Define What "Great" Looks Like

Between the Nos and the Points

Ted Williams, The Science of Hitting (1970):

A good hitter can hit a pitch that is over the plate three inches. But a great hitter waits for the pitch he wants.

Chapter 2 gave you the pitches you never swing at. Dealbreakers are the balls out of the zone: hard nos, machine-checkable, applied without fatigue. This chapter is about the pitch you’re actually waiting for: the one you’d pull for a stand-up double, the one that matches your swing. Great hitters don’t swing at everything in the zone. They wait.

Most prior treatments of “know what you want” are either aspirational fluff (mood boards, vision boards, “dream role” listicles) or reduced to a title-and-comp fill-in. Neither survives contact with an actual sweep. What survives is a small number of ranked criteria with a paragraph-length definition of “great” for each one, sitting cleanly between the dealbreakers that reject and the rubric that scores. You do not need a finished picture of the dream job before you start. You need circles you already own, and a way to keep drawing.

Why criteria come between dealbreakers and the rubric

Dealbreakers filter. The rubric scores. But if you go straight from filter to score, the weights float. You end up with a 100-point rubric whose top dimension is “compensation” because that’s what every template leads with, even when compensation isn’t actually your #1 at this stage of your life.

Criteria are the missing middle. They answer three questions the dealbreakers don’t:

Of the things you’d say yes to, which matter most? The priority order.
What does “great” look like for each one? The definition, not the adjective.
What are you willing to trade off against what? The order of the list is itself a decision.

Only after these are written does the rubric have anchors. “Compensation is worth 25 points because” needs a because that comes from your criteria, not from a template.

The clarity interview, honestly done

You can run this exercise on yourself, but most people can’t. Not because they’re not thoughtful. Because self-interview without pushback collapses into the answers you’d give a hiring manager. What produces real criteria is a structured interview with someone (or something) that refuses to accept vague answers.

I ran mine as fourteen questions across four groups:

Background and direction: career history, target titles, target industries, skills to lean into or step away from.
Work content: what energizes you, what drains you, what makes work “worth doing.”
People and environment: management style, team shape, culture and values.
Structure and lifestyle: flexibility, compensation, trajectory, visibility, dealbreakers.

Fourteen questions is not a magic number. It’s about right for a single sitting where you’re forced to articulate things you’ve never said out loud. Fewer and you skate. More and you drift.

The friction of the interview matters more than the questions. Every time I answered vaguely, the interviewer pushed back: reflected what I’d said, named the pattern I was avoiding, asked why I’d said “the chapter I liked least” without saying what was actually wrong. That friction is the whole exercise. Compliance kills it.

AI is now able to fill this role, but only if you instruct it to push back rather than affirm. “Be my career coach. Ask me these questions one at a time. When I answer vaguely, reflect what I said back and ask a sharper follow-up. Don’t move on until my answer is specific.” That prompt, or one like it, is the whole trick. Left to its defaults, most AI will nod through your answers. Ask it to be the friction.

Draw the Owl

“Draw the Owl” is a core value popularized by Twilio, from a 2010 meme:

Step 1: Draw some circles.
Step 2: Draw the rest of the owl.

The joke is the gap. The instruction is the point. When you’re building something new, there is rarely a manual. You put down what you do know, take ownership, and figure out the rest in motion.

A job search is the same kind of blank page, with worse marketing. Job boards will happily draw something that looks like an owl if you feed them titles and buzzwords. Auto-apply tools will draw a hundred of them overnight. None of that is your owl. Your owl is the role that matches who you actually are under pressure: the floor you won’t drop, the craft you won’t fake, the life constraints you won’t hide, the ceiling you’re honestly optimizing for right now (not the one a career coach would invent for a generic senior candidate).

Owning your own search means refusing to wait for a finished picture before you start. It also means refusing to outsource the picture to volume.

