My twins were seven years old when they wandered into my office, opened Lightroom, and deleted a folder of client proofs I’d delivered two days before delivery day. I wasn’t in the room. I heard the click. I came in, looked at the screen, and felt that specific cold dread that every photographer knows. Forty-three images, gone from the catalog. I had them restored from my backup drive in 90 seconds flat, and the only people who lost any sleep that night were the twins.

That story isn’t about luck. It’s about structure. A catalog that’s built correctly is a catalog that survives mistakes, software crashes, corrupted cards, and small children who don’t know what “client files” means. A catalog that’s thrown together by default settings and good intentions is a liability waiting to mature.

Why Your Catalog Is Doing More Work Than You Think

Most photographers treat their Lightroom catalog like a database of pictures. It’s actually a database of decisions. Every crop, every color grade, every keyword, every star rating, every export preset lives in that single .lrcat file. The raw files themselves sit untouched on your drives. What gets corrupted, bloated, or lost when things go wrong is the catalog, not the raws. That distinction matters enormously.

A standard Lightroom catalog file for a working photographer doing 30-40 shoots per year will run anywhere from 800MB to over 3GB depending on how many previews you’re storing inline. Smart Previews add roughly 1-2MB per image on average. If you’re shooting 2,000 images per job and building Smart Previews for everything, that math compounds fast. Knowing this helps you make intentional decisions about what you’re actually storing and where.

One Catalog vs. Many: The Decision That Shapes Everything

There are two schools here, and both are defensible. I run a single master catalog for all commercial work, and I’ve done it that way since 2011. The argument for multiple catalogs, one per client or per year, sounds organized until you need to search across three years of work to find a specific shot from a campaign you half-remember. A single catalog with disciplined folder structure and consistent keywording beats multiple catalogs with loose naming every time.

My folder structure inside Lightroom mirrors my hard drive exactly: Year, then Client Name, then Job Date and Job Name. A corporate headshot session from this spring looks like this on the drive: 2025 / Hartley_Commercial / 2025-04-11_Meridian-Realty-Headshots. Every folder gets three things before I move on: a color label (green for delivered, yellow for in progress, red for archived), a five-star minimum on selects, and a keyword set that includes client name, shoot type, and location. This takes about four minutes per job. It saves 40 minutes when a client calls a year later asking for a specific image.

Storage Architecture and Where the Catalog Lives

Here is something I see photographers get wrong constantly: they store their catalog on the same drive as their image files. Do not do this. Your catalog should live on your internal SSD or a dedicated external SSD with fast read/write speeds, 500MB/s minimum. Your raw files live on a separate high-capacity drive, 4TB to 8TB spinning drives are fine for storage. These are different jobs for different hardware.

I run a 1TB Samsung T7 for my active catalog and previews. My raws live on a Seagate 8TB desktop drive. Backups run to a second 8TB Seagate and a third offsite drive that rotates monthly. Lightroom’s built-in catalog backup is set to “Once a week, upon starting Lightroom,” which backs up to a folder on my backup drive automatically. That’s not optional. That’s baseline.

The catalog backup Lightroom creates is a compressed .lrcat file. I keep 12 weeks of these, then delete the oldest. At roughly 400MB per compressed backup, that’s about 5GB of catalog history sitting in rotation at any given time. That’s nothing. Keep them.

Keywording Is the Part Nobody Wants to Do

I know. It’s tedious. I didn’t keyword consistently for the first six years of my career and I have 80,000 images from that period that are essentially unsearchable. I’m not going to pretend that’s not a real cost I paid for being lazy.

The practical system I use now: every image that makes it past the cull gets three keyword tiers applied via a saved preset. Tier one is client name. Tier two is shoot category, portrait, product, event, editorial, architectural. Tier three is location, city and state minimum. That’s it. I’m not tagging emotional descriptors or color palettes. This isn’t Instagram. I need to find the image, not feel something about it. A saved keyword preset in Lightroom takes about 20 seconds to apply to an entire folder. There’s no excuse for skipping it.

For face tagging, I use Lightroom’s built-in People feature for client catalog searches but I don’t rely on it for anything mission-critical. It’s useful, not essential.

The Catalog Audit You Should Run Every Quarter

Four times a year I open my catalog and run three checks. First, I verify that every folder listed in Lightroom still shows a connected hard drive. Missing folders show with a question mark icon. Fix these immediately; don’t let them accumulate. Second, I run a smart collection filtered by “has no keywords” to catch anything that slipped through without tagging. Third, I back up and then optimize the catalog via the Catalog Settings dialog. Lightroom’s optimize function rewrites the database file cleanly and recovers space from deleted records. On a 2GB catalog this takes about three minutes and keeps query speeds fast.

If you do this quarterly, your catalog stays lean, your searches stay fast, and you never have that moment where you open Lightroom and wait 45 seconds for the interface to respond because the database has become a swamp.

The single most important thing I can tell you about catalog management is this: the structure you build in the first month of using any system is the structure you’ll be living with for a decade. Build it like it matters, because when something goes wrong at 11pm the night before a client delivery, it absolutely will.