Volume workflow
The difference between clearing a backlog and abandoning it is usually not effort. It is order of operations. This is the sequence that holds up at volume, the parts that get slower as you scale, and where a tool genuinely helps rather than just moving the work around.
The core mistake
Photograph one item, edit it, write the listing, publish it, then start the next. Every switch costs a setup change: camera position, lighting adjustment, background reset, mental context. At ten items it feels productive. At two hundred the switching cost is larger than the work.
Set the station once and photograph every item. Then do all the image work. Then do all the listing. Each stage is done with one setup held in mind, and the expensive parts — lighting, framing, decision-making about condition and price — happen once rather than two hundred times.
Concretely, this is the order that works: triage, station, shoot everything, capture details while shooting, then process and list in batches. The rest of this page is those steps in detail.
Step 1
Decide what is worth listing before you spend time photographing it. Sorting first is the single largest saving at volume, because it removes items from every later stage.
List now, list later, and do not list. Be decisive. The "later" pile is where backlogs come from, so keep it small and genuine — items needing cleaning or a missing part, not items you feel ambivalent about.
Books together, clothing together, small hard goods together. Similar items need the same framing, the same lighting and often similar listing fields, so grouping removes per-item decisions from the shoot itself.
Shoot the easiest and most homogeneous items first. They build the rhythm and give you a realistic rate per hour before you reach the awkward items that will slow you down.
Expect triage to take roughly one to two hours per hundred items. That sounds like a delay. It is the opposite: it is the cheapest hour you will spend, because it stops you photographing things you will never list.
Step 2
The goal is that adding the next item takes one movement and no decisions. Anything you adjust between items is a tax paid two hundred times.
A perfectly lit photograph you have to recreate every time costs more than a good-enough photograph you can recreate instantly. Consistent lighting also makes background removal behave predictably, because the separation between item and background stays the same across the batch.
If the item does not fit the station, that is a signal to handle it separately rather than to rebuild the station mid-batch.
Step 3
The most common cause of a stalled backlog is going back. If you finish photographing and then discover you cannot remember which item had the scratch, or what the measurements were, you have to find the item again and the batch breaks.
Flaws, wear, missing parts, anything a buyer would ask about. Photograph the flaw as its own frame rather than trying to describe it later from memory.
Labels, makers' marks, model numbers, size tags and barcodes. Scanning a barcode takes a second while the item is in your hand and saves re-identification entirely when you list.
Do all measurements while the item is on the station, not during listing. Group items with similar dimensions — clothing in sizes, books as a stack — so the ruler comes out once.
A practical rhythm is three to five frames per ordinary item: a main shot, one or two detail shots, any flaw, and the identifier. Homogeneous items can drop to two. Complex or high-value items will need more and should be shot separately.
Step 4
Keep these two stages apart. Image work is visual and fast once you are in the rhythm; listing is administrative and uses different attention. Switching between them per item is what makes an evening disappear.
Crop, straighten and remove backgrounds across the whole batch. If you are removing backgrounds by hand, this stage is where volume becomes unmanageable — it is the one step that does not get faster with practice, only more tedious.
Titles, descriptions and the fields each marketplace requires. At this stage the item details you captured during shooting are the input, and the work is review and correction rather than transcription.
Where the time goes is worth stating plainly, because it determines what to optimise. For most sellers at volume:
| Stage | Rough share of total time | Does it get faster with practice? |
|---|---|---|
| Triage and sorting | A small one-off cost | Yes, with clearer rules |
| Photography | Moderate | Yes, substantially |
| Background removal and image edits | The largest single block | Barely. It is mechanical work |
| Listing copy and fields | Substantial | Somewhat, with consistent categories |
| Publishing and correcting | Small but disruptive | Yes, with fewer destinations per item |
These are order-of-magnitude shares from a volume workflow, not benchmarks from a controlled study. Treat them as a prompt to measure your own split rather than as targets.
Where it breaks at scale
Manual masking is linear: twice the items, twice the hours. This is the clearest case for automating the step, because the output is objectively checkable — either the background is clean or it is not.
Titles, descriptions and category fields are mostly transcription and pattern, which is exactly what software handles well and humans handle poorly at volume. The judgement that still matters — price, condition, whether a claim is safe — stays with you.
Product Agent covers exactly these two stages on Android: it removes backgrounds server-side and drafts the title, description, price, condition and category from the item evidence, then holds every listing for your review before it goes to a store. It connects to eBay, Shopify, WooCommerce, Square and Wix, and batch runs process up to 100 items on the paid plans. Etsy support is coming soon; timing and availability may change.
Three limits worth knowing before you rely on it: it does not sync stock or inventory between stores, so overselling remains your own problem to solve. It is Android only, with no desktop version. And every listing is reviewable rather than published automatically, which is deliberate but is still work you have to do.
A realistic first day
Sort into piles, group by similarity, build the station. You will photograph very little. Resist the urge to start shooting before the station is fixed.
Once the station is set, shooting a whole batch in one sitting is much faster than it sounds. Stop when your framing starts drifting, because inconsistent photos cost more later than they save now.
All image work in one pass, then all listing in another. Review as you go, and deliver in groups rather than one at a time so a failure does not cost you the batch.
If the first day overruns, the usual cause is skipping triage or rebuilding the station repeatedly. If it finishes early, you have probably listed items you should have set aside — which is a fine problem to have.
The user guide walks through capture, review and delivery step by step, so you can judge whether the process fits how you already work.
Read the user guide