Dear readers,

This Wednesday we look at the less discussed side of building a startup: the moments when things break, money runs out, teams disappear and founders have to decide whether to keep going.

Our main story follows Ervin Kalemi and Publer, an Albanian startup that came close to dying several times. From losing customers after Facebook’s Cambridge Analytica crackdown to taking on debt, betting on AppSumo, and absorbing a hefty monthly Twitter API bill among other challenges, our featured story is about constantly rebuilding when the previous version stops working.

Elsewhere, we also look at the uncomfortable gap between the founder mythology and the employee reality, and why the relentless commitment celebrated in startup stories can sometimes mean very different things for the people working underneath it.

Enjoy!
Bojan Stojkovski
Editor-in-Chief, IT Logs

There is a tendency to tell startup stories backwards. We start with the revenue number, the growing team, the product, the users and the milestones, and then work our way back to the founder, as if success was always the logical conclusion. The story of Ervin Kalemi and Publer is considerably messier than that.

Publer went from a small tool built for a family member's travel agency to a global social media management platform on track to surpass $3 million in annual recurring revenue. During that journey, Kalemi nearly killed the company himself, then watched external events threaten to finish the job. 

There were health problems, family tragedy, debt, lost customers, failed fundraising attempts, a team that disappeared and a costly lesson about what happens when a bootstrapped startup starts behaving like a corporation. It is, in many ways, less a story about startup success than one about survival.

Another life before Publer

Kalemi's own account of the journey begins long before Publer. He grew up with a computer at home in 2000, but says he was not the stereotypical child who immediately started taking hardware apart or learning how software worked. Instead, he designed covers for pirated video games and created elaborate PowerPoint presentations, including one that functioned almost like a website.

His path into technology was also accidental. He initially enrolled at the University of Texas at Austin to study advertising, believing his interest in graphic design would make it a natural fit. 

Then, during his second semester, he took an introductory programming class, and switched to computer science. The transition was difficult, as many of his classmates had been coding since high school, while Kalemi was only starting at 19. But persistence mattered more than being the best student in the room, and that eventually produced his first experiments with social media.

In 2011, Kalemi volunteered to build the frontend for a university class project and began experimenting with Facebook Page tabs and applications. By graduation, he had created two Facebook apps. One, Suggest to Friends, allowed visitors to invite all their friends to like a Facebook Page with a single click. It went viral and became a source of revenue through advertising and paid plans. Another, MemeGen, allowed users to create memes and publish them to Facebook.

Then came the idea that would become Publer. Kalemi's brother, Marsel, had opened a small travel agency and needed a way to schedule posts for its Facebook page. Kalemi built one. Publer was born over Thanksgiving weekend in 2012, using PHP, MySQL, HTML, CSS and a little jQuery.

The first version was extremely simple - users could log in with Facebook and schedule text posts. There was no sophisticated dashboard and not even a way to see what had been scheduled. Users essentially had to trust that the system would publish their posts.

Six months after introducing a paid plan, the first paying customer arrived. Kalemi was still working full-time as a developer, so Publer became a second job. He would come home from work and continue building the product, often until two or three in the morning, before getting a few hours of sleep and doing it again.

At 25, a LinkedIn birthday message reminded him of an entrepreneurship class and reignited the idea that he had abandoned. Kalemi took it as the sign he had been waiting for and bought a one-way ticket back to Albania.

With around $60K saved, most of it generated by his earlier apps, he moved back in with his parents and committed himself to Publer, in what proved to be a first real reset.

He built a proper website, started a blog and launched an ambassador program that would eventually become one of Publer's strongest organic acquisition channels. He also studied competitors and identified features they were missing, adding bulk scheduling, CSV imports, auto-scheduling, watermarking and signatures.

By 2018, Publer had hundreds of paying customers and was generating between $2,500 and $3,500 a month. Then Kalemi decided to rebuild the entire product from scratch, since the original code had become what he describes as "spaghetti code", so Publer was rewritten using Ruby on Rails, React and MongoDB/PostgreSQL.

The technical rebuild was painful, but what came next was far worse.

When everything started falling apart

One day, Kalemi saw police outside his father's currency exchange business. A major fraud had resulted in €140K being stolen. The money did not belong to the family, but Kalemi gave his savings to his father to help. 

At almost exactly the same time, Facebook's Cambridge Analytica scandal triggered a manual review of third-party applications, and as a result, Publer's group-posting functionality was suspended for six months.

The consequences were devastating as Publer lost roughly half of its paying customers to competitors. Even after Facebook restored the functionality, many of those customers did not return. 

The team that had spent years rebuilding Publer eventually disappeared and Kalemi was alone again. Fundraising attempts failed, and pitch competitions ended in rejection. He even began considering getting another job and moving back to the US.

It was at this point that the difference between a founder and an entrepreneur became particularly important. Kalemi had started Publer as a technical founder, but if the company was going to survive, he had to become something else.

