How to train your eye for culling

Nobody gets better at picking photos simply by picking more photos. Practice works when it repeats a small task, gives you an answer quickly, and pushes just past comfortable, and an ordinary culling session does none of those things. Here are six drills that do, a schedule to run them, and how to tell whether your eye is genuinely improving.

Why culling ten thousand frames a season does not make you better at it

Practice only improves a skill under specific conditions. The task has to be small enough to repeat, the feedback has to arrive quickly enough to connect to the decision that caused it, and the difficulty has to sit slightly above comfortable.

A normal culling session fails all three. The task is enormous, 2,000 to 4,000 frames in a sitting. The feedback is nonexistent, because nobody tells you the gallery would have been better with different frames in it. And the difficulty sits below your level and repeats for four hours, which makes you numb rather than sharp.

So volume makes you faster at the mechanics and no better at the choices. That gap is why selection stays weak in photographers a decade into the job, and it is the argument in culling is a skill and nobody teaches it. The fix is to build the missing conditions on purpose, using material you already own. Everything below takes twenty minutes or less.

Drill one: the retro, and the answer key you already own

Start here. It is the only drill that scores you against reality rather than your own opinion.

You know things about your past galleries that no course could teach you: which frame the couple printed large, what went in the album, which image became a profile picture, which one the mother of the bride emailed about. That is ground truth about your own selection, sitting in your inbox and your print orders for free.

How to run it

Pick a gallery you delivered at least six months ago, long enough that you no longer remember the day frame by frame. Open the full take. Everything you did not pick is still on the drive, which is exactly why nothing should ever be thrown out. Set a timer for twenty minutes and pick it again cold, without opening what you sent. Then lay three sets side by side: what you delivered, what you picked today, and what the client used. Write down the disagreements and name them.

What you are looking for

Three patterns show up in almost everyone. Frames you now recognize as padding, near duplicates that survived because you could not choose. Frames you left out that you would now fight for, usually the softer emotional ones. And a gap between what you pick and what the client used, which is the most useful of the three because it is the only one that is not your taste arguing with itself. Run this monthly and it recalibrates everything else.

Five more drills, twenty minutes each

Each isolates one part of the decision. Run one at a time, not all in an evening.

Drill two: the sixty second set

Put thirty frames on screen from a shoot you have not delivered. Set a timer for sixty seconds and pick six. When the timer goes, the picks stand.

This trains commitment, which is what most photographers are missing. The rewind loop, forward, back, forward again, is not carefulness. It is a decision you keep declining to make. Under a timer you find your first instinct was usually right and the extra four minutes bought nothing.

Drill three: fit to screen only

Take a burst of six to ten frames. Judge them at fit to screen with zooming forbidden and pick the winner. Only then zoom to 100 percent and check whether your winner was also the sharpest.

Keep score over ten runs. Most people find their favorite frame and their sharpest frame are the same one well under half the time, which is a fast and personal demonstration of why your best photo is not the sharpest one. It also builds the habit of ranking first and verifying second, rather than letting a technical check make an editorial decision. The speed version is in choosing the best frame from a burst.

Drill four: the reverse pass

Instead of picking keepers from a set of fifty, name the ten you would cut first, and give a reason for each in four words or fewer. "Blinked." "Same as the last." "Nothing happening." "Hand in the way."

This is faster than picking, and it drags your real standards into the open in your own words. Collect those four word reasons over a few sessions and you have the first draft of a written standard, which is the job of building your own culling criteria.

Drill five: one frame per block

Take one section of a shoot, the prep, the ceremony, the speeches, and allow yourself exactly one frame. Not five. One.

This is uncomfortable and that is the point. It forces you to decide what the block was about before you decide which frame shows it, which is the hierarchy a real gallery needs and a flat pass never produces. It is the same muscle album work uses, where every page has to earn itself.

Drill six: somebody else's take

Cull a set you did not shoot: a second shooter's cards, a friend's session, or a trade with another photographer.

You have no memory of the day, so you cannot keep a mediocre frame because the moment was wonderful. You judge what is in the picture, which is all a client ever gets to do. Most photographers find they are harsher and faster on someone else's work, and that gap measures how much sentiment sits in a normal cull. With a second shooter it doubles as calibration, which is the point of culling second shooter photos.

