The decision took ten minutes. The handover took two months.
In early 2024 I looked at SmmPres's numbers and decided to get out. Nobody suggested it. The metrics did not demand it. The business had never been in better shape, which is the only reason the decision is worth writing about. Leaving something that is failing is housekeeping. Leaving something that has just started working is a decision.
SmmPres was a social media marketing panel. If you are not familiar with the SMM world: it is a marketplace where people buy social media engagement. Followers, likes, views, streams, comments. Influencers use these platforms. Small businesses use them. Marketing agencies use them. The product sits in a grey area of the internet, not quite illegal, not quite ethical, not something you would bring up at a dinner party unless you wanted the conversation to get uncomfortable.
We built it and launched into the Turkish market as a straightforward consumer product. Anyone could sign up, top up a balance, and buy.
That version of the business had a ceiling, and I hit it fast.
Panels compete almost entirely on price. Everyone buys from the same upstream providers, adds a margin, and sells. Users compare panels and pick the cheapest. A fraction of a percentage point is enough to lose a customer.
So I tried to be cheaper, and found the thing that actually shapes that market: I could not go below what my own supplier charged me. My floor was somebody else's price list. Every panel in the country was in the same position, all of us decorating the same wholesale rates with slightly different interfaces and fighting over the difference.
The margin was not in the shop. It was one layer up.
So I stopped trying to win the retail game and went after the supply instead. I picked the service where the money actually was, which was Spotify, and worked on becoming the upstream provider for it rather than a customer of one. Twitter, Instagram and Facebook followed.
That changed the economics completely, because it changed what I was selling and to whom. As a provider I was not competing with two hundred panels on price. I was the thing they bought from.
The customer changed with it.
I went after the biggest resellers in the market and turned them into my customers rather than my competitors. The panel stopped being a shop and became a supplier: business to business, and through them to everyone downstream.
The clearest signal of that shift was the minimum deposit. I set it high enough that an individual buying followers for their own account simply could not use the platform. That was the point. I was not accepting end users any more. If you could afford the minimum, you were a reseller or an agency, and you were buying to sell on.
It worked. At its peak the business was doing four figures a month in dollars, which in Turkish lira was a number that did not feel real to me at the time, and it was still climbing. That is the uncomfortable part of this post and the reason I am writing it. The strategy was correct, the repositioning was correct, and the business was in the healthiest state it had ever been in when I walked away from it.
Here is what it actually did, underneath all of that.
Someone buys ten thousand followers. Those followers are not real people who found an account and thought, "I enjoy this content." They are manufactured numbers. The account holder displays a bigger figure, brands see the bigger figure and assume organic popularity, and deals get signed on the basis of an audience that does not exist. On the music side the same mechanism moves streams, and streams are attached to real money and real chart positions.
The entire value chain rested on a fiction.
I am not going to pretend I agonised about this from the start. The market existed, the technology was straightforward, and the revenue validated the decision. I did not spend long considering what the product meant in a broader sense.
Going upstream is what forced the question. As a panel I was retailing something somebody else manufactured, and it is easy to tell yourself you are only the shopfront. As a provider I was the manufacturer. The supply that all those other businesses resold was mine. The fiction was not something I passed along any more; it was something I produced.
That is a harder thing to be relaxed about, and I did not stay relaxed about it.
There is a lesson in that sequence which I did not expect, and it is the opposite of the one I thought I was learning at the time.
I had assumed the problem was that the market had no moat. It was not. The moat was real and I built it: control the supply and you stop competing on price, because you set it. What I learned is that a moat is a multiplier and nothing else. It does not make a business worth running. It makes whatever the business already is harder to stop.
I had spent a year making something defensible without asking whether it deserved defending. The better the moat got, the more clearly that question stood out.
Now the part I would rather write differently than it happened.
I did not switch the supply off. I sold it. A Russian upstream provider bought the service operation, and I told my resellers what was happening so they could move their business across to them before anything went dark. The buyer got the machine. I kept the brand.
So the honest accounting is this: I stopped doing it, and it did not stop. The followers still get manufactured, the streams still move, the deals still get signed against audiences that do not exist. The only thing that changed is that the margin goes to somebody else now, and my name is not on it.
I have gone back and forth on whether that counts. The case for it is that the market was not mine to end, that a service with that many businesses depending on it does not vanish because one supplier develops a conscience, and that leaving my resellers stranded would have been a worse thing to do to people who had done nothing wrong. All of that is true and all of it is also exactly what you would say if you wanted the money and a clear conscience at the same time.
What I will not do is tell the version where I heroically shut it down, because I did not. I found the exit that cost me the least and I took it.
The thing nobody tells you about leaving a product is how good it feels afterwards.
Not immediately. Immediately it feels expensive and disruptive and mildly nauseating. But a week later the engineering time was going to work I actually cared about, and the weight of maintaining something I was not proud of was gone.
Every product you maintain is a commitment. Code needs updating, servers need monitoring, customers need supporting. When the product is something you believe in, that maintenance is fuel. When it is not, it is weight. SmmPres was weight, and it got heavier the better it did.
What it left behind is a question I now ask before starting anything: would I be comfortable describing this to a room full of people I respect? If the answer is not obviously yes, the project does not start. It is a cheap test to apply at the beginning and an expensive one to apply three years in, which is the whole point of applying it early.
I still hold the brand. I could bring it back tomorrow and I am fairly sure I never will. It sits there as a reminder that I was good at this, that being good at it was not the same as it being worth doing, and that I worked out the difference a year later than I should have.