The box left on a Friday. It’s Tuesday now. Inside is a plant somebody searched for across months of sold-out listings, packed carefully by a woman who grew it in her spare bedroom, and it has spent four days in a truck in the dark.
We all know what’s in that box.
Nobody did anything wrong. The seller wrapped the root ball properly. The buyer paid promptly. The carrier did what carriers do over a long weekend, which is nothing. And now two people who have never met are about to have the worst conversation of their brief relationship, and one of them has never handled a customer complaint in her life.
That box is the whole business.
Plants break the assumptions ecommerce is built on
Start with the obvious: a plant is alive. It is also seasonal, and it is fragile in a way that nothing else you buy online is fragile. A dropped phone is a cracked phone. A dropped philodendron three days into transit is compost.
Then look at what commerce software assumes. Inventory is stable. A unit is a unit. Shipping is a solved problem someone else handles. All three are false here, simultaneously, every day.
I’ve built product in regulated healthcare, where being wrong is expensive and everybody knows it. Plants are harder. In healthcare the constraints are written down. Here they’re biological, seasonal, and they change depending on whether it froze in Ohio last night.
Demand was never the problem
We started Plantly during the pandemic, when plants were having a serious moment. People were home, staring at their walls, and buying monsteras with money they’d have spent on flights.
If you’d asked me before we launched what would break, I’d have said demand. I’d have been wrong. Demand showed up. Everything underneath it is what needed building.
The sellers are not businesses, and that is not a flaw
The biggest surprise had nothing to do with plants.
Plantly connects buyers with independent micro nurseries and side hustlers, and that phrase carries more weight than it looks like it does. Most of our sellers are people who love plants and found a way to make a little money from it. They are not businesses. They are not built to scale, and a lot of them don’t want to be.
Every marketplace platform on the market assumes the opposite. It assumes a merchant with inventory systems, someone who thinks in units and margins and has done customer service before. Build on that assumption and you ship a seller dashboard that a person with four hundred plants and a day job opens exactly once.
So the product question changes shape. You’re not building tools for merchants. You’re building scaffolding for enthusiasts — and it has to hold them up without demanding they become someone else first.
Chicken-and-egg is the famous problem. Credibility is the real one.
Everyone who’s looked at a two-sided marketplace knows the chicken-and-egg trap: no buyers without sellers, no sellers without buyers. It’s real and it’s hard, but it’s survivable. You seed a side, you subsidize, you grind.
Nobody warns you about credibility.
When someone buys a rare plant from an independent seller they’ve never heard of, you’re asking them to hand a hundred dollars to a stranger for a living thing that has to survive a journey. That seller has no brand. No reviews yet. No storefront anyone’s seen. They’re probably excellent — most are — and the buyer has no way to know it.
Which means the marketplace’s actual early job is lending credibility it hasn’t fully earned either. Every policy, every guarantee, every line of interface copy either builds that or spends it.
The version-one question was never which features to build
Back to the box.
Our seller now has to handle a situation that would test a trained support team. Most people have no framework for it. They freeze, or they apologize too much, or they go quiet — and quiet is the one that ends things. A buyer who gets a dead plant and silence never comes back, and tells people.
You cannot fix this with a policy page. You have to walk the seller through the moment: here’s what happens next, here’s what the buyer needs to hear first, here’s the part that isn’t your fault and here’s how to say so. The sellers who stayed are the ones somebody carried through their first disaster instead of linking to a help article.
That reframed what our first version was actually for. The temptation with any v1 is to ship what demos well — search, filters, a catalogue that photographs nicely. But nothing in that list determines whether a seller is still here in March. What determines it is what happens on their worst day, and we had to decide to build for the worst day before we’d had one.
The problem I still don’t have a clean answer for
Rare plants break the model, and I haven’t solved it.
Someone buys a specific rare plant. Not a category, not something similar — that one, from that seller, because they’ve been looking for months and it finally appeared. It dies in the truck, or the seller goes to the shelf and it isn’t there.
The uncomfortable fact is that the buyer purchased that particular plant. There’s no warehouse behind it. A micro nursery with one specimen of something unusual can’t restock on request, and won’t have another for a season if ever. Every standard remedy — refund, replace, substitute — is either inadequate or faintly insulting.
You can refund the money. You can’t refund the looking.
I don’t think that’s a bug in our implementation. I think it’s what happens when you take a model designed for interchangeable units and point it at things that are emphatically not interchangeable. Plants aren’t the only market like that. Anyone building in one of the others is hitting the same wall, and I’d genuinely like to know what they did about it.
Keno Vigil built Plantly. He writes about product and platform decisions at thinkkeno.com.