He started thinking about marketing, customer retention, product development and leadership. He began building his personal brand on LinkedIn, despite criticism and jokes about his accent. He appeared on camera after launching new features and effectively became the public face of the company. It worked, and then he borrowed €50K to hire again.

The AppSumo bet

The next major turning point came in 2020, when AppSumo approached Publer. Kalemi had actually contacted the company a year earlier and received no response. The difference this time was the collection of customer reviews Publer had accumulated, including reviews encouraged through discounts.

Through AppSumo, Publer offered its highest-tier plan for a one-time $39 payment. On paper, it looked like a terrible SaaS deal, however instead, it became one of the company's most important moments.

The campaign ran for three months, followed by another limited lifetime offer around Black Friday. Together, the campaigns generated around $250K in profit and, more importantly, introduced Publer to a much larger audience.

Kalemi reinvested the money into the company - Publer expanded its product, rebuilt its website and dashboard, added integrations with Instagram, Pinterest, YouTube, WordPress and Telegram, and introduced features such as calendar views, media libraries, link-in-bio functionality, browser extensions and mobile applications.

The company eventually reached breakeven in August 2021, and by the end of 2022, Publer was approaching $1 million in ARR. Kalemi was finally debt-free and paying himself a proper salary after years of financial uncertainty. Then Elon Musk bought Twitter.

A $42,000-a-month problem

The resulting API changes created another existential threat. Publer was told access to Twitter's enterprise API would cost $42K per month. Around $20K of Publer's monthly revenue at the time came from customers using X.

Dropping X would have protected the balance sheet, and staying meant taking a substantial financial hit. Kalemi chose to stay, but the API bill was only part of the problem. Publer had grown beyond 20 employees and accumulated department heads, hierarchies, larger offices, paid marketing, SEO and influencer campaigns.

Despite generating $1.5 million in revenue in 2023, Publer ended the year with a $140K loss. So Kalemi did something that is often harder for founders than raising money or hiring people: he reversed coursе. That meant removing layers of hierarchy, letting people go and stopped paid marketing. The company returned to a flatter structure and focused on doing more with less.

The results appeared in 2024 when Publer returned to profitability and generated around $2 million in sales, roughly 40% more than the previous year. The company improved its analytics, added hashtag and competitor analysis, expanded its integrations to 13 platforms and continued developing its mobile application.

Publer began building free, functional social media tools as part of a programmatic SEO strategy. Those seemingly simple tools generated millions of visitors and app installs at a fraction of the cost of traditional marketing.

The lesson was very different from the one Kalemi had learned during Publer's early years - growth did not necessarily require spending more. Sometimes it required understanding the product, the customer and the channels around it better.

That is particularly important for a bootstrapped company, since Publer could not simply solve every growth problem by raising another round. It had to find distribution mechanisms that worked economically, then reinvest the returns into the product.

From social media scheduling to automation and AI

By 2025, Publer was expanding again. The platform was being localized into German, Italian, Spanish, French and Albanian. Automation was becoming a larger part of the product, while its API was opening Publer to external workflows through tools such as Zapier, n8n, Make and Pabbly.

Its AI assistant was also becoming more sophisticated, adapting to a user's tone of voice and learning from post performance. And then there was Linkie, a separate product that emerged from one of Kalemi's earlier attempts at diversification. 

Back then, Kalemi had launched Kibo while Publer was still maturing, only to realize that splitting attention between two products was hurting both. He eventually shut Kibo down and returned his focus to Publer.

This time, the expansion appears more planned, with the difference being that Kalemi learning about the effects that attention, capital and organizational complexity can have on a second product.

The company did not follow the conventional startup trajectory of raising large rounds, hiring aggressively and chasing growth at all costs. Yet each time with every new version, Kalemi went back to the same thing: the product, the customers and the question of whether Publer could become a sustainable business. The headline number is now approaching $3 million ARR, but the more important story is everything that happened before that number became possible.

For a founder who once worked until 3 am because he had a product he could not stop thinking about, the ambition is still there. An additional hard truth of building a company is also that commitment does not disappear once the product gains traction. If anything, the stakes get higher: there are employees depending on you, customers expecting you to deliver and a business that can quickly fall behind if you stop pushing.

Creators lock in their holiday content calendars 90 days out, before most ecommerce brands finalize their commission strategy and way before Black Friday and October deal events.

Get ahead of the seasonal rush with The 90-Day Holiday Sprint, a practical guide for brands that want creators driving holiday demand while competitors are still recruiting:

Your 90-day countdown starts now.

Get the Free Guide

IT Logs:  What helps startups win faster: moving quickly or building something truly deep?

IT Logs: Do you think AI-first startups will clearly beat those just adding AI on top?

IT Logs: What is slowing down founders the most: hiring, fundraising, or regulation?