Choose Your First Plant Accessories From the Orders You Already Ship

Most shops choose their first accessories from a supplier catalogue, a trade show, or the shelf by a competitor's register. All three are somebody else's answer to a question about somebody else's customers.

There is a better source and it is already sitting in your order history. Before you buy a single accessory line, your last couple of hundred orders will usually tell you what to stock, in what size, and in what quantity. Two hundred is not a researched threshold, just a workable starting point; if you ship less than that in a season, widen the window rather than shrinking the sample. The work is not gathering data. It is reading the data you already have in four specific places, and knowing what each place can and cannot tell you.

Read the combinations, not the items

The first place to look is what people bought together. Not what they bought, what they bought together.

Pull the last couple of hundred orders and write down every order that contained more than one item. Skip the orders that were a single plant. What you are looking for is repetition: the same two or three things appearing in the same basket often enough that it stops looking like coincidence.

Two things make this reading useful rather than misleading.

First, count orders, not units. If twenty orders each contained one small bag of potting mix, that is twenty signals. If one order contained twenty bags, that is one signal, and it may be a large order but it is not proof of a wholesale customer. Check the buyer and the repeat pattern before you label it. Counting units will tell you a big order exists. Counting orders tells you a pattern exists, and it is the pattern that belongs in your catalogue.

Second, separate the bundles you created from the bundles customers built. If you already sell a starter kit, every order of that kit reflects demand for the kit as a whole, and it does not by itself tell you whether anyone would buy those components separately. That is not the same as saying it tells you nothing, because a kit that sells well is telling you something about the plant and the pot together. What it cannot tell you is whether the saucer would sell on its own. The useful data is in orders where the customer added things themselves. Those baskets are the ones nobody planned, which is exactly why they are informative.

This is harder to do than it sounds, because the two can look the same in a flat export. Shopify's own Bundles app records the individual component SKUs on the order rather than the bundle's own SKU, and derives weight from the components. That does not mean the bundle's identity disappears. The platform groups bundle orders and provides bundle sales reporting, so the relationship is available. What happens is that a flat line-item export can drop it. Check what your export actually contains before you conclude you have lost the distinction.

If the export does not carry the bundle reference, note kit orders as they are placed, while you still know which is which. Without that, a season from now your own packaging decisions will be sitting in the same column as customer demand, and you will read the wrong one as a signal. If your export does carry the reference, you already have what you need and can skip the tagging.

What comes out of the exercise is usually unglamorous. A plant, a slightly bigger saucer than the one that came with it, and a bag of soil. That is not a product idea anyone would pitch. It is also a combination that repeats, and the usual reason is that the plant arrives in a pot too small for what the buyer intends to do with it, so the buyer is already planning the repot before the parcel lands. That explanation may not hold for your customers, but the pattern is worth checking against a few of those orders before you discount it.

Read the questions that arrive after delivery

The second place is your inbox, and specifically the messages that come in the first two weeks after a parcel lands.

These messages are the most honest product research a plant shop will ever get, because they are written by people who already own the plant and have just discovered what they need. They are not browsing. They are not speculating about what they might want some day. They have a specific problem and they are asking you to solve it.

Read them in bulk rather than one at a time. Sort the last six months into types. You will find that a small number of questions account for most of the volume, and that the same three or four questions keep arriving in slightly different words from people who have never met.

Two cautions apply here.

A question that arrives once is a person with a problem. The same question arriving repeatedly is a signal worth investigating, and it may point to a missing product, to a gap in your listing, or to instructions that are not being read. Most shops never count, because the messages arrive spread across months and feel like noise. There is no validated number of repeats that turns a question into a product decision, and fifteen is not a threshold anyone has established. Count, and then ask what the repetition means.

The second caution is about what these questions do and do not tell you. They tell you what to stock or what to fix. They do not tell you what to put in a cart at checkout. A buyer asking in week two how often to water a self-watering pot is telling you that the pot needs clearer guidance and possibly a refill product. That same buyer, at the moment of purchase, did not yet know they had that question. If you put the refill in front of them at checkout, they may not take it. If you stock it and put it in the follow-up email, some of them will. How many is something you find out by offering it, not by assuming the answer.

