A macro keyword set has to answer a question the photograph does not: what will the buyer call this? Most macro files are keyworded with what the photographer sees, which is the species, the lens and the magnification. The downloads tend to arrive on words like texture, detail, natural background and copy space.
Why the obvious macro words earn the least
Macro is the one stock category where the photographer's vocabulary and the buyer's vocabulary barely overlap. The person who made the picture knows it is a hoverfly at roughly 2:1 on a tube-extended 90mm, stacked from eleven frames. The person who licenses it is building a pharmaceutical brochure and needs a clean, shallow-focus close-up of something alive with space on the right for a headline.
Neither of them is wrong. But only one of them is typing into the search box, and a keyword set built entirely out of the first vocabulary will sit undiscovered next to weaker images that happened to use the second. This is not an argument for dropping accurate terms. It is an argument for the accurate terms being the floor of the set rather than the whole of it.
There is a second reason macro suffers more than other categories. A landscape photograph announces its own subject, so even a lazy keyword set lands somewhere useful. A frame-filling close-up of lichen is genuinely ambiguous: it is a plant, a texture, a background, a colour field and a science illustration all at once, and which of those it gets found as depends entirely on what you wrote.
The four layers a macro keyword set needs
Rather than brainstorming words, work through four questions in order. Each one produces a different kind of term, and a set that is missing a layer is invisible to a whole class of search.
Layer one: what it literally is
The subject, named at two or three levels of precision. For an insect: the species if you are confident, the common group, and the broad category. Hoverfly, fly, insect. For botanical work: the plant, the part, the category. Sunflower, stamen, flower. This layer is where accuracy lives and where a wrong guess does real damage, because a mislabelled species is a metadata error a buyer can catch.
Layer two: what it looks like
The visual properties of the frame, independent of subject. Close-up and extreme close-up. Texture, pattern, detail, surface. Shallow depth of field, bokeh, selective focus. The dominant colour, named plainly. Symmetry or repetition if the frame has it. This is the layer most macro contributors skip entirely, and it is the layer that designers search.
Layer three: what it will be used for
The job the image does in a layout. Natural background, wallpaper, banner, copy space, negative space. The concepts a macro frame carries without stating them: growth, fragility, detail, precision, purity, pollination, biodiversity, decay. A dew-covered leaf is a picture of a leaf and an illustration of freshness, and the second is what a skincare brand is searching for.
Layer four: the market it belongs to
The industry that licenses this kind of frame. Macro sells into science and education, agriculture, food, cosmetics, medical and environmental communication, and those are real search terms, not decoration. A close-up of a water droplet on skin belongs to cosmetics; the same technique on a crop leaf belongs to agriculture. One or two terms from this layer is enough, and they have to be honest, because irrelevant keywords are a named rejection reason at every major agency.
Species names: scientific, common, or both
This is the question macro contributors argue about most, and the answer is both, in a specific order. The common name goes in the high-value early positions; the Latin binomial goes later in the set.
The reasoning is about who searches which. A designer picking an image for a magazine spread types ladybird or ladybug. A textbook publisher, a university press or a specialist natural-history buyer types Coccinella septempunctata, and that is a small but high-intent audience with almost no competition on the term. You want both doors open, but only one of them has volume behind it.
Two cautions. Include the regional variants of the common name where they differ, because ladybird and ladybug are different searches in different markets. And if you are not certain of the species, do not write a binomial: a confident wrong Latin name is worse than an honest common one, and it is the kind of error that gets an image flagged by the one buyer who actually knows.
Order carries more weight than count
Adobe's own guidance for contributors is explicit on this point: place the most important and relevant keywords in the first 10 positions, because those have the greatest influence on search ranking. The field holds up to 49 keywords, each used once, and Adobe asks for only relevant keywords to avoid content refusal. The full set of field rules is in the Adobe Stock field reference, but the ranking consequence is what changes how you write a macro set.
For macro that means the first ten slots are a budget, not an inventory. A workable split is roughly three literal subject terms, three visual-property terms, three use or concept terms, and one market term. Everything else, including the Latin binomial, the equipment context and the long-tail variants, lives in positions eleven onwards where it can still be matched but is not competing for the ranking weight.
