Several times a year, I’ll don my helmut & headlamp, strap on the old leather harness, grab some ropes and carabiners, and rappel into the deep, dark archives of my photographic back catalog. There’s a treasure trove of material down there, just begging to be unearthed.
Lightroom makes these expeditions exceedingly effortless — as long as you’re clever enough to leave a trail of breadcrumbs each time you launch the program. My own breadcrumb trail — comprised of a carefully considered (though simple) keyword strategy and an intricately devised collection of flags, labels and stars — took nearly four years to construct, and I’ve been following it slavishly since 2010.
This article has very little to do with the art of photography and everything to do with the technical aspects of workflow and file management. As such, it’s really more of a “nerdicle” than an “article.” Fans of linguistic brevity might want to give this one a pass — but for anyone looking to improve their Lightroom cataloging, databasing and tagging methods, it might just be worth enduring all these words.
Overview of the ULTRAsomething Strategy
Most photographers tend to think of Lightroom as a raw processor and photo editing application. Which it is — but it’s also a powerful, configurable, photography-specific database program. And like most good database programs, it’s only as effective as the amount of work you put into designing and maintaining that database. As is usually the case, there’s a fine line between creating a database with too much data, and one with too little.
I started using Lightroom 10-years ago, early in its year-long extended beta development period. From 2006 – 2009, I designed, re-designed, and re-re-designed my databasing tactics dozens of times — constantly assessing exactly what worked and what didn’t. Some techniques proved too burdensome to maintain. Others were too confusing, not robust enough, or too inflexible to support different photographic directions. By early 2010, I’d finally perfected my cataloguing method, which I then applied retroactively to the entire database. I’ve adhered to it ever since.
For those photographers still struggling to find a cataloguing system that works for them, I’ll offer a glimpse into mine. Like most systems, it seems complicated when relegated to print. But when put to actual use, it’s so simple and un-intrusive that it’s nearly invisible. As with any database, success requires strict adherence to a set of cataloging rules. Below is a brief outline of the ULTRAsomething system components, each of which I’ll expand upon later in the nerdicle.
After trying all sorts of seemingly more beneficial folder hierarchies (such as grouping photos into location, subject or event-themed folders), I settled on the simplest and most consistent one imaginable: the calendar. Every photo I take gets filed into a folder that corresponds to the month it was shot. 12 monthly folders are filed in a yearly folder. 10 yearly photos are filed in a decade folder.
Before photos are filed into their permanent storage folders, they’re kept in a special holding folder, which I unimaginatively call “Unfiled.” This is the destination folder for everything that gets imported into Lightroom, and is where all photos sit until properly tagged with all the requisite metadata.
I don’t use any GPS location aids, so I don’t record exact coordinates or specific addresses for my photos. Instead, my location tagging technique involves simply filling in the Sublocation, City, State/Province and Country fields for every single photo.
Keywording is the most tedious part of the entire cataloging process, which is why I personally use a rather limited number of them. It’s essential to establish a definitive set of all possible keywords and then stick to it. A shorter keyword list means having less granularity when it’s time to search your library, but there are two substantial benefits: 1) It’s less onerous to tag photos with keywords when there are fewer words to select from (meaning you’re more likely to actually do it), and 2) when searching your library in the future, there is less ambiguity over which keywords you might or might not have used.
90% of my photos draw from the same subset of 50 keywords. Having only 50 commonly-used keywords makes the tagging process quite painless — particularly since I committed these words to memory years ago. Yet there’s still enough granularity that I’m able to quickly find specific photos — particularly when I add additional metadata fields to the search.
This is basically the only totally arbitrary value judgement to which I subject my photos. Using the P, X and U keys, I rate every photo as “good” (P), “Bad” (X), or “mediocre” (U). Flagging photos is a continuous process — it’s something I do when I view a photo for the first time, and it’s something I do with every subsequent viewing — for eternity. It’s essential that I constantly evaluate photos, because my own opinion of a photo’s merit changes over time. At least, I hope it does. Because without change, there can be no growth…
I don’t believe it’s possible to consistently grade the relative merits of one “good” photograph in comparison to other “good” photographs. So, although Adobe likely intended the star-rating tag as a way for users to grade their photos, I don’t actually use it for this purpose. Instead, I use the star-ratings to indicate a photo’s work-progress and/or publication status. As such, I only begin applying stars to photos once I’ve started to edit my photos, and only if I think a photo might be on the path to publication.