This is the distinction that separates a line that sells from a line that sits. Post-delivery questions choose the line. Checkout questions choose the add-on. Both come from the same source and they are not interchangeable.

Read the returns for the mismatch

The third place is the least pleasant and the most valuable: the returns and the complaints.

Most shops read a return as a cost. It is also a measurement. A return tells you that the buyer's expectation and the item diverged, and the reason field tells you which way.

Sort returns into three kinds, because they point to three different places to look. The reason field is a starting point, not a diagnosis, and it is often written by a customer who is describing a symptom.

A return because the item was damaged in transit points at packing, and it could also be the item. Look at the damage itself, the condition of the box, and whether the same item has arrived damaged before, before deciding which.

A return because the item was the wrong size or the wrong version points at how the versions were presented, and it could also be a picking error. Check the ordered SKU against the shipped SKU before you rewrite anything.

A return because the buyer did not need it points at customer selection, and it may equally point at a listing that described the buyer badly.

None of these three labels is a conclusion on its own, and a single return is not enough evidence to act on. What makes them useful is the count across a period and the cross-check against the order record.

Where returns do give a clear answer is in the contrast between the wrong-version returns and the not-needed returns. A run of returns for the wrong version says the item is being wanted and the presentation is failing. A run of returns for not-needed says the item is reaching people who were never going to use it, and stocking more of it will not improve either number. Damaged-in-transit returns tell you about packing, which is a different problem with a different fix. Getting the sorting right is what stops you applying one fix to another problem.

Ordered but not offered

The fourth place takes the least effort of all, and it is the only one where the buyer writes the product brief for you.

Somewhere in your inbox is a message asking whether you also sell something you do not sell. The rest of that sentence is a product request in the buyer's own words, from someone who was ready to pay that day and went elsewhere instead. Those messages are the ones to go and find.

This reading deserves its own place rather than being folded into the post-delivery questions, because it captures a different moment. Post-delivery questions come from people who already own the thing. These come from people who are still deciding whether to buy at all, and who would have taken the accessory with the plant that same afternoon if it existed. That is the easiest sale in the shop, and right now it is going to whoever happened to stock it.

The search itself takes an afternoon, and it does not work the way normal inbox searches do. These questions rarely announce themselves in a subject line. Search the body text for phrases like “do you sell”, “do you carry”, “do you stock”, “any chance you have”, and run each phrase separately so the results stay readable. Add “bigger” and “version” as single words, since people use them in different sentences.

Keep a running table as you go: the phrase the buyer used, the item they were asking for, the date, and whether they bought anything. That last column is the one that turns a request into evidence, because a request from someone who then ordered anyway is a stronger signal than a request from someone who disappeared.

Logging matters because of the same threshold that applies to post-delivery questions. One request is a person. The same request three times across a year is a gap in the range that nobody filled and everybody has been quietly working around.

What the four readings add up to

Put the four side by side and you get a shortlist with a reason behind each item, which is more than a catalogue can give you.

An accessory belongs on the shortlist when it appears in more than one source. It repeats in customer-built baskets and it answers a question that arrives after delivery. Or it answers a post-delivery question and the returns show people buying the wrong version rather than returning it unused. Overlap across two sources is a better signal than volume in one, but it is not a guarantee. The four sources are not independent: the same buyers appear in all of them, and a listing that is confusing produces mismatched baskets and wrong-version returns and post-delivery questions from the same confusion. Overlap can mean a real demand, or it can mean one bad listing showing up four times. Confirm with a small trial before you commit stock.

An accessory belongs on the list to think about later when it only shows up in one place, and the shape of that single appearance tells you what to do next. Something that keeps turning up in baskets people built themselves, but that nobody has ever emailed about, is usually a pair that is convenient rather than needed. Convenience pairs are often the first to stop selling the moment a competitor throws them in for free, though you will only know after it happens.

