert it and cite others for it, ran no study and reported no effect size, and Locke and Latham disputed it in print.
The warning they attach for managers setting goals applies directly to a quarter: "Be sure that short-term efforts to reach a goal do not harm investment in long-term outcomes." Note what they do with the core finding. They accept it, writing that compared to vague, easy goals, "specific, challenging goals boost performance." Their argument concerns dosage and side effects. They accept that goals work.
The side effects argument has a specific translation into a quarter. A number, pointed at, pulls the whole firm toward it, and what is not on the number gets starved. That is the narrow focus item, and it is the reason the guardrail is a written thing. Next to each goal, write down what it must not cost: which non-goal areas keep their budget, which long-term commitments keep their investment. Review the guardrail in the same review that reads the number. And check the number itself: a goal can point at a figure that is not actually holding the business back, and the cheap discipline of asking where the real bottleneck is is the way to catch that before the quarter is spent.
When the work is new, a learning goal
Locke and Latham put a specific condition on the finding. On a complex task, a performance goal can interfere with learning the task, and setting a specific, difficult learning goal instead restored the advantage. Gerard Seijts and Gary Latham extended the point in 2005: when the work requires acquiring knowledge instead of more effort, the goal should be framed around discovery. Their example is the frame itself: "find ten ways of developing a relationship with end-users of our products," set against a performance frame of "decrease costs by 10 percent this quarter."
In the business simulation they report, participants given a learning goal reached almost twice the market share of those given a performance goal, and the performance goal group did no better than the group simply told to do their best. The translation for the quarter: when one of the firm's lines of work is new, one of the three goals should be a learning goal, and it is still a goal with a number. Ten ways found, three pilots run, five clients onboarded. The number is the count of discoveries, which the weekly review can read like any other number.
The plan for the moment it slips
Peter Gollwitzer and Paschal Sheeran's 2006 meta-analysis pooled 94 tests with 8,461 participants and found that adding an if-then plan to a goal, naming the when, where and how in advance, improved goal attainment with a medium to large effect, d = .65, with a 95% confidence interval from .60 to .70. A goal cannot do this part on its own. The goal says what to hit, and the plan says what to do when the number stops moving. For a firm of your size, the plan is one line per goal, written in the first week: if the pipeline has not moved by Thursday's review, then we do one hour of follow-up on Friday before the next client work. The slip will come, and the point of the if-then plan is that the decision is already made before it does.
The shape of a quarter
A 1981 experiment by Albert Bandura and Dale Schunk is the small study the quarter sits on. In it, children given a per-session subgoal completed 74% of the material, against 55% for those given the identical workload as one end-of-programme goal, and Bandura and Schunk concluded that "distal goals had no demonstrable effects." Read that as an instruction to an owner: the end-of-year goal does the least work, and the subgoal does it. Split the year into quarters, and split the quarter into weekly numbers, because the weekly review reads the weekly number. That is the same rhythm as a steadier week: a number, a review, a decision, every week, thirteen weeks in a row.
The shape we recommend for a quarter has five parts, and they are cheap to write down:
- Three goals at most, one of them a learning goal when the work is new, with the number being the count of discoveries.
- A number in each that the weekly review can read. If the review cannot read it, rewrite it until it can.
- An if-then plan for the moment it slips, one line, written in the first week.
- A written note of what it must not cost, the guardrail the side effects research asks for, reviewed in the same review that reads the number.
- Where the goal touches shared work, framed as the person's contribution to the shared outcome, the framing Kleingeld and colleagues' data supports.
Exhibit 3
A quarter needs three goals at most, each with a weekly number, an if-then plan and a written guardrail.
What a quarter is good for is correction. Thirteen weeks is long enough for the number to move, and the same thirteen weeks is short enough that a wrong target, or a strategy that has stopped working, shows up on the number and can be changed before the quarter is spent. Building that shape, the review, the guardrail, the plan, is part of the work we cover in management consulting.