My job title says Growth Marketing Manager. In practice, it means I'm the blog, the SEO person, the social media team, the video editor, and the person who sends the newsletter.
I'm the only person handling marketing, which means I wear pretty much every hat.
So when people ask how I manage to do it all, the honest answer is that I don't. I focus on a few things consistently and build systems that allow every piece of work to be reused in multiple ways.
Here's what that looks like, what has worked, and what hasn't.
First, What Happened
I'll get a few numbers out of the way so the rest makes sense:
- Our open-source repository crossed 10,000 GitHub stars.
- I organised a hackathon, bringing together developers to build real-world applications.
- We've seen growing interest from inbound leads, with content playing an important role in building awareness.
- We've been working on improving our visibility across traditional search engines and AI-powered search platforms.
- I manage multiple X accounts, some of which have seen significant growth, with individual posts reaching hundreds of thousands of impressions.
None of this happened overnight, and none of it came from doing everything at once.
Pick the Two Things That Compound
When you're working alone, the to-do list is practically infinite. There are always more platforms to post on, more content to create, and more channels to experiment with.
I started asking one question about every marketing activity:
Will this keep working after I stop working on it?
A blog post can bring traffic months after publishing. A well-ranked comparison page can continue attracting relevant visitors. A GitHub README can keep introducing developers to a project.
Some marketing activities have a much shorter lifespan. A social media post might get most of its engagement within a day. A paid campaign stops generating results when the budget runs out.
That's why I prioritised long-term content and our open-source presence. Everything else either supported these priorities or had to justify the time it took.
What didn't work: Trying to be everywhere during the first few months. I experimented with too many platforms because it felt like something a marketing team should do. Engagement was flat, and I was spreading myself too thin.
Dropping activities that weren't working was one of the best decisions I made.
The Content Engine: One Post, Five Outputs
Most of my content starts with a blog post. But publishing the article is only the beginning.
I break it down into multiple pieces:
Blog Post
|
+------------+------------+
| | |
X Thread Single Posts Short Videos
| | |
+------------+------------+
|
Newsletter
|
Founder Content
For example, a blog post about automating a social media app with an AI agent contains several smaller stories.
- The main argument becomes an X thread.
- An interesting statistic becomes a standalone post.
- The actual workflow becomes a 30–40 second screen recording.
- The key takeaway becomes a newsletter section.
- The opinion behind the article becomes a founder post.
One piece of research can give me five different content outputs.
Writing the original article takes the most time. Repurposing it takes considerably less. That's the whole trick.
If I had to create five completely separate pieces of content from scratch, I'd probably publish one-fifth of what I do now.
My approach is simple: create once, distribute everywhere, but adapt the format to the platform.
What didn't work: Sharing a blog link with a caption like "New blog post, check it out."
Nobody has a reason to click just because you've published something. The social post needs to be valuable on its own. The link should be an extension for people who want to learn more.
Writing for Developers
Our readers are developers and people building AI agents. They can spot marketing copy from a mile away.
So I try to write the way an engineer would explain something to a friend.
Short paragraphs. Real examples. Actual commands. Specific numbers. And no pretending a product can do something it can't.
A few rules I follow:
- Show the mechanism. Explaining how something works is more useful than calling it intelligent or powerful.
- Include code. Even a simple command like
pip install package-name makes an article feel more like a useful resource than an advertisement.
- Be fair about alternatives. If another tool is better suited to a particular use case, say so. Developers appreciate honesty.
- Problem first, product second. Start with something the reader cares about. Introduce the product only when it naturally fits into the discussion.
What didn't work: My early drafts were too polished. They sounded like brochures.
The content that performed best felt like it was written by someone who had actually tried the technology, run into problems, and figured things out.
SEO and Showing Up in LLM Answers
This is one of the areas I'm most interested in because search behaviour is changing.
People don't just Google things anymore. They ask ChatGPT, Perplexity, and other LLMs questions like:
"What are the best tools for automating mobile apps with AI agents?"
If your product isn't mentioned in those answers, you're potentially missing out on people actively looking for a solution.
A few things have helped us build visibility:
- Comparison pages. People searching for alternatives or comparing tools are often close to making a decision. Comparison articles help answer those questions directly.
- Consistent positioning. Maintaining consistent messaging across different channels makes it easier for people and search systems to understand what you're building and where it fits.
