Flags, star ratings, or color labels - which culling system should you use in Lightroom Classic?

By David · August 24, 2026 · 8 min read

Every new Lightroom Classic user gets the same advice: use flags, or use stars, or use color labels to sort your photos. What almost nobody explains is that these aren't three ways to do the same thing - they're three different data types solving three different problems, and using just one is why culling a real shoot always feels slower than it should. I build plugins that read and write Lightroom's catalog directly, including one that syncs client picks from a delivery gallery back into a photographer's catalog, and that round trip forced me to figure out exactly which of these three systems survives leaving Lightroom at all.

What do flags, star ratings, and color labels actually do in Lightroom Classic?

All three live in the same place - the Attribute row of the Filter bar and the flag/star/label icons under each thumbnail in the Filmstrip and Grid view - but they store fundamentally different kinds of data.

A flag is binary triage with a neutral middle state: a Pick is a Pick whether it's the best photo of the day or barely cleared the bar. A star rating is built for gradation, a 0-5 scale meant to express how good or how important a photo is relative to the others around it. A color label is neither - it has no built-in order at all, red isn't "worse" than green, it's just a tag whose meaning is whatever you decide under Metadata > Color Label Set.

The mistake most workflows make is treating these as interchangeable. Trying to do triage with stars (is a 2-star photo a reject or just mediocre?) or trying to grade quality with color labels (is blue better than yellow?) fights the tool instead of using it.

Why does a single rating system usually break down on a real shoot?

Pick a wedding, a 400-frame ceremony sequence, and watch what happens with only one axis. Stars-only culling stalls because you're forced to make a quality judgment on every frame, including the ones that are obviously out of focus or mid-blink - you end up hovering between 2 and 3 stars on borderline technical failures that should have taken half a second to reject. Flags-only culling loses information the other direction: once you've picked your 120 keepers out of 400, a flat list of Picks tells you nothing about which twelve are the hero shots you'll deliver first. Color-labels-only culling can't express "better than" at all, so you end up inventing an ad hoc ranking inside labels never meant to carry one.

The fix isn't a better single system - it's using each one for the job it's actually built for, in a specific order.

The two-pass workflow: Pick/Reject flags first, star ratings second

This is the workflow I've settled on after watching where single-system culling breaks down, and it maps directly onto what flags and stars are each good at.

1. Pass one, fast, flags only. Loupe view, keyboard only. Hit X on anything technically unusable - blur, blinks, unrecoverable exposure. Hit P on anything worth a second look. Leave the rest unflagged.
2. Filter to Picks. Use the Attribute filter bar to show only your Picks, so you're no longer spending attention on rejects.
3. Pass two, slower, stars only. Within the filtered Picks, grade 1 through 5: 1-2 stars for "include but not featured," 3 for "solid, main gallery," 4-5 for "hero shots, portfolio candidates."
4. Build Smart Collections off both fields. A rule like Flag is Pick AND Rating is greater than or equal to 4 gives you a highlight reel with no manual re-filtering - the same mechanism a Smart Publish Collection can automate, covered in our post on automating publishing with Smart Collections.

The key discipline is not skipping straight to stars. Rating a photo you're about to reject is wasted attention, and it's the single biggest reason a first cull drags into hours it shouldn't take.

Where color labels genuinely earn a place

Color labels get dismissed as a leftover from an older Lightroom era, but they solve a real problem: workflow states that don't fit on the pick/reject axis or the quality axis at all.

• Retouch flags. Tag a color (say, red, renamed to "Needs retouch") on any 5-star keeper that still needs a spot-removal or skin pass, so you can see at a glance which keepers aren't delivery-ready yet.
• Portfolio candidates. A label separate from your delivery rating lets you mark "possible portfolio addition" without disturbing the star rating driving the client's gallery ranking.
• Client special requests. If a client asks for "the one where Grandma is laughing," a distinct color keeps it findable without bumping its star rating and misrepresenting how you'd otherwise have ranked it.

What makes labels useful here is exactly what makes them a poor fit for grading: they carry no order, so they represent a pure category without implying one photo outranks another.

What survives once you publish or deliver: Smart Collections, publish filters, and client galleries

This becomes obvious once metadata has to leave Lightroom Classic and be read by something else - a Smart Publish Collection rule, an exported file another editor opens, or a client-facing delivery tool.

Star ratings write to a standard XMP field (xmp:Rating) with one unambiguous meaning: a number from 0 to 5. Any tool that reads XMP - another copy of Lightroom, Bridge, most DAM software - reads that number the same way you set it. Color labels also write to XMP, but the label-to-meaning mapping lives in your catalog's Color Label Set and doesn't travel with the file. Open the same photo in a colleague's catalog and the color is still there, but "red means needs retouch" isn't - to them it might just be Adobe's factory default of "select."

Pick and Reject flags are the least portable of the three. They're fundamentally a Lightroom catalog concept with no equivalent standard field most external tools read at all - a photo flagged Pick in Lightroom generally shows up as an ordinary, unflagged photo the moment it leaves your catalog.

I ran into this directly building Keptfolio, our Publish Service that turns a Lightroom collection into a hosted client gallery. When a client swipes through a delivery gallery and marks their favorites, that pick has to sync back into your catalog somehow. We write it back as a star rating, because it's the one axis that round-trips cleanly: one plain numeric field with a meaning every tool agrees on, so a client's picks land directly inside whatever Smart Collections you already built around ratings, with nothing to remap.

What this means for your own system: if you plan to automate anything downstream of culling - a Smart Publish Collection, an export preset filtered by rating - lean on star ratings for anything that needs to survive outside Lightroom, and treat flags and color labels as the fast, in-catalog triage tools they were designed to be.

Frequently asked questions

Should I use flags or star ratings first when culling a shoot?

Flags first. One fast pass with Pick (P) and Reject (X) separates technical keepers from failures - blur, blinks, bad exposure. Only once you have a clean set of Picks should you spend the slower attention star ratings need, grading 1 to 5 for delivery tier. Rating a frame you're about to reject wastes time.

What is the difference between a flag, a star rating, and a color label in Lightroom Classic?

A flag is a three-state triage field: Pick, Reject, or unflagged. A star rating is a 0 to 5 numeric field meant to grade quality or priority. A color label is one of five arbitrary tags (red, yellow, green, blue, purple) with no inherent order - its meaning is whatever you or your catalog's Color Label Set says it means. Flags answer keep-or-not, stars answer how good, and labels answer which bucket.

Can I rename what a color label means in Lightroom Classic?

Yes. Go to Metadata > Color Label Set > Edit and rename each of the five colors to whatever workflow states you use, like Retouch, Client Pick, or Uploaded. The catch is that this mapping lives in that one catalog. Open the same files in a different catalog or hand them to another editor and the color still shows, but the meaning you assigned does not travel with it.

Do flags, ratings, and color labels survive when I publish or export?

Star ratings write cleanly to a standard XMP field most software reads the same way. Color labels also write to XMP, but the label-to-meaning mapping does not travel with the file. Flags are the least portable of the three - a Lightroom catalog concept with no widely read export field, so most publish targets can't see flag state at all.

Why would a client gallery send picks back as star ratings instead of flags?

Because a star rating is one plain numeric field with an unambiguous meaning everywhere it is read, while a Lightroom flag has no standard external representation and a color label's meaning is only defined inside your own catalog. Writing a client's picks back as a star rating means they land in your existing Smart Collections and Library filters without you remapping anything.