What Percentage of Wedding Photos Should You Keep?

Why a keep percentage is the wrong benchmark

Keep-rate questions assume every raw take is made of similar material, but it is not. A raw take that is mostly single, deliberate frames has little redundancy to cut, so a high keep percentage there is normal. A raw take built from long bursts through the ceremony and reception is, by construction, full of near-duplicates that were never going to survive the cull, so a much lower percentage is equally normal for equally good work. Comparing the two percentages tells you about shooting style, not about culling quality.

What actually drives your personal keep rate

• Burst habits. Heavier burst shooters generate more frames per moment, which mechanically lowers the share that survive, since most of a burst exists to guarantee one good frame, not to all be kept.
• Personal quality standard. Some photographers cull tight editorial galleries; others favor generous, documentary-style delivery. Neither is more correct, and the difference alone can shift keep rate substantially.
• Coverage type. A deliberate portrait or formals session naturally keeps a higher share than a fast-moving, candid-heavy reception.
• Client expectations and contract. Some markets and couples expect leaner, highly curated galleries; others expect more generous ones.

A concrete example

Two photographers each shoot a similar eight-hour wedding. One works mostly in singles and doubles, rarely holding the shutter down; her raw take of 2,200 frames yields a gallery of around 900, a high share, because there was little redundancy to begin with. The other bursts heavily through every moving moment, ending up with 6,500 frames from an equivalent day, of which perhaps 800 survive, a much lower share, because most of those frames were insurance against blinks and timing, not distinct candidates. Their final galleries are similar in size and quality; only the percentage differs, and it differs for a mechanical reason that has nothing to do with either photographer's skill.

A better approach than chasing a percentage

Rather than aiming at a target percentage, aim at a process that surfaces every frame worth reviewing and lets you decide each one on its merits. A structured pass, technical rejects, then duplicates, then judgment, then coverage, produces a defensible final set without ever needing an arbitrary cutoff number. A review queue that routes genuinely uncertain frames to you, rather than forcing a blanket threshold across the whole take, keeps the decision on merit instead of on hitting a percentage.

Honest limitations

Even a good process will not tell you whether your keep rate is "right," because there is no external answer key; the only real check is whether your galleries consistently satisfy your clients and your own standards over time. Chasing someone else's percentage, high or low, risks either padding a gallery with weak frames or cutting frames a family would have treasured.

How Cull AI Studio handles this

Cull AI Studio does not target a keep percentage. It scores each frame locally on your computer for sharpness, exposure, faces, and eye state, groups bursts and near-duplicates so the strongest frame of each set surfaces, and routes anything below the confidence bar to a REVIEW queue instead of forcing every frame into KEEP or REJECT. Your corrections, including how generously or tightly you resolve REVIEW frames, are recorded with reasons and feed your personal model once about 1,000 reviewed decisions accumulate, so the software gradually reflects your own keep rate rather than applying someone else's.