- Structured information for AI crawlers. Files like
llms.txt can provide AI systems with a concise overview of a website, its products, and important documentation.
- Third-party mentions. Your own website saying you're great is one thing. Independent articles, community discussions, launch posts, and recommendations provide another layer of visibility.
What didn't work: Chasing high-volume keywords where we had little chance of ranking.
A small developer-focused product probably won't beat established publications for a broad keyword like "mobile testing". But it has a much better chance of ranking for a specific query like "AI agent for automating Instagram on a real phone".
I'd rather attract 100 relevant visitors than 10,000 people who aren't looking for what we're building.
Managing Multiple X Accounts Without Losing My Mind
Managing one social media account is already a job. Managing multiple accounts requires a slightly different approach.
The trick is making sure each account has a distinct voice and purpose.
Different audiences respond to different types of content. Some are interested in product updates and demonstrations, while others care more about opinions, industry observations, and personal experiences.
I try to adapt the content accordingly rather than publishing the same thing everywhere.
I've also noticed that demonstrations and strong opinions tend to travel much further than generic product announcements.
A 30-second video showing an AI agent completing an actual task can do more than a long list of features.
My routine also includes spending time replying to relevant conversations and engaging with people in the AI and developer communities.
Replies have played a bigger role in audience growth than I initially expected.
What didn't work: Scheduling an entire week of posts and then disappearing.
Social media isn't just about publishing. It's also about being around when people respond.
I now batch-create content but try to stay active around the time a post goes live.
Organising a Hackathon
Organising a hackathon was one of the most demanding things I've taken on as a solo marketer.
It involved much more than announcing an event and getting people to register.
There was promotion, coordination, onboarding, participant communication, judging, and making sure teams could actually start building.
A few things made the process easier:
- A focused theme. The challenge needs to be specific enough that participants know what to build within a weekend.
- Easy access to tools. Give participants access to the required infrastructure so they don't spend half the event figuring out setup.
- A starter repository. Provide a working starting point rather than expecting everyone to begin from scratch.
- A setup guide. A short walkthrough video can eliminate a surprising number of repetitive support questions.
- Clear communication channels. Make it easy for participants to find help, connect with other teams, and submit their work.
The real value of an event doesn't end when the winners are announced.
The projects can become demo videos, social media posts, tutorials, and even future blog content. Participants may continue building, contribute to open-source projects, or become long-term members of the community.
One event can create content and community long after the event itself is over.
What I'd do differently next time: Start promotion earlier, involve more mentors, and spend more time making the onboarding experience smoother.
Turning Content Into Business Impact
Publishing content and getting impressions is great, but marketing eventually needs to contribute to the business.
I try to make sure every piece of content has a natural next step.
It could be:
- Trying the product.
- Reading the documentation.
- Exploring a particular use case.
- Getting in touch with the team.
The CTA doesn't need to be aggressive. It just needs to make the next step obvious.
I've noticed that content demonstrating practical applications tends to attract more meaningful engagement and business interest.
Someone reading about a problem they personally struggle with is much more likely to explore a solution than someone reading a generic feature announcement.
I also think it's important to remember that marketing doesn't operate in isolation. Understanding the audience, their problems, and the questions they ask helps create better content and more useful conversations across the business.
What didn't work: Gated content.
We experimented with putting resources behind email forms. The number of signups was low, and the people who did sign up weren't necessarily qualified leads.
Making useful content freely accessible, with a clear next step, worked better.
What I'd Tell Another Solo Marketer
You can't outwork a team of ten. But you can out-focus them.
If I had to summarise everything I've learned, it would come down to a few things:
- Choose channels that compound. Prioritise work that continues creating value after you've moved on to something else.
- Repurpose everything. Don't create five pieces of content when one good idea can give you five outputs.
- Write like a human. Especially when your audience is technical. Specifics build more trust than marketing language.
- Give every channel a purpose. More accounts and platforms don't automatically mean more reach.
- Build systems, not just habits. A repeatable process makes it possible to maintain consistency without constantly starting from scratch.
- Measure business impact. Impressions and engagement matter, but meaningful conversations and qualified leads matter too.
Being a team of one doesn't mean doing everything yourself, all the time.
It means figuring out which work deserves your attention, building systems around it, and making every hour count.
That's how I've learned to run marketing without a marketing team.