A practice week that fits around real work

None of this survives as a burst of enthusiasm followed by nothing. Here is a rotation that adds up to under an hour a week.

Two rules make this survive a busy season. Never run a drill inside a production cull, because the goals conflict and the deadline wins. And never drill when you are too tired to do a real cull, for the reasons in why culling is so exhausting. Practice done while numb only teaches you to be numb faster.

Habits that turn every real cull into a practice rep

Drills are the gym. These make the job itself build the skill instead of only draining it.

Say the reason out loud, four words maximum. "Her face, right there." "Hands." "Only wide of the room." Speaking forces a real reason to exist. When the only sentence available is "it's sharp," you have caught yourself using the floor as a ranking.

Keep a one that got away note. At the end of every edit, write down the frame you nearly left out and ended up loving. One line, one file name. After ten jobs that list has an obvious pattern in it, and the pattern is your specific blind spot.

Record two numbers per job. Frames shot and frames picked. That ratio is your keeper rate, and it predicts your edit week. Track your own trend before comparing it to any published culling benchmark, because genre and shooting style move it more than skill does.

Never make the first and final pass in one sitting. Even a few hours restores the ability to see a set as a set rather than a queue.

Keep the conditions constant. Same screen, same brightness, roughly the same time of day. A dim room and a tired eye will change your keeper rate more than your taste will, and if the conditions move you cannot tell which one caused the change.

How to tell whether your eye is actually improving

You need signals you can read on your own work, with no benchmark from someone whose shoots look nothing like yours. Five that work:

Agreement between passes. What share of your first pass picks survive to the final gallery? As that climbs, your instinct is getting trustworthy and your second pass gets shorter.

Rewinds per run. Count how often you scroll back within one burst. This is the clearest single indicator, and it usually falls first.

Keeper rate stability. Not the number, the spread. Three similar weddings at 620, 640 and 610 picks mean the standard lives in your head reliably. 400, 900 and 550 mean it does not.

Time per thousand frames on your own gear. Track it over three shoots with a stopwatch, and compare it only to your own earlier number.

Client agreement. Of the frames the client chose for prints, the album or social, how many were in your own top fifty? It is the only external score you get, and it is free.

Expect the first two to move within a month and the last to take a season. Progress here is slow and durable rather than dramatic.

The practice that happens before you get home

A trained eye changes what you capture, not just what you keep. Each of these exists to make the later decision easier.

Shoot the release, not the build. A laugh builds, peaks and collapses into something softer that is usually the better photograph. Most photographers stop the burst at the peak. Hold two frames longer.

Recompose instead of repeating. Seven identical frames give you one option. Seven frames with a change of height, angle or focal length give you seven. The run should offer a real choice, not a coin flip.

Take one deliberate exit frame. On formals, keep shooting for two seconds after you say thank you. That is where the unmanaged expressions live, and in the review it is often the only frame in the setup that is about anybody.

Pre-pick in your head. After each block, name the frame you think will make the gallery. A voice memo works. Check yourself in the review. Being wrong is useful, being right saves you the block entirely.

Notice when you are shooting to feel safe. Frames taken because you are nervous, rather than because something is happening, are the ones that clog a cull later. The cure is not shooting less, it is knowing which of the two you are doing at the time.

Spend your reps on judgment, not on triage

There is a limited number of decisions in you on a Tuesday night, and the question is where they go.

Checking whether a frame is in focus is real work, but it does not make you better at anything. It has a right answer, and the thousandth repetition teaches you nothing the tenth did not. The decisions that improve your eye are the ones with no right answer, and most photographers reach them three hours in and out of judgment.

This is what we are building Kepla to change. It checks every frame from a shoot against clear criteria, focus, open eyes, expression, near identical neighbors, and hands you a shortlist. From a 2,000 photo wedding that is roughly 800 picks, and it takes minutes. The floor pass arrives done, so the attention you have left goes to the ranking, which is the half that responds to practice.

Nothing is ever deleted, moved or renamed. Kepla only marks what it picked. Every frame stays exactly where it is, one tap away, which is also what makes the retro drill possible five years from now: your practice material has to still exist for you to practice on it.

Straight status: the picking app is still being built for iPhone, iPad and Mac, with the first pass designed to happen on the drive home rather than at the desk at midnight. The booking page is live and free today. Founding members get the picking app the day it ships.

FAQ