Retail is the one stock niche where the most descriptive word available to you is usually the one you are forbidden to type. A shopping frame is full of logos, packaging and store signage, and both Adobe Stock and Shutterstock treat brand names in a title or keyword field as a reason to refuse the file. Keywording retail well means describing the act of shopping rather than the shop.
A retail frame is an intellectual property problem first
Adobe's intellectual property guidelines say two separate things that both land on retail work. The image side is familiar: remove or avoid recognizable logos, designs or artwork, and understand that trade dress counts, meaning distinctive shapes, colours and packaging, not only a printed mark. The metadata side is the one people miss. Do not mention brand names in keywords or titles, and do not use artist names or trademarks to increase search visibility. A frame cleaned of logos can still be refused on the strength of its keyword list.
The keyword guidance repeats it in plainer words: keep trademarks, brand names, and personal info out of your submission, and for titles specifically avoid references to intellectual property, artist names or real people. There is no editorial exemption hiding in that sentence for the keyword field.
Adobe's known restrictions list, which it describes as updated periodically and not all-inclusive, is worth reading once if you shoot retail at all. Among the entries that bear directly on shopping scenes: Build-A-Bear retail stores; public transit names, logos, signage and vehicle designs such as the MTA, BART and the London Underground; and distinctive trade dress including the Burberry check, the LV monogram and Tiffany blue on gift boxes. Specific packaging is named too, from the Absolut and Heineken bottle designs to Hershey's Kisses foil and the Duracell battery design. Grand Central Terminal appears when its signage is visible.
The literal layer: name the act, the place and the object
Shutterstock's metadata requirements define keywords as individual or compound words that describe the content's subject matter, and reject content when the metadata does not relate to what is depicted. So the first pass is inventory, not interpretation. What is physically in the frame: shopping trolley, wire basket, checkout counter, barcode scanner, card reader, receipt, paper bag, shelf, aisle, price tag, fitting room, queue.
Retail has an unusually strong vocabulary split between English variants, and it costs downloads if you only pick one side. Trolley and cart, till and checkout, queue and line, shop and store, shopfront and storefront are all live search terms depending on who is buying. A frame that says trolley but not cart is invisible to half the market. These are not synonyms stuffed in for volume, they are different buyers describing the same object, which is exactly the relevance test both agencies apply.
Name the role too, because retail buyers search by person as often as by object: shopper, customer, shop assistant, sales associate, cashier, store manager, delivery driver. And name the format of the retail space, since a buyer looking for a small independent shop will not take a hypermarket aisle: boutique, grocery store, supermarket, convenience store, market stall, pharmacy, pop-up shop, warehouse club.
The brand-shaped hole, and what goes into it
Once the brand name is off the table, the temptation is to leave the gap empty, and that is where retail portfolios lose their search surface. The substitute is a category word plus an attribute, which is usually what the buyer typed anyway. Nobody licensing a stock image searches for a specific soft drink brand; they search for canned drink, soda can, beverage aisle. The named handbag becomes leather handbag or designer-style tote. The named coffee chain becomes takeaway coffee cup or coffee shop counter.
There is a second-order version of this that gets files refused: describing the brand without naming it. Writing red and white cola can, or the three stripes, uses the keyword field to point at the trade dress the image is supposed to be free of. Adobe's wording covers it directly, since using trademarks to increase search visibility is listed alongside naming them. If the shot genuinely needs the mark to make sense, that is a signal about the shot, not the metadata.
The same supermarket frame, keyworded twice
A supermarket aisle frame keyworded two ways. The rejected version names brands and packaging, repeats keywords and adds unrelated commercial filler; the accepted version swaps each brand for a category word plus an attribute, covers both trolley and cart vocabulary, and keeps the title under 70 characters as Adobe recommends.
