How I built this platform twice.

2013 — A tool for myself

In 2013 I had a problem. I was selling digital goods online and nothing available did exactly what I needed. So I found someone brilliant enough to build it.

That's how Keysender started — a side project built by a talented developer I hired, solving my own selling needs. No investors, no grand vision. Just a tool that worked, built by someone who cared about getting it right. Then other sellers started asking to use it too.

I learned to code alongside him. Over the years, working next to someone that good, I went from knowing nothing about development to being able to build integrations myself. That education would later become the only thing keeping Keysender alive.

Eight years

Over eight years we grew Keysender into a working platform with live marketplace integrations, real customers, and real revenue. What started as a tool for myself became something others depended on.

The numbers tell the story better than I can. Starting from 91,000 orders in 2014, Keysender grew to processing 4.5 million orders a year at peak. By the time of the acquisition, the platform had processed over 15 million orders in total. Pure organic growth, no investors, no outside team — just a product that worked because the people building it were also using it every day.

Four integrations powered all of it: PayPal, eBay, Mercado, and Allegro. That was the entire platform when it was acquired.

2021 — The decision I'd regret

A company called ZeroOne eCommerce approached me with an acquisition offer. I'll call them "Zero" from here. They valued themselves at $2 million. They had a team, a platform, and a vision for the future of digital commerce.

And at first glance, they looked the part. Modern design, slick website, YouTube introduction videos. Built on React — latest technology. A design team, a sales team, a development team. They looked like the kind of company that could take Keysender to the next level. I believed them. I signed over my product, my IP, and eight years of work for a stake in "Zero".

The reality revealed itself quickly. After the acquisition, the design team was gone. The sales team was gone. The development team was gone. What remained was two founders, and the developer who had built Keysender with me from the start.

They proposed a $3,500 monthly salary. I turned it down. I knew that kind of fixed cost would eat into revenue, so I negotiated a percentage of revenue instead — with a $3,500 upper cap. I was confident enough in what I had built to bet my income on it. That cap was never reached.

Within months, their platform was scrapped. Not because they ran out of time — but because the architecture was so overcomplicated that developing anything on top of it was nearly impossible. After six months of trying, we shipped our first integration. It was hopeless. The $2 million company turned out to be worth exactly nothing.

What they did get as part of the acquisition — beyond Keysender itself — was something the developer had built before we ever signed anything. A mobile code scanner that let sellers point their camera at a code instead of retyping it manually, uploading directly to the database. At the time, scanning text from an image on mobile was genuinely ahead of its curve. Real innovation, solving a real problem for the exact customers we were serving.

"Zero" claimed it. Shelved it. Scrapped it. Never used, never maintained, never built upon. The same fate met the Keysender mobile app on Google Play — it died from neglect. He left not long after. I don't blame him.

The customers noticed too. During my time, all I needed to support 600 active users at peak was a Gmail inbox. Fast replies, fast fixes, problems solved. No support team, no ticketing system, no process. Just someone who cared.

The work nobody asked me to do

I had two choices: walk away or build. I built. I was hired as a product owner. Not a developer. But with a team that couldn't deliver, the choice was simple — build it myself or watch it die.

I could deliver a single integration in one to two weeks. Each one built by one person, outside their job description.

At first I took the fastest route available — a paid third-party integrator that charged per request. It wasn't good and it wasn't cheap, but it got things moving. "Zero" was satisfied. I wasn't. So I went back and rebuilt each integration from scratch using direct APIs — cleaner, faster, free of per-request fees. Some were built twice because I refused to leave poor quality in production. Nobody asked me to. Nobody noticed. I just couldn't leave it broken.

We lost Etsy. Not because the integration was technically difficult — but because of an insecure email exposure that got our developer account flagged. All we needed was to contact Etsy and request reapproval. A single email. It never got sent. A live integration, serving real customers, died because nobody followed up.

Then Shopify. They were rolling out a mandatory billing API requirement and our integration needed updating to comply. Their developer couldn't deliver it. Deadline passed. Then another. Rather than watch the integration break completely, I built a workaround myself — customers would provide their own Shopify developer keys as a temporary solution. It held for almost a year until the proper update was eventually done. A product owner, running a workaround for a year, because nobody else could fix it.

The deals we should have won

The incompetence wasn't just technical. It was commercial. Over four years we had real opportunities land on our table. Partnerships, integrations, enterprise deals — the kind of growth that should have transformed a small platform into something significant. Not a single one was closed properly.

During the acquisition they talked about an eBay partnership — our fraud screening and digital delivery capabilities were exactly what eBay needed to open digital goods to more sellers. It never happened.

PayPal was another open door. We had PayPal Smart Buttons functionality already built — and running on code I wrote in 2013 that they still use today. A formal partnership could have unlocked powerful new payment flows for thousands of sellers. Nobody pursued it.

We were even testing eSIM sales — a product I was personally selling on my own Allegro account, willing to hand over to the company the moment they set up their own seller account. They never managed to get their Allegro account verified. I kept selling as myself.

One stands out as particularly painful. We developed a supplier integration for a company — built it, connected their stock, got it working. Then, through sheer miscommunication and negligence, we started selling their inventory without authorization. Their own stock, listed and sold without their knowledge. That relationship ended immediately and badly.

Then there was the Spanish company. They came to us, believed in what we were building, paid $3,000 for a custom integration. What they received was so far below what was promised that they never used it. Not once.

At some point I started counting. Meetings to discuss what we should build. Meetings to discuss the meetings where we discussed what we should build. Retrospectives on sprints that delivered nothing. Every hour spent in a pointless call was an hour I could have spent shipping. A team that had perfected the art of motion without movement.

Four years. Hundreds of meetings. Ten integrations built by one person who wasn't hired to build them. Deals lost. Clients burned. A $2 million company that was worth nothing on arrival. I told myself one thing: if after four years a team is only moving forward because of me — I don't need the team. So I left.

Building the right way

I had now spent over a decade in digital commerce. I knew the digital fulfillment problem inside out — I had built Keysender from scratch in 2013, grown it, sold it, and then watched the buyers struggle to maintain what I had handed them.

But physical fulfillment was a different world entirely. Warehouses, couriers, pick-and-pack, inventory management across multiple channels — this was an order of magnitude more complex than sending a customer an email with a code. And complexity I don't understand is complexity I won't build well.

So I didn't open my laptop and start coding. The same approach had worked before. When I built the original digital delivery platform I didn't start with code either. I started selling digital goods manually — no platform, no automation, just the raw process. Once I understood exactly what needed to happen, exactly where the pain was, exactly what the solution had to look like — then I built it. Learn first. Build second.

So I started selling physical products myself. Four warehouse moves in a single year — not because I planned it that way, but because understanding the problem kept revealing a bigger problem worth solving.

Four moves in one year.

Today other companies don't just use SendItNow to manage their own warehouses. They ask me to store their stock and fulfill their orders for them. The software became a service became a business — because I learned the problem before I built the solution.

At some point the orders overtook the development. I was spending so much time picking, packing and shipping that I had no time left to build. So I hired someone. The developer who couldn't find time to code had become the warehouse operator who needed someone else to handle the floor while he got back to coding.

It's a slower approach. It's also why what I build actually works. Keysender was built the same way. Which is exactly why it still runs on my code — code I wrote starting in 2013 — with nothing new shipped in the year since I left.

Some people build slides. I build things. If you're selling physical goods across multiple channels and you're tired of tools built by people who've never touched a warehouse — you know where to find me.