The circles are already in your hands. Comp floor. Remote. The side work you won’t quit. The shape of daily practice that doesn’t drain you. The culture you survive contact with as yourself. Chapter 2 made the hard nos machine-checkable. Those nos are circles: incomplete, load-bearing, yours. The clarity interview above is how you name more of them out loud.

The rest of the owl is not a mood board. It is ranked criteria with paragraph-length definitions of “great,” written cold, then pressure-tested by real postings. You will not finish that picture in one sitting. You will get sharper because the sweep keeps returning roles that almost fit, and you have to say why they almost fit. That is how the dream job stops being an adjective and becomes a set of paragraphs you can score twice the same way when you’re tired.

If the project in front of you feels impossible because no one handed you instructions, ask three questions:

What is the vague or daunting thing (the search, the next title, the “what do I want”)?
What circles are already clear, even if they’re basic?
What unknown are you treating as a reason not to draw at all?

My own longer attempt at drawing the owl lives here:
https://blog.codemarc.net/series/duck-to-dss

This book is the same move applied to the search. Dealbreakers and criteria are the circles. The rubric, the sweep, and the morning report are how you keep drawing without letting the market redraw you. Everything after the circles is the owl, and nobody else can finish it for you.

Rank before you refine

Here is the move most people skip: the order of the categories is itself a decision.

I ended up with twelve categories. Comp. Autonomy. Mutual commitment from leadership. Cultural fit. Side-project coexistence. Role shape. Work worth doing. Team. Industry. Geography. Visibility. Trajectory. Every one of them mattered. But not equally.

At 32, my top would have been role shape and trajectory. At 62, with a co-founder role I won’t quit, elderly parents, a kid still in college, and a retirement account under active repair, the top is different. The ranking answers “what would you trade what for, if you had to.” A comp floor beats a role you’d love. A leadership partner who trusts you beats an extra thirty thousand a year. Autonomy over your schedule beats title. Ranking forces you to admit those trades before a compelling posting invites you to un-admit them.

Rank cold. Not against a specific role. Cold ranking is durable; warm ranking is a rationalization.

Define “great” as a paragraph, not an adjective

“Great compensation” is not a criterion. It’s a wish. A criterion is:

Floor: $230K + bonus, or hourly equivalent $125–$150/hr at meaningful utilization. Target: $300K gross. Strong-yes: $350K+ for the twelve-month window when the retirement account needs replenishment. Cash > equity at this stage; equity welcome but non-substitutive. Healthcare currently through spouse: meaningful tiebreaker, not a gate.

That paragraph is machine-checkable in the same way a dealbreaker is, but it’s scored, not passed/failed. The rubric in the next chapter turns it into points; this chapter makes sure the paragraph exists in the first place.

The test for a criterion: could a tired version of you evaluate a role against it in thirty seconds and come out with the same answer twice? If the “great” definition needs interpretation every time you apply it, it’s still an adjective in a paragraph’s clothing.

A few patterns that work:

Adjective (useless)	Paragraph (usable)
Great comp	Floor $X, target $Y, strong-yes $Z, cash vs. equity preference, benefits stance
Great manager	Peer-in-the-foxhole vs. sponsor-from-above; matching work ethic; substantive vs. metric-only
Great culture	Founder-led vs. committee; overhead tolerance; specific cultural filters that would create friction
Great flexibility	Hours-vs-outcomes measurement; schedule autonomy specifics; disclosure requirements
Great work	The three-sentence description of the shape of daily practice you’re here to do

Every one of these is a paragraph because you’ll refer to them a hundred times over a search. Adjectives don’t survive that many applications. Paragraphs do.

A note for senior tech people

If you’re reading this past forty, or circling it (which in tech often starts earlier than anyone admits), this chapter is for you as much as anyone. The industry treats “senior” like a title ladder. Maureen Wiley Clough named the quieter truth with It Gets Late Early: the clock on how rooms read you starts ticking long before retirement, often right around forty, and the career playbook that got you here was written for a game that no longer pays out the same way.