The conceptual layer buyers actually search
Retail imagery is bought almost entirely to illustrate an idea, which makes the conceptual layer worth more here than in most niches. The same supermarket frame can be licensed for cost of living, inflation, household budget, consumer confidence, supply chain, food security or impulse buying, and none of those words are visible in the picture. They are still legitimate, because the relevance test is whether the keyword describes the subject matter, and a person comparing prices on a shelf is subject matter about price comparison.
The line to hold is whether the concept is readable from the frame without a caption. Price comparison is readable if the shopper is holding two products. Inflation is readable if a price label is legible. Neither is readable from a wide empty aisle, and adding them there is the kind of reach that gets a whole keyword list treated as unreliable.
Retail also carries a commercial-vertical layer that other niches do not. The buyer of a checkout frame is frequently a payments company, a point-of-sale vendor, a logistics firm or a retail software business, and they search in their own vocabulary: contactless payment, point of sale, self-checkout, inventory, stock take, click and collect, last mile delivery, omnichannel retail, loyalty card. These terms pull a different and better-paying set of licences than shopping alone, and they cost nothing to include when the frame genuinely shows them.
A retail scene and a product shot are keyworded on different axes, which is easy to lose when the shoot produces both in one session. The scene sells on activity and setting; the isolated product sells on object, material and background. The approach for the second half of that output is covered separately in keywording product and still life stock photos.
How many keywords, and which ten carry the weight
Adobe accepts up to 49 keywords per submission and states that the first 10 positions carry the most weight in search ranking, so order is a ranking decision rather than a tidiness one. Its other three rules are short enough to memorise: use only relevant keywords to avoid content refusal, use each keyword once, and keep the title ideally under 70 characters. The Adobe keywording rules post works through how those interact on a full submission.
Shutterstock sets a minimum of 7 keywords and a maximum of 50, allows a description of up to 2048 characters, requires one category with a second optional, and treats location as optional but wants it specific down to neighbourhood, town or city. Retail frames benefit from that location field more than most, because a buyer sourcing a market stall in a particular region is searching geographically. The platform-specific detail sits in the Shutterstock keywording rules.
For a retail scene, the first ten are almost always the same shape. Two words for the setting, two for the object, two for the person, two for the action, two for the strongest readable concept. That ordering survives the trip between agencies, because the terms earning those slots are the ones a buyer would type regardless of which search box they are standing in front of.
Check a retail title and keyword set
Paste a title and keyword list to check it against the published rules used in this article: Adobe's limit of 49 keywords, its recommendation to keep titles ideally under 70 characters, the first-10-positions ranking weight, the use-each-keyword-once rule, and Shutterstock's floor of 7 keywords with a ceiling of 50.
The description is a factual sentence, not a pitch
Shutterstock defines the description or title as a factual sentence that describes the content's subject matter, and asks that it be free of spelling errors, grammatical errors, emojis and spam. Retail descriptions fail this more often than other niches because the subject invites sales language. Young woman enjoying a great shopping experience is not a factual sentence. Young woman comparing two cereal boxes in a supermarket aisle is.
Editorial submissions carry two extra description requirements at Shutterstock: the date the content was taken, with only the year or decade needed for anything shot before 2000, and the geographic location. That is a formatting rule rather than a creative one, and it is the single most common reason an otherwise sound editorial retail caption comes back.
Editorial is a different keyword set, not a looser one
Contributors often treat editorial as the amnesty route for a frame with a storefront in it, and partly it is: branding that cannot be removed generally has to travel as editorial, and the caption then carries the named place and date rather than hiding it. What editorial does not do is relax the commercial keyword rules in the other direction. An editorial file keyworded with commercial concept terms is a mismatch in the opposite sense, and it is still metadata that does not relate to the depicted subject matter. Identifiable people and identifiable private property keep their release requirements either way, which the guide to attaching releases to stock uploads sets out per agency.