Like star-ratings, I also use these as status indicators, with different colors indicating different gates through which a photo has passed. Combined with star-ratings and flags, color labels let me create detailed and accurate searches without ever having to resort to making relative “value” judgements about photos.
With every photo in my library tagged with the metadata mentioned above, I make judicious use of Lightroom’s Smart Collections to arrange photos in all manner of useful groupings. For example, I can create a smart collection that says “show me all the photos I’ve taken in the last year that are on the road to possible publication but that haven’t yet been published.” Or I can say “show me all the abstract photos I took in 2011 that I flagged as “good,” but that haven’t yet been edited in any way.”
The remainder of this nerdicle will discuss each of these tactics in more detail. Anyone who’s struggling with an efficient way to organize their Lightroom catalogs might find value in studying my methods. Even if all my techniques don’t specifically apply to what you need to accomplish, you just might glean some new ideas or methods that you can adapt to your own system.
For example, a photo taken in June 2012 gets stored in a folder named “2012.06” along with every other photo shot that month. That folder (along with folders for the remaining eleven months) are stored in a year folder, named “2012.” Every yearly photo from 2010 through 2019 is stored in a decade folder, named “201x.”
There are generally only three exceptions to this rule:
1) Every year has a thirteenth folder, which is named “Unfiled.” This is the destination folder for everything that gets imported into Lightroom. It’s where every digital photo sits until it’s been keyworded, flagged, geotagged, and put away into its proper monthly folder.
2) The only exception to the “group by month” rule occurs when I shoot an assignment, in which case I create one or more client-specific folders. For example, if a client wants me to shoot an event, these photos will all go into an event-specific folder. When the event is over and the client is satisfied, I’ll export the entire folder (and its photos) as a stand-alone catalog and remove it from my master Lightroom catalog. My Lightroom catalog is for my photos — not a client’s.
3) Film photos are organized somewhat differently. Because I have film loaded into so many cameras at any one time, I rarely have any idea when any particular photo was actually taken. So, instead, I group scans from each film roll into their own folder, which I then group by month developed, rather than by month shot.
My choice to separate photos into monthly folders is based solely on the quantity I shoot. When I first built my Lightroom catalogs, I divided each year into four quarters (Jan-Mar, Apr-Jun, etc.). Ultimately, I realized each folder contained too many photos for me to visually digest, so I switched to a monthly grouping. Were I to suddenly start taking massive quantities of photos (which I hope never happens), I’d probably adopt a weekly folder structure in its place.
Some photographers prefer using folders to group photographs by event or by place. For example, some people might have folders for “Buster’s Birthday Party,” “Iowa Vacation” or “Donny & Marie Tribute Concert.” This has not worked for me, as it tends to overlay an unnecessarily arbitrary filing system over top of a better and more detailed version that already exists courtesy of the metadata. What if you vacation in Iowa more than once? Which Iowa vacation folder are you looking for? What if you go to several Donny & Marie tribute concerts? What if you attend a Donny & Marie tribute concert while vacationing in Iowa? In my experience, these are all questions that are best left to the metadata and not one’s filing structure.
Within a week of importing photos, I tag them with basic location data. Specifically, I fill in the fields for country, state/province, city, and sublocation.
None of my cameras have built-in GPS, nor do I bother using an external geo-tracker running in time-sync with the camera. I simply go by memory — which is precisely why this process needs to be completed ASAP. It’s easy for me to remember exactly where I took every photograph this past week, but nigh impossible to remember where I shot last week. Also, since each day’s photos were often shot at only one or two locations, it’s relatively simple to shift-click all the related photos, and apply the same geotag to all of them at once.
Because I’m not shooting for National Geographic and also because I tend to abhor “vacation” photos, location data is rarely a defining characteristic in my own photos. However, when searching for specific photos, I do have a tendency to remember exactly where I was when I shot it — so having this data available and searchable means it’s much easier for me to find a particular photo in the future.
I usually don’t add keywords until I move a photo from my “Unfiled” folder into its permanent monthly storage folder. Moving a photo to its proper place is a deliberate, one-time act — much like keywording. So I like to combine the two actions.