The common mistake is spending all ten on layer one. Hoverfly, fly, insect, bug, wings, eyes, antenna, compound eye, pollinator, nature is an accurate ten and a wasted ten, because every one of those words describes the same thing and competes against thousands of better-known frames of the same subject.
Order a macro keyword set
Drag a mixed macro keyword set into priority order. Adobe's contributor guidance places the most important and relevant keywords in the first 10 positions because those have the greatest influence on search ranking, with up to 49 keywords allowed and each used once.
Technique words: which ones buyers actually use
Macro contributors tend to put production vocabulary in the keyword field, and most of it does not belong there. The test is simple: would a person licensing an image type this word? Macro, close-up and extreme close-up pass, because buyers describe the look that way. Magnification ratios, focus stacking, tube extension, diffuser, rail and bellows all fail, because those describe how you made it and nobody shops by method.
Shallow depth of field, bokeh and selective focus are the interesting middle case. They are production terms that crossed over into buyer vocabulary, because a designer specifically wants a background that will not fight with overlaid type. Keep those. Drop the number that produced them.
One term worth adding deliberately is copy space, along with negative space. Macro frames are unusually good at providing a clean, out-of-focus area, and buyers searching for an image to put a headline on use those words directly. It takes one keyword slot and it is the difference between appearing in a general macro search and appearing in a layout search.
The filler-word trap
Macro attracts aesthetic adjectives, and they cost more than they return. Beautiful, stunning, amazing, incredible and gorgeous describe your opinion of the frame rather than its contents, and they are exactly what agency guidance means by irrelevant keywords. Adobe's documentation ties irrelevant keywords directly to content refusal, and Vecteezy names excessive repetition and irrelevance as a rejection reason in its own right. A keyword that cannot be verified by looking at the picture is a liability.
The related habit is padding to fill the cap. A 49-slot field does not need 49 words. A tight set of twenty-five verifiable terms outperforms forty-nine where the last twenty are singular and plural variants of the first ten, and on some agencies the padded version is refused outright.
Which tools actually help on macro files
Macro exposes a specific weakness in automatic keywording: vision models are good at visual properties and weak at species identification. A tool will reliably return close-up, insect, green, detail and shallow depth of field. It will guess at the genus and it will occasionally be confidently wrong. That asymmetry, rather than overall quality, is what should decide your choice. The wider tool landscape for this niche, including the stacking side, is covered in the macro tooling comparison.
Agency-native keyword suggestions
Adobe and several other agencies suggest keywords and categories inside their own contributor portals. Strongest point: free, and tuned to that agency's own taxonomy, so the suggestions will not violate its field rules. Main limitation: it works on one agency at a time and gives you nothing portable, so a contributor delivering to several destinations does the work repeatedly.
Lightroom or Bridge keyword hierarchies
Manual, but underrated for macro specifically. Strongest point: you build the controlled vocabulary yourself, so a species list you have verified once is reusable forever and never hallucinates. Macro portfolios are repetitive by nature, which suits a hierarchy. Main limitation: it is entirely your labour, there is no generation step, and the hierarchy only covers subjects you have already encountered.
PhotoTag.ai
A focused keyword and title generator. Strongest point: it does one job and the interface is simple enough to pick up in minutes, which suits someone who wants metadata and nothing else. Pricing follows a credit and subscription model. Main limitation: it stops at metadata, so delivery to the agencies remains a separate manual step.
Xpiks
A desktop tool with a long track record among high-volume contributors. Strongest point: it runs locally, so your originals never leave your machine, and the keyword editing and spellcheck tooling is genuinely built for bulk work on a controlled vocabulary. Licensed as desktop software rather than a monthly service. Main limitation: it expects you to know what you are doing, and the learning curve is real.
StockSubmitter
The long-standing multi-agency submitter. Strongest point: breadth of destinations and years of accumulated handling of each agency's quirks, which is not a small thing. Paid desktop software. Main limitation: the interface shows its age, and the metadata side is weaker than the delivery side, so you are still writing the keywords yourself.
Wirestock
A managed pipeline rather than a tool. Strongest point: it is genuinely hands-off, and for a contributor who dislikes the admin side that has real value. Main limitation: the commercial model takes a share of royalties, which compounds over a portfolio's lifetime, and you hand over control of how your metadata is written.