What has to come out before you submit
Four categories of word do more damage than an absent keyword. Brand and retailer names, in the title as well as the list. Descriptions of trade dress that identify a brand without naming it. Repeated terms, including the near-duplicates that creep in when a list is assembled from two sources. And commercial filler such as stock photo, download, high quality, commercial use or copy-paste marketing words, which describe the transaction rather than the picture and are precisely what the relevance rules exclude.
One more is specific to retail and easy to miss on a batch: a legible price in the frame dates the image and localises it to a currency. That is not a metadata violation, but keywording it with evergreen concepts like household budget sends a file into searches where its visible price will look wrong within a year. Either keyword it to the moment, or reframe so the label is not readable.
A repeatable order for a retail shoot
Work the same sequence on every file and the first ten positions take care of themselves. Clear the frame of marks and packaging you cannot carry, or decide now that the file is editorial. Inventory what is physically there. Add both English variants for the objects that have them. Name the person and the format of the retail space. Add only the concepts a buyer could read from the frame without your help. Add the commercial-vertical terms the shot genuinely supports. Then order the list so the ten most searched terms sit at the front, cut it to the agency's ceiling, and write the title as one factual sentence under seventy characters.
At volume the hard part is not writing one good list, it is applying the same banned-term rules across a whole shoot and across agencies that each publish a different ceiling. Rastock AI reads the image, writes the title, description and keywords from what is in it, enforces per-agency banned and mandatory keyword lists so a brand name cannot survive into a submission, and delivers over FTP or SFTP to 10+ agencies with rollback if a batch goes out wrong. The comparison of retail workflow tools covers how the alternatives handle the same job, and the feature list has the detail.
No revenue share, no claim on your files, IPTC, CSV and XML export stay open, and the 14-day free trial needs no card. What no tool decides for you is whether a concept is readable from the frame, and that judgement is still where retail keywording is won.
Frequently asked questions
Can I use a shop or brand name in retail stock keywords?
No. Adobe's intellectual property guidelines say not to mention brand names in keywords or titles, and not to use trademarks to increase search visibility. Its keyword guidance repeats the instruction to keep trademarks and brand names out of the submission. Describing the trade dress instead, such as red and white can, falls under the same rule because it points at the mark without naming it.
How many keywords should a retail stock photo have?
Adobe accepts up to 49 per submission and Shutterstock requires a minimum of 7 with a maximum of 50, so a list in the twenties to low forties satisfies both. Order matters more than count at Adobe, which states the first 10 positions carry the most weight in search ranking. Padding past the point of relevance works against you, since only relevant keywords avoid content refusal.
Should I keyword both trolley and cart?
Yes, where both describe what is in the frame. They are not synonyms added for volume, they are different buyers searching for the same object in different English variants. The same applies to till and checkout, queue and line, shop and store, shopfront and storefront. Both terms pass the relevance test because both accurately name the depicted subject.
Do I need an editorial licence for a photo taken inside a store?
Often, yes. Branding that cannot be removed generally has to travel as editorial, and Adobe's known restrictions list names specific retail environments, including Build-A-Bear stores and public transit signage. Editorial is a reclassification, not an exemption: Shutterstock requires the date and geographic location in an editorial description, and identifiable people and private property keep their release requirements.
Which concept keywords are safe for a shopping scene?
The ones a buyer could read from the frame without a caption. Price comparison works if the shopper holds two products; inflation works if a price label is legible; cost of living, household budget, consumer confidence and impulse buying all work when the behaviour is visible. On an empty aisle none of them are readable, and adding them there is the kind of reach that makes a whole keyword list look unreliable.
Why do payments and logistics keywords belong on a retail photo?
Because the buyer frequently is not a retailer. Point-of-sale vendors, payment companies, logistics firms and retail software businesses license checkout and stockroom frames, and they search in their own vocabulary: contactless payment, self-checkout, point of sale, inventory, stock take, click and collect, last mile delivery. Those terms are legitimate whenever the frame genuinely shows what they name.