The whole keyword process is a slippery slope, and requires a lot of forethought and fine-tuning to get just right. If your list of possible keywords is too long and detailed, then keywording a photo becomes prohibitively oppressive. If your list of possible keywords is too short, then it can be difficult to find exactly what you’re looking for at some future date.
Unless you’re shooting for stock (does anyone still do that?) or a specific client, then each photographer should create his or her own custom list of possible keywords. And then stick to it. Religiously. My list cannot and should not be your list. Keywords exist to help you find your photos. My keywords won’t help you do that.
My own keyword list is somewhat sparse compared to convention. Basically, if I’m looking for a needle in a haystack, my keywords won’t pinpoint exactly where that needle is. But my keywords will point to a small section of that haystack where, with a tiny bit of manual searching, I will soon uncover that needle.
Obviously, subjects that appeal to me have a more nuanced list of possible keywords than subjects that don’t. For example, let’s look at the subject of birds. If I snap a photo of a bird, then I have exactly one keyword in my standard list to describe it — “bird.” It doesn’t matter if it’s an owl or a finch or a cormorant or a swallow — I’m tagging it with “bird.” It would be pointless for me to increase the granularity of my ornithology keywords because I’m neither an ornithologist nor a bird photographer. If I take a photo of a bird, the fact that it’s a bird is probably all that matters to me. So why spend all that extra time identifying and maintaining a list of bird species? 15 years from now, if I inexplicably need a photo of a kingfisher, then it’ll probably take me an extra minute or two to cull through my bird photos to find it. But for me — a most un-natural photographer — the amount of time I save by using only a single keyword to describe any bird will greatly surpass whatever I might have to spend searching for something in the future.
Wildlife photographers might scoff at such shiftlessness, but consider this: While my keyword list is woefully lacking in ornithological keywords, it does contain numerous keywords to describe different attributes of “abstract” photographs. And what I consider to be an “abstract” photo is what the average wildlife photographer would call a “reject.”
So the point is, we all have different photographic tendencies and requirements. And it’s up to each of us to decide the level of coarseness we need when setting up a keyword list. Basically, if you find yourself spending too much time and effort keywording each photo, then you’ve got too many keywords in your master list. Conversely, if you often have trouble locating photos of a particular type, then your master list probably doesn’t have enough keywords (or perhaps not the right keywords).
ULTRAsomething’s master keyword list contains fewer than 200 keywords (see note 1), of which only about 50 are required to describe 90% of my photos. As I’ve mentioned, this provides enough detail that a keyword search points me to the right section of the haystack without specifying exactly where the needle is. When combined with searches on additional metadata fields, I rarely need to spend more than a minute before I find that needle — and my Lightroom photo library goes back several decades!
(Note 1: I have an additional set of 100 keywords that are related solely to the technical aspects of film photography — words that enable me to identify film types, developers, agitation techniques, and all those sorts of things (lens, camera model, ISO speed) that would normally be embedded in a digital camera’s EXIF file, but that I must add manually to a scanned film photograph.)
Flags, Colors & Stars: A Comparison
Most photographers treat flags, color labels and star-ratings as if they were three different means to the same end: grading the relative “quality” of each photo.
- Flags: This is the simplest method of grading photos. With Lightroom, you can assign every photo a white flag (good), a black flag (bad) or no flag at all (neither good nor bad). Some may find this lack of nuance liberating; others may find it limiting.
- Star-Ratings: Many photographers like to assign star-ratings (from 0-stars to 5-stars) to their photos. When used in conjunction with flags, star-ratings allow you to subdivide your “good” photos into six levels of “goodness.”
- Color Labels: Photographers often use color labels — which provide a similar level of granularity to star ratings — to further gradate, rate and distinguish their photos from one another.
Similar to my keywording strategy, I tend to deviate from “normal” practice when it comes to using flags, colors and stars.
Specifically, I do not believe it’s possible to consistently grade each and every photo, nor is it possible to assess their relative value against one another. What strikes me as a “2-star” photo on Tuesday might seem more like a “3-star” on Saturday.
