AI Guides › Playbooks
By Nigel Guy · 7 min read
Most people set the "realistic" goal: last year's number plus ten per cent, the promotion that is next in line, the client list that is slightly longer. It feels sensible because nobody can call you reckless. The catch is that everyone with a sensible goal reaches for the same sensible methods, so you end up competing on effort in the most crowded lane available.
The rule: Write the goal that would force you to change the method, then prove it is ambitious rather than fantasy by building a ladder of checkpoints underneath it.
There is a real idea behind this, and it is worth knowing where it comes from so you do not oversell it. Astro Teller, who leads X (Google's "moonshot factory"), wrote that it is easier to get people to work on making something 10X better than 10% better, because aiming for 10X means you cannot just slog through; you have to find new ways of doing things. Larry Page put the same stance more bluntly in a 2013 Wired interview, as reported by secondary sources: a 10% improvement means doing the same thing as everybody else.
Read that for what it claims. It is an argument about method: a 10% goal invites you to polish the existing process, whereas a 10X goal makes the existing process obviously insufficient. It is not a claim that big goals are easier to hit, and it comes from a well-funded research lab, not from a controlled study of your job or business.
The goal-setting research adds a useful second half. In their 2002 American Psychologist summary, Locke and Latham report that, across many studies, specific and difficult goals produced higher effort and performance than easy goals or "do your best". They also state the limits: performance levels off or drops when the limits of ability are reached or when commitment to a very hard goal lapses, and the effect is smaller on complex tasks where you still have to discover the right strategy. So "harder" helps only while the goal is specific, you are committed to it, and you can see how you are doing.
The ladder has four rungs. You fill them in order, and the goal does not count as chosen until all four exist.
| Rung | What you write | Pass test |
|---|---|---|
| 1. The sensible goal | The goal you would have set by default, with a number and a date | You could hit it by doing more of what you already do |
| 2. The unreasonable goal | A version 5 to 10 times larger on the same measure, same date | Doing more of the same visibly cannot reach it |
| 3. The mechanism | The one thing that would have to be different (channel, price, product, partner, process) | You can name it in a single sentence and it is not "work harder" |
| 4. The first rung of proof | A cheap test, within two to four weeks, that would show the mechanism has any pull at all | It has a number and a pass line set before you run it |
Do it deliberately. It becomes your floor and your control: if the big goal collapses, you have not lost the realistic outcome, and you know what "normal" looked like.
Multiply on the same measure and deadline. Keep multiplying until a straight-faced "just do more of it" answer stops working. For many goals that is around 5X; for some it is less. The point is not the multiplier, it is that the method has to change.
Ask what would have to be true for the large number to happen. Sales example: ten times the revenue is not ten times the calls; it might be a partner who already holds your buyers, a higher-priced offer, or a product that sells itself. If you can only answer "more hours", the goal is a bigger chore, not a bigger idea. Go back to Step 2 and lower the multiplier or find the mechanism.
Design a small test that tells you whether the mechanism has any pull. Write the pass line before you start, as in a test card: "If three of ten partner conversations lead to a second meeting, I continue; if none do, I change the mechanism." That keeps you committed to the goal while staying honest about the method, which is exactly the combination Locke and Latham's limits point to: commitment plus feedback.
Run these four checks on the unreasonable goal.
Two or more "no" answers mean fantasy. Fix the checks, not your enthusiasm.
Priya runs a one-person bookkeeping practice with 12 clients and wants "to grow". Her sensible goal is 15 clients by next September. Her unreasonable goal is 60, which no amount of extra networking could deliver. Her mechanism: serve one niche (independent physiotherapy clinics) through the two professional bodies that already reach them, with a fixed monthly price. Her first proof rung, set for the next three weeks: ask both bodies for a short slot or newsletter mention, and pass if at least one agrees and produces five enquiries. She fails the control check on the full 60 only partly (the bodies decide), so she keeps the 15-client floor and adds a direct-outreach route she controls. Her goal is now large, specific, testable and survivable.
Fill in the square-bracket placeholders with your own goal, deadline and constraints.
You are a sceptical but constructive planning coach. I want to test whether a bigger goal is worth choosing over my sensible one.
My context: [YOUR_ROLE_OR_BUSINESS_AND_SITUATION]
Measure I care about: [THE_METRIC]
Current level: [CURRENT_NUMBER]
Deadline: [DATE]
Resources and limits (time, money, skills): [CONSTRAINTS]
My sensible goal: [SENSIBLE_GOAL_WITH_NUMBER]
Work in this order:
1. Restate my sensible goal and say what method it implies.
2. Propose an unreasonable goal on the same measure and deadline, large enough that the current method visibly cannot reach it. Show the multiplier.
3. Name up to three different mechanisms that could plausibly close the gap. For each, say what would have to be true.
4. For the best one, design a test I can run in two to four weeks with a pass line and a stop line, both set before I run it.
5. Score the goal on four checks (mechanism, feedback within a month, my control, acceptable floor if I miss by half) as yes or no with a one-line reason.
Rules: If any input above is missing or vague, ask me before answering rather than guessing. Do not invent statistics, market sizes or benchmarks; if a number would help, tell me what to look up. Do not accept "work harder" as a mechanism. Before you answer, check that every step uses my numbers and that nothing relies on a fact you cannot support.
Format: a short table for steps 1 to 3, then a bullet list for the test, then the four-check score.