An accessory that only appears in returns, bought by mistake and never asked about, deserves a hard look, and it is not automatically a product to delete. It may be a real item with a broken listing. Fix the listing, measure again, and only then decide. That item is not proven to be an underperformer. It is a case where you do not yet know which of the two it is.

The wrong version is a product decision, not a service problem

One pattern deserves a closer look, because it costs more than its return rate suggests: items that get bought in the wrong version.

If buyers keep ordering a 4-inch saucer for a 6-inch pot, the tempting response is to absorb it, send the correct one, and treat the mismatch as the cost of doing business. That is generous to the customer and expensive to the shop, and it never touches the cause.

The expense is larger than the return postage, which is the part shops count. A wrong-version return may cost the return label, a second outbound shipment, and the packaging consumed twice, and which of those you actually pay depends on how the return is handled. Some shops refund and let the customer keep the item. Some pay return postage. Some send a replacement without a return. Work out what your own policy costs before you treat the number as fixed. Only the first of those shows up in a returns report. The rest appears in the reason nobody buys a second time, which you will never see in a metric.

The usual place to look is how the versions were presented. Two saucers that differ by two inches look nearly identical in separate photographs, each shot on its own against the same background at the same framing. A single photo showing both beside the pot removes the ambiguity that generates some of these returns, and it costs one afternoon.

If the wrong version keeps being ordered even after the listing is unambiguous, check the rest of the chain before you change the range. Look at the variant selector and which option is default, at what size basis the listing uses, and at what is actually being picked and shipped. A two-inch difference in nominal pot size does not tell you whether the two items serve different jobs, and a buyer who picks the wrong one has not demonstrated that the range is too fine. If the sizes do genuinely serve different plants, both belong on the shelf and the problem is upstream of the range.

Only after that review is worth asking whether two versions two inches apart are one product. Sometimes they are, and consolidating removes a source of error. Sometimes they are two genuinely different products that happen to photograph alike, and consolidating removes a sale.

Start with fewer items than you think

The most common first mistake is not choosing the wrong accessory. It is choosing eight of them.

Every accessory you stock consumes three things: cash tied up in stock, shelf or bin space, and the attention needed to keep the listing accurate. That last one is the one beginners underestimate, and the reason is simple. It does not appear on an invoice and it does not take up shelf space, so nothing ever prompts you to count it. It only shows up later, as a listing with an out-of-date photo and a size note that no longer matches what arrives in the box.

Two or three accessories, chosen by overlap across the four readings, will teach you more in a season than a broad range will teach you in a year. You will learn which ones actually attach, which ones generate questions, and which ones come back. That learning is the asset. The stock is not.

Costco is often cited as evidence that a narrow range works. Its own corporate description of the business rests on a limited selection of nationally branded and private-label products, which concentrates buying power and shelf attention. That is a different business with different customers and it does not transfer directly to a plant shop. What does transfer is the underlying logic: a narrow range concentrates purchasing and shelf attention, and both of those are scarce in a small shop. No SKU count, floor area comparison or sales figure is needed to borrow that much. Comparing a trial with mixed product sourcing is a more useful exercise than comparing yourself with a warehouse club.

Start from what you already ship

Your first accessory line should come from the overlap between things you already own: baskets customers built themselves, questions that keep arriving after delivery, mismatches that keep coming back as returns, and requests for items you do not sell.

None of these require a new tool, a consultant, or a subscription. They require one afternoon with your own order history, and the discipline to count orders rather than units, to check what your export actually contains before tagging anything, and to sort messages into types rather than answering them one at a time.

What you will almost certainly find is that the right first accessory is dull. It is the bigger saucer, the refill bag, the correct-size pot. Those do not make a good photograph and they do not impress anyone at a trade show. They do repeat in the data, and repetition in your own orders is worth more than a supplier's bestseller list, provided you confirm it with a small trial rather than reading a pattern as a verdict.

Turn your shortlist into a practical first order. Share the plant lines, pot sizes and two or three accessories your customers ask for, and explore a mixed order with Scarecrow Garden.