A similar problem with rating and ranking photos, is that we tend to grade our photos on a curve. For example, I might take a bunch of photos I think are 5-star. Tomorrow, I might take some photos that are even better! But Lightroom doesn’t have a 6-star option, so I’m forced to also rate the new photos as 5-star, even though I think they’re better. Maybe I’ll go back and demote the old 5-star photos to 4-star. Maybe I won’t. But if I do decide to downgrade some old 5-star photos, will I then go through my entire library and see if all the other 5-star photos need downgraded? And am I then going to see if the newly-downgraded 4-star photos are now better than those previously designated as 4-star? And do some of those previous 4’s need to be downgraded to 3’s? It’s a process that never ends. Within a year, you’ll inevitably have 1-star photos that should be 3-stars; 4-star photos that should be 2-stars; and 5’s that are not really better than many of your 2’s.
All that time spent grading photos; all that struggling over how they rank amongst one another… and for what? In the end, all you can say for sure is, “if a photo has any stars at all, then it’s probably a keeper.” At this point, you realize you could have just used flags, saved yourself a ton of time, and achieved the same result.
Now, because I said it’s impossible for me to grade photos with any granularity, you probably assume I use only flags, and don’t bother with color labels or star-ratings. You would assume wrong. In fact, I use all three methodologies to tag every single photo — and the reason I do this is not to increase granularity, but to compensate for the lack of it. Because I find it impossible to categorize a photo beyond the level of good, bad or mediocre (using flags), I use star ratings and color labels as status indicators, rather than value judgements.
What follows is a detailed description of how I use flags, colors and stars.
Flags are my most basic tagging technique. Every photo offers three choices:
- Good — “P” key (white flag)
- Bad — “X” key (black flag)
- Mediocre — no key (no flag). Note that the “U” key can be used to remove a “Good” or “Bad” flag from a previously tagged photo.
Any photo deemed worthy of “possible future exploration” gets a Good flag. Any photo not worthy of the hard drive space it occupies gets a Bad flag. Every mediocre photo (which is the vast majority of them) gets kept, but has no flag at all.
Because all new photos arrive in the Unfiled folder, and because I see the contents of this folder every single day, my flagging process is dynamic. I am constantly looking through this folder — analyzing it, absorbing it. Consequently, it’s not uncommon for me to reclassify bad photos as mediocre. Sometime mediocre gets promoted to good, and sometimes good gets demoted to mediocre. A photo will fly many different flags while it gestates in the Unfiled folder.
Note that I never actually delete a bad photo until its time to move photos to their permanent monthly storage folders. In general, a photo has to be pretty darn awful to get deleted — mostly because I’ve learned that whatever judgment I possess today might not match what I possess in the future.
Flagging is neither a chore nor a task — it’s simply something that happens thoughtlessly and automatically. Any time I look at a photo, I have an immediate response to it — it’s either good (P), bad (X) or mediocre (U). My fingers are constantly pressing one of these three keys every time I view a photo. It’s an automatic process.
I use color labels for certain very distinct (but ultimately unrelated) purposes. They’re essential for helping me find specific photos and for creating Smart Collections.
- Blue: Any time I edit a photo outside Lightroom (such as with Photoshop or a processing plugin), I apply a blue label to the edited photo, then group it with the original, uncolored RAW file in Lightroom. This allows me to quickly scan a collection and see which photos have had “outside” edits. Blue photos might eventually get tagged with a different color as their status changes, but blue is the first (and most basic) color label I apply.
- Red: Any time I publish a photo (blog, vBook, etc), I assign a red label to the published version.
- Yellow: Any time I publish a photo, I re-label all its preliminary edits with a yellow label. Usually, every published photo has several variations — editing “stages” that the photo endured between capture and publication. For example, my library might contain 4 or 5 versions of a single published film photo. One version might be the initial raw scan, while the second one might be that same scan after removing the dust and repairing the scratches in Photoshop. A third version might contain basic contrast adjustments, plus some dodging and burning. The fourth version might be the one I used for posting on the web, while a fifth version might contain one that’s been modified for printing. Yellow photos are always subservient to Red photos, meaning no photo will get a yellow label until some version of it has been published — at which point I’ll label the published version as red, and all grouped sub-versions as “yellow.”
- Purple: I use the purple label to tag my favorite photos of family and friends. These are the ones that get uploaded to Dropbox and saved to thumb drives — they’re important to me. Because these photos are rarely published (but are of equal or greater importance than the usual ULTRAsomething oeuvre), I need something other than flags or stars to indicate which are my favorites. That way my “personal” world doesn’t have to intersect with my “ULTRAsomething” world.
