Dropshipping Growth Operator

This is a custom GPT: ChatGPT running a set of instructions I wrote, published in OpenAI’s GPT directory. You need a ChatGPT account to open it. I don’t see your conversations, and I make nothing from it.

Open Dropshipping Growth Operator

What it does

You bring a product decision that’s stuck, and it takes the other side of it.

The instructions push it through five moves in order: state the assumption you’re making, name the number that would have to be true for the plan to work, name the cheapest way to test that number, name the failure you haven’t priced in, and only then give an opinion. On sourcing questions it asks about defect rate, lead-time variance and who answers when a shipment arrives wrong, because those are the costs that never appear in a unit-price comparison.

It earns its time on:

• A product you’re about to test, where you can describe the demand signal but haven’t written down what you’re assuming about it.
• Two or three supplier quotes that differ on more than price.
• A price you picked because it felt right, attached to a margin you haven’t worked all the way through.
• A product page that reads well and doesn’t convert, where you need someone to commit to which section is the problem.
• A list of twelve possible fixes that has to become an ordered list of three.

What it’s bad at

Anything that needs live data. It cannot see your store, your ad account, your supplier’s real lead times, current freight rates, or what is selling this month. If it produces a specific market figure, it produced it from nothing. Ask where the figure came from and it will usually admit it has no source.

Picking winners. It has no way to know whether a product will sell. It can tell you your reasoning has a hole in it. That is a different service and a smaller one.

Customs, tax, product safety and compliance. It will answer. Do not rely on the answer. Those questions need a person who is liable for being wrong.

Vague inputs. Give it “is this a good product” and a link, and you get generic ecommerce advice, because that is all the input can support. Give it your landed cost, your target price, your test budget and two real quotes and it starts being worth the session.

Being neutral about sourcing. I work for the company behind Yakkyofy, a sourcing platform. I wrote the instructions. I’ve kept the comparison criteria generic on purpose, but read anything it says about sourcing providers knowing where I work.

Who shouldn’t bother

If you already have real order volume and someone who does analysis, this will slow you down. It’s built for the stage before there’s data, when the only thing left to improve is the quality of the decision.

If you want a tool that tells you what to sell, this isn’t it, and I’d be wary of anything that claims to be.

If you don’t have a ChatGPT account and don’t want one, there’s nothing here for you. It only runs inside ChatGPT.

Things worth asking it

• “Here’s my landed cost, target price and test budget. What am I assuming that has to be true, and which assumption is most likely to be wrong?”
• “Three supplier quotes below. Rank them by the risk I’m not seeing, not by price.”
• “This product page. Which single section does the most damage, and why that one?”
• “Argue against testing this product. Give me the strongest case, not a balanced one.”
• “Twelve things I could fix, listed below. Cut it to three and defend the cut.”

The last two work better than they should. Asking a model to argue one side produces sharper output than asking it to be even-handed.

How to check its work

Treat every specific claim as unsourced until you’ve checked it. Its arguments about your reasoning are the product. Its facts about the world are not. When it gives you a number, ask where the number came from. When it gives you a verdict, ask what would change it.