Every marketplace shows sellers their numbers in its own dashboard. Orders, stock, reviews and shipments each live in a separate place. Sell on two marketplaces, and putting the full picture together is your job.
On May 18, the owner of LABÁ cosmetics came to Ardor with a simple question: can you get to the data in our seller account?
Four months later, the brand had a system of its own. Sales from both marketplaces sit in one set of reports. The warehouse has a tool that packs products into boxes and prints the labels. Review analysis runs with a single phrase. The owner didn't write code. They explained what the business needed, chose between options, and checked every result against their own numbers. Ardor did the rest.
Here's how it happened.
Day one: from a question to a database of their own
Ardor started by checking access and showing what the seller account could provide. Then it offered a few ways to use that data: one-off exports, regular collection, or reports on top. The owner replied with a very specific question:
“How many sales did SKU LC5 have yesterday?”
Then came the requests: connect to the marketplace, set up a database, add a reporting tool. Ardor had all three running within 28 minutes. The first reports and charts followed.
From then on, the owner didn't have to ask about one product at a time. They could see the whole business at once, and decide for themselves which reports they needed.
Data from both marketplaces has kept flowing into that database ever since. On the day we checked, four months in, it was still arriving.
Reports you change by asking
Once the data was in, the owner started shaping the reports. Every change took one message:
- Show orders as a chart by day
- Add a target line. Actually, take it out
- Use our SKUs instead of the marketplace's internal numbers
- Compare ads: impressions, clicks, spend, orders
- Put both marketplaces in one report
Some simple requests hid a lot of work. To show familiar SKUs instead of marketplace numbers, Ardor loaded the product catalog and matched one to the other. For the owner, it was one line in a chat.
By September there were 27 reports: the customer's path from viewing a product to ordering it, sales by product, each marketplace's share, product-group comparisons, and one sales trend across both marketplaces.
Review analysis in one phrase
Reviews tell you what customers love and what annoys them. Reading every one of them, every day, takes time.
On May 21, the owner asked Ardor to read the day's reviews and list five reasons customers were happy, five reasons they weren't, and five problems. Then they refined it: how many reviews came in, how many gave one, two or three stars, and what share was negative. They asked for one consistent format. And then they wrote:
“Let's turn the last four tasks into a clear skill you can run.”
Ardor saved it as a command. The next day, the owner asked for an analysis of yesterday's reviews and got a finished summary without explaining anything again.
The owner set the rules: what counts as negative and what belongs in the report. Ardor handles the rest. It pulls the reviews and writes the summary, every time.
The warehouse: from a product list to box labels
Before products ship to a marketplace warehouse, someone has to pack them into boxes. Every box gets a label: what's inside, how many, the expiry date, and which warehouse it's headed to.
In June, the owner asked Ardor to help with this. They explained the rules only someone who packs shipments would know: a blank quantity means zero, each shipment numbers its own boxes, and every label needs an expiry date and a warehouse.
The result is a working tool. Pick a shipment, and it pulls the contents from the seller account. Upload a product list, and it packs the products into boxes, prints the labels, and prepares a file to upload to the marketplace.
“Great, it works.”
That was the owner's verdict, followed right away by a request for new label fields.
Then the prototype grew into a real service, with its own page and sign-in for five employees. It now works with both marketplaces, and one button switches between them.
What to restock, and when
Sales reports tell you what happened. In July, the owner wanted to know what comes next: how long the current stock will last, and what needs restocking.
The owner asked Ardor to start with clarifying questions. Together they settled on average sales over 28 days. The new report shows, for each product, how many sell per day, how many are left, and how many days that stock will last.
None of it started from scratch. Ardor offered a choice: a new project or an extension of the existing one. The owner chose to build on what was there, so the forecast sits in the same database as the May reports.
And a lot more
In the same workspace, the owner took on work that had nothing to do with data:
- A presentation for colleagues
- Product images based on an attached photo
- A builder for product listing images
- A business model built from four audio recordings of a strategy discussion
- A financial model and a profit and loss statement
- A “Show account balances” command: every account's balance, grouped by legal entity
- Sales, returns and costs for every product in one summary
- One table of SKUs, barcodes and logistics details instead of several scattered ones
- A box-folding guide for packaging, with an animation
Every task builds on the last
On their own, these are small tasks. Together, they're one system.
The warehouse tool runs on the same marketplace connection Ardor built on day one. The stock forecast lives in the same database as the May reports. Each new task started from what was already there, not from zero. That's how the system grew along with the business.
Building software is now open to everyone
Not long ago, a system like this meant hiring a team: an analyst, a developer, and someone to keep it all running. First you wrote a spec. Then you waited for months.
The owner of LABÁ talked to Ardor in the language of their business: SKUs, boxes, expiry dates, days of stock. Ardor turned that into working tools, and before each important step it laid out the options: where to keep the data, when to build on what exists, and when to start something new. The owner made the calls.
The business got value on day one: an answer to its question, a database of its own and a reporting tool. After that, the system grew one message at a time.
You don't need to understand databases or code to do the same. You need to know your business and be able to describe the task in plain words. Ardor will do the rest.