- Green: Like the purple photos, green is used to tag photos that have meaning only to me, but have no publication or business potential.
- No Color: Every photo starts its life without a color label, and most will rot in a monthly folder for all eternity without ever having one applied.
I use star-ratings as status indicators. Specifically:
- 1-Star: This means a photo is undergoing editing and, with further work, may one day be worthy of publishing.
This is not the same as the Blue label. Blue labels indicate that a photo has been processed through an external app — but just because a photo has gone through Photoshop, that doesn’t mean I currently consider it worthy of additional work or potential publication. So a photo can be “blue” but have no stars.
Nor is this the same as a white-flagged “good” photo. Flags are applied fairly liberally. If a photo interests me in any way, it gets flagged as “good,” but not every flagged photo is worthy of continuing effort. For example, I may shoot a subject from various angles, and ultimately flag three different versions as “good.” Odds are, however, that I’m only going to start editing one of those three picks — and the one I’ve chosen to edit is the one to which I give one star.
- 2-Stars: Two stars are for photos that I’ve worked on that are now finished and are good enough to publish, but that haven’t yet been published. Perhaps they’ll be published one day. Perhaps they won’t.
- 3-Stars: Three stars are for published photos. At first glance, this is the same as applying a Red label, since a photo needs to be published in order to get a third star and obtain a red label. However, there are 4-star and 5-star photos to consider…
- 4-Stars & 5-Stars: These are reserved for future designation. Currently, any photo that gets published gets a red label and 3-stars. But there are varying degrees of publication — Basically, if something gets published by ULTRAsomething (either via the website, a book, or a vBook), I consider it “published,” and thus a 3-star photo. But what if I end up getting my own show at the Museum of Modern Art? Or Steidl wants to make a book? Any such photos would surely supersede the value of those that had appeared only in an ULTRAsomething-based publication, and would thus get a higher star rating (while maintaining the red, “published” color). Unlikely? Probably. But what’s the point of all this photography without a few lofty goals to help drive me? Might as well build room for those goals into my system…
Searching the Catalog
The advantage of using so much metadata is that it’s easy to apply all sorts of nested boolean searches that would be difficult (if not impossible) were I to use only a few different data fields.
By standardizing what goes in each metadata field (and limiting the amount of data available to that field), I’m able to tag photos effortlessly, while maintaining the ability to search my library every which way I need.
And searching is, obviously, the entire point of maintaining a database in the first place. You may have 100,000 photos on your hard drive, spanning 25 years — but without a method to find the ones you want to see, you might as well not have any photos.
Lightroom lets you easily filter and sort your photo library in all manner of ways. Sometimes your need to find a particular photo is a one-time thing — “Show me all those photos I took of dogs at that Yaletown agility show in 2006.” Other times, you might need to filter and sort your library the same way again and again — “Show me all the photos I’ve taken in the last year that are on the road to possible publication, but that I haven’t yet published or finished editing.”
It’s this latter requirement — these sorting techniques that I tend to apply daily — that require extensive use of Adobe’s Smart Collections feature. Smart Collections let me create complex, nested boolean search and filtering algorithms, and then save those algorithms as a preset. Because of this feature, I’m never more than one-click away from seeing every flagged photo from 2011 that hasn’t yet been edited; or seeing every photo that I published in 2010; or… well… you get the idea.
Because I’ve adhered to this exact same methodology since 2010 (and because I eventually revamped all the earlier photos to conform to these techniques), it’s a system that’s thoroughly ingrained in my photography process. To recap, here’s what happens throughout the life of an ULTRAsomething photograph:
- Import photos into the Unfiled folder.
- Within a week, geotag them, and flag them as “good,” “bad” or “mediocre.”
- Every day, when new photos get added, look through all the other photos temporarily housed in the Unfiled folder, and re-flag them as necessary.
- After a photo has been in the unfiled folder for a month or two, add some keywords to it and move it to its proper monthly storage folder.
- Whenever a photo gets edited in an external application (like Photoshop or a plugin), apply a blue label to it and group it with the original raw file.
- Whenever a photo gets edited (either within Lightroom or with an external editor) and that photo seems destined to possible publication, assign it 1-star.