I care specifically about the real relation between AI and experienced workers. Not the board-deck story that AI replaces veterans, and not the consolation story that AI is just autocomplete for juniors. Experience is judgment under pressure: pattern memory, known failure modes, taste about what not to automate. Used honestly, AI amplifies that. Used as cover for headcount theater, it discards it. Criteria written only for the young-hungry-mobile template will score the wrong roles and hide the trade-offs you actually face: caregiving, side work you won’t quit, cash over lottery equity, sustainability over a capstone.

Write your “great” paragraphs for the search you’re in now. The rest of this chapter is how.

The two-truths problem

Most criteria exercises assume you’re operating from one context: young, hungry, mobile, optimizing for upside. Real people have overlapping contexts. Late-career people especially.

My exercise had to hold two truths at once:

The pragmatic now: income needed to land within one to six months, a co-founder role I wouldn’t quit, care obligations that constrain schedule.
The longer arc: the shape of work I actually want to be doing, the culture that doesn’t wear me down, the trajectory I’m optimizing for at sixty-two (which is not the trajectory I optimized for at thirty-two).

A criterion that ignores either is useless. Ranking that pretends only one exists produces a search that either sells out or starves.

If your context has two truths (pragmatic and long-arc, or family and career, or health and ambition), write the criteria to hold both. The ranking should reflect which one wins when they conflict. That’s the trade-off you’re actually making.

Sustainability vs. capstone

One more meta-move that changed my criteria: I stopped optimizing for a capstone.

At forty, I’d have written criteria that pointed at a defining last act: a company, a book, a category role. At sixty-two, honestly examined, I’m optimizing for sustainability and replenishment, not ascension. That’s not a smaller ambition. It’s a different one. It’s the ambition of staying intact, staying solvent, staying in work I respect, and not being forced to make hard choices about my home or lifestyle.

Naming that changed my ranking. Comp moved up because stability matters more than upside. Autonomy moved up because life requires it. Capstone-adjacent categories (visibility, external recognition, board seats) moved down because I don’t need them to validate an identity I’ve already settled.

Your version of this question is not “capstone or sustainability?” It’s “what am I actually optimizing for right now, and does my ranking reflect it?” If your ranking is the ranking a career coach would produce for a generic senior candidate, it’s not yours.

How criteria feed the rubric

Every ranked criterion with a “great” paragraph maps into the next chapter cleanly:

The category becomes a scored dimension.
The priority order becomes weights. Not literally: a #1 category isn’t automatically 25 points and a #12 automatically 2, but #1 outweighs #12, and the gap is legible.
The “great” paragraph becomes the max-score anchor. A role that clears the paragraph earns the top of that dimension.

The rubric in chapter 4 is where this becomes numeric. This chapter just makes sure the ranking and the paragraphs exist. Without them, the rubric is a template. With them, it’s yours.

The worksheet

Fill this in before you touch the rubric. A sanitized copy lives in [templates/criteria.md](https://github.com/codemarc/i-built-my-own-recruiter/tree/main/templates/criteria.md). Replace the categories, the ranking, and the paragraphs with your own. The whole point is that generic criteria produce generic scoring.

#	Category	“Great” looks like (paragraph, not adjective)
1	your top-priority category	paragraph
2	next	paragraph
3	next	paragraph
…

Guidelines that kept mine from becoming a wish list:

Rank cold. Not against a role you already want. Not on a good day. Not on a bad day.
Paragraphs, not adjectives. If “great” needs interpretation every time, it isn’t defined yet.
Include the trade-offs. Which criterion wins when it conflicts with which: write that down.
Honor the two-truths problem if you have one. Pragmatic now and longer arc are not always aligned.
Revisit after every thin sweep (chapter 10). Criteria that produced no matches for three weeks aren’t sacred. They’re a hypothesis that failed.
Do the interview with friction, not compliance. A partner, a coach, or an AI told to push back: anyone whose job is to refuse the vague answer.

The nos protect you from lowering the floor. The criteria protect you from optimizing for the wrong ceiling. Chapter 4 turns the criteria into points.

Next: Building the Scoring Rubric →