Rastock AI
Metadata generation and FTP or SFTP delivery to 10+ agencies in one pass, with each agency's field rules and banned or mandatory keyword lists applied on the way out, and a rollback if a batch goes wrong. Subscription, no revenue share, no ownership claim, and IPTC and CSV export are unrestricted. Main limitation on macro specifically is the same as every vision-based tool: it will not reliably identify a species, so the binomial is still yours to supply.
When to use which
If you shoot a narrow range of macro subjects and submit to one agency, build a Lightroom or Bridge keyword hierarchy and use the agency's own suggestions on top of it. The vocabulary problem solves itself after fifty files, and you have paid nothing.
If you want keywords generated but you are happy uploading manually, a dedicated generator like PhotoTag.ai is the smallest sensible purchase, and you supply the species names yourself.
If you are delivering the same macro batch to several agencies and the duplicated metadata work is the bottleneck, that is the case for a pipeline that generates once and reshapes per destination. Xpiks if you prefer local software and a one-off licence; Rastock AI if you want the delivery and the per-agency field rules handled in the same pass; StockSubmitter if destination breadth matters more than metadata quality to you.
If you would rather not touch the admin at all and accept a smaller share, a managed route like Wirestock is a legitimate choice. Run the arithmetic over the expected life of the portfolio rather than the first month, because a royalty share and a subscription cross over at a volume that is easy to misjudge.
Whichever you pick, one rule holds across all of them on macro work: never ship a generated species name you have not checked. Every other layer of the set can be automated with reasonable safety. Layer one cannot.
A pre-upload check for a macro batch
Run this over a batch before it goes out rather than over each file as you write it. Most macro metadata faults are batch-wide habits, not individual slips.
Macro keyword set pre-upload check
A pre-flight check for a macro batch covering the four keyword layers, the species-name decision, production vocabulary that does not belong in the field, and the agency rules on relevance, keyword order and the 49-slot limit.
Where the time actually goes
The four-layer method takes a few minutes per file the first time and almost no time once it is a habit, because macro portfolios repeat. What does not shrink with practice is reshaping the same set for each agency's field rules, and that is the part worth handing to software. Rastock.ai generates the set once, applies each destination's limits and banned or mandatory keyword lists, and delivers over FTP or SFTP to 10+ agencies with the batch status in one view. You can see how it works and what it costs; the 14-day trial does not ask for a card, there is no revenue share, and your IPTC and CSV exports are yours to take anywhere.
Frequently asked questions
How many keywords should a macro stock photo have?
Fewer than the cap, and all of them verifiable. Adobe's field holds up to 49 keywords and asks that each be used once and only relevant ones included, but a tight set of around twenty-five accurate terms usually outperforms a padded forty-nine. Padding with singular and plural variants of the same words reads as repetition, which is a named rejection reason at some agencies.
Should I use scientific names in macro keywords?
Yes, but place them later in the set. The common name belongs in the early high-weight positions because that is what designers type; the Latin binomial belongs from position eleven onwards, where it still matches the small, high-intent audience of textbook and natural-history buyers. Only write a binomial you have actually verified, since a confident wrong species name is a metadata error a buyer can catch.
Does keyword order matter on macro photos?
It matters more than keyword count. Adobe's contributor guidance states that the most important and relevant keywords should go in the first 10 positions because those have the greatest influence on search ranking. For macro that means spreading those ten across subject, visual property, use case and market rather than spending all ten on different words for the same subject.
Should focus stacking be in the keywords?
No. Focus stacking describes how the image was produced, and buyers do not shop by method, so it occupies a slot without returning anything. Keep the terms that crossed over into buyer vocabulary, such as shallow depth of field, bokeh and selective focus, because a designer genuinely searches for a background that will not fight with overlaid type.
Can an AI keywording tool identify the species in a macro photo?
Not reliably. Vision models are strong on visual properties such as close-up, texture, colour and depth of field, and weak on species identification, where they will guess at the genus and sometimes be confidently wrong. Treat the generated layers as a starting point and supply the species name yourself. It is the one layer of a macro keyword set that should not be automated.
Is copy space worth a keyword slot on a macro image?
Usually yes. Macro frames are unusually good at providing a clean out-of-focus area, and buyers looking for an image to put a headline on search copy space and negative space directly. One slot moves the file from general macro searches into layout searches. The condition is honesty: only add it when the frame actually has usable empty space.