- After completing any and all post-processing, any photo deemed “publication worthy” (but that hasn’t yet been published), gets assigned a 2-star rating.
- When a photo gets published, assign it 3-stars, and label it red. Group any preliminary edits or alternate versions with the published photo and label them all yellow.
- Check my email to see if Gerhard Steidl, or anyone from MoMa or the Tate Modern has contacted me. If so, get ready to finally use those 4- and 5-star ratings.
- Remember that I have personal photos too. Be sure to tag any particularly good photos of family and friends with a purple label, and to tag any other subjectively important photos (personal documents? insurance photos?) with a green label. Don’t forget to export these as JPEGS so they can be backed up to thumb drives and uploaded to Dropbox.
Every time I open my Lightroom catalog, one or more of the previous steps needs to be performed. The process is simple, unvarying, and ceaseless. I consider it the equivalent of brushing my teeth — it’s something one does because it’s necessary, and because of this, I never give it a second thought.
I know this has been a particularly long and wordy nerdicle, and I would normally feel compelled to apologize for boring everyone silly. However, in this particular case, I suspect anyone who fell victim to ennui has long-stopped reading, and would thus not make it to the apology paragraph.
For those of you who did make it to the end, I suspect one of two things: either 1) you actually derived some useful information from it all (which is, of course, my sole reason for writing it), or 2) some sense of loyalty compelled you to read it, even though it didn’t apply to you (which means, as always, I owe you a massive “thank you” for indulging me).
©2016 grEGORy simpson
ABOUT THE PHOTOS:
”Light Spelunking” provides irrefutable evidence that any photo — no matter how bad it might be — is “publishable” if you have a strong enough reason to do so. This particular shot was taken over ten years ago with a Canon PowerShot S400 — a camera that was already hopelessly outdated by 2006. So why use it? Well, would you take one of your good cameras on a five hour cave descent that included having to rappel down a 7-story underground waterfall? Exactly. One must forgive the total lack of satisfying composition — photography in a cave involves pointing your camera at where you think your wife might be, then warning her to not look down, lest the flash might blind her.
”Insurance Policy” was photographed with a Leica M8 and a 28mm Summicron lens. It’s published here, not because it’s a great photo, but because it makes a point. As an exercise, I decided to empty my mind and think of a photo that I’d never published, and hadn’t seen in a very long time. Immediately I thought of a photo I’d taken of a fire hydrant wearing a motorcycle helmut. I remember this photo because, when I first saw it, I thought “Gee… that could have been so much better. I need to spend more time exploring alternate angles when I have the opportunity.” I’m rather certain I haven’t looked at this photo since then, so I figured it would be an ideal candidate for testing the effectiveness of my cataloging methods. How long would it take me to actually find this photo in my massive Lightroom library? I didn’t have much to go on. I remembered it was shot in Yaletown and… well… that was about all. I had no recollection which camera I might have used, nor could I remember exactly when it was taken — though I assumed it would have been at least 5 years ago. Because I have an immutable keyword strategy, I thought about how I would keyword this photo today, knowing I would have keyworded it exactly the same way back then. So I fired up Lightroom’s search filter and went to work. First, I went with my hunch that it was taken sometime prior to 2010. So I decided to first search only the contents of my 200x decade folder (which includes every photo taken between Jan 1, 2000 and Dec 31, 2009). I next told Lightroom to display only those photos shot in Yaletown, and I then started typing the keywords I would normally use to describe such a photo. By the time I’d finished typing the first keyword, my list of possibilities was down to only about a dozen images. I spotted the desired fire hydrant photo immediately. So how long did it take to find this needle in the haystack? Less than 10 seconds!
“Keywording Conundrum” was shot with a Leica M Monochrom (Type 246) and an undocumented (and un-remembered) lens. “Pre Google” was also shot with a Leica M Monochrom (Type 246), though the lens is now known: a 28mm Summicron. Both are obviously included for their vaguely relevant illustrative purposes.
“Polarity” was shot with a Ricoh GR, and has no illustrative purpose whatsoever. It was just one of those 2-star photos that I was looking for an excuse to publish…
REMINDER: If you find these photos enjoyable or the articles beneficial, please consider making a DONATION to this site’s continuing evolution. As you’ve likely realized, ULTRAsomething is not an aggregator site — serious time and effort go into developing the original content contained within these virtual walls.