Validating via services before building a $27k MRR product
After Romàn Czerny's first business got acquired for seven figures, he decided to build a new business around the growth tactic that was responsible for that success.
He started by doing it manually for clients, then created a product to automate the process. Now, Gojiberry AI is at $27k MRR.
Here's Romàn on how he's doing it. 👇
A 7-figure exit
My name is Román. I’m 30 years old, I'm currently based in Lisbon, Portugal, I have a weakness for pastéis de nata, and I’m the cofounder of Gojiberry AI.
I started doing online business in 2019, first as a freelancer, then through affiliate marketing. In 2023, I discovered a passion for SaaS, leading me to build my first product, CocoAI, which I later sold for seven figures.
Today, I’m working on Gojiberry AI, an intent-based LinkedIn outreach tool. It helps B2B companies find and convert customers on LinkedIn by leveraging buying signals and reaching prospects when they’re ready to engage. We reached $27k MRR at the beginning of December, and we made $35,000 in sales the same month.
Scratching his own itch
Our previous SaaS sold WhatsApp solutions for e-commerce. We saw that e-commerce founders actively engaging on LinkedIn, especially around topics like email marketing for e-commerce, converted significantly better.
So we started doing it manually. We searched for e-commerce founders interacting with specific keywords and topics. It worked extremely well, but it was incredibly time-consuming.
After selling that SaaS, we asked a simple question: Why not turn the system that helped us sell our previous SaaS for seven figures into a product?
That is how Gojiberry AI was born, built on the same approach that powered CocoAI's growth and exit.
Validation through manual services
To build the MVP, we did not write a single line of product code. Instead, we first focused on getting customers. We told them we would deliver high-intent leads, and then we manually searched for those leads on their behalf.
Finding 100 qualified prospects often took several hours, sometimes up to five. We spent an entire summer doing this manually, with the help of multiple virtual assistants based in Pakistan and India. It was painful and completely unscalable. But it proved two important things: Real demand existed and customers were genuinely happy with the results.
That validation pushed us to turn the process into an actual product with Gojiberry AI.
That wasn't easy. We had to automate a very human, intuitive process. Identifying real buying signals on LinkedIn is not just about scraping data. It requires context, judgment, and timing. Translating this into a reliable system without losing quality took time and many iterations.
Tech stack
I’m not technical at all; my CTO handles all the technical stuff. But we use:
AWS
Framer
Firebase
Gemini
Claude
ChatGPT
Growth and consistency
Our growth strategy was multi-dimensional:
Reddit: We built early trust by sharing real data and transparent results in SaaS communities.
LinkedIn: We ran a dual engine, viral inbound content combined with high-intent outbound campaigns using our own tool.
Cold Email: We scaled volume, sending thousands of emails daily using leads sourced from Sales Navigator and Gojiberry AI.
Influencers: We partnered with trusted B2B voices to borrow credibility and reach high-quality users.
Other tactics: We layered the above with YouTube tutorials, newsletter sponsorships, webinars, and building in public on X.
No single channel was a silver bullet, but we found working with B2B influencers on LinkedIn particularly helpful.
Our core customers already spend time on LinkedIn, so partnering with influencers who speak directly to that audience gave us instant credibility and distribution. It allowed us to reach the right people, in the right context, with a message they already trusted.
And overall, the real secret was the compounding effect of showing up everywhere, every single day.
Expansion revenue
Our business model is subscription-based SaaS. We sell Gojiberry AI on monthly and annual plans. As teams grow and rely more on intent-based outreach, they naturally upgrade their plans, which creates built-in expansion revenue.
From the beginning, we made sure customers could see results quickly, meaning qualified conversations and booked meetings, not just data or dashboards. That made it easier to convert trials, reduce friction in sales, and drive word of mouth.
Narrow down your ICP
Here's my advice: Pick one clear ICP and one clear use case. Trying to serve everyone slows everything down. A narrow focus makes product decisions, messaging, and growth much easier.
Initially, we tried to serve too many profiles at once. Once we focused on the users who truly valued intent-based outreach, everything became easier. If I had to start over, I would validate pricing and positioning earlier and narrow the ICP faster.
Also, try all marketing channels. There is never enough marketing!
What's next?
Our main goal is to make intent-based outreach the default way B2B teams generate pipeline.
On the product side, we want Gojiberry AI to become the standard system for detecting real buying signals and turning them into qualified conversations across LinkedIn and other channels.
That means deepening signal quality, automation, and integrations, while keeping the experience simple and outcome-driven.
You can visit gojiberry.ai and start a free trial. If you want 14 days instead of 7, send us the word INDIE14 in the in-app chat and we’ll extend your trial.
Leave a Comment
Grateful for Professional Crypto Recovery Support
After falling victim to an online scam last year, I lost 1.3 Bitcoin, which was devastating and incredibly stressful. For weeks, I felt stuck and unsure if recovery was even possible. Instead of giving up, I continued researching legitimate recovery options and eventually came across a specialist who took my case seriously.
The process required patience, documentation, and clear communication, but I was impressed by the professionalism and transparency throughout. To my relief, I was eventually able to recover my stolen funds, something I honestly thought might never happen.
I’m truly thankful for the guidance and effort provided during such a difficult time. If you’ve experienced a crypto-related scam and are looking for professional assistance, it’s worth reaching out and exploring your options. (RECOVERYCOINGROUP AT GMAIL DOT COM)
The manual validation approach really hits home. I'm on Day 4 of launching MeetDone - a tool that turns meeting transcripts into follow-up emails.
Before building, I lived the pain myself: spending 20-30 mins after every client call writing a professional follow-up. Now it takes 30 seconds.
Still at 0 paying customers, but the problem is real - just need to find the right people. Your point about narrowing the ICP fast is something I need to focus on. Freelancers and consultants seem like the right fit, but maybe I need to go even narrower.
Thanks for sharing the journey - encouraging to see the path from manual pain to $27k MRR.
I am working on marketing part of my SaaS, quite challenging at early stage, without huge financial input.
Finding 100 qualified prospects often took several hours, sometimes up to five. We spent an entire summer doing this manually, with the help of multiple virtual assistants based in Pakistan and India. It was painful and completely unscalable. But it proved two important things: Real demand existed and customers were genuinely happy with the results.
How did this initial idea come about? And for indie developers who want to validate their product in the early stages but struggle to find 100 potential users—do you have any advice? Thanks a lot!
This resonates. Years of freelance work showed me exactly what problems people pay to solve. The tricky part was turning that into a product instead of just more client work.
Getting paid while you figure out what to build is underrated.
Strong write-up. The part that resonates most is validating behavioral friction before building anything.
In my experience, the biggest trap isn’t lack of ideas, it’s confusing “sounds useful” feedback with actual workflow pain. Until someone is willing to change how they currently operate (or give you their email/money), the signal is mostly noise.
Out of curiosity — in your early validation, what was the most misleading positive feedback you received before you saw real traction?
The manual validation phase really stood out to me. Spending an entire summer doing painful, unscalable work before writing code is something many founders skip.
Curious — during that manual phase, what was the biggest signal that convinced you people would keep paying long-term, not just try it once?
Great breakdown. The “manual services first” part really stands out.
Curious - during the manual phase, what was the strongest signal that convinced you this was worth turning into a product? Was it retention, referrals, or clients asking to scale volume faster?
Love this. Services are the fastest way to validate demand + pricing before you burn months building. The service → product path is underrated.
Love it !
We have been customers of gojiberryAI for a while now and the tool is amazing.
Same here — tools like that are the best when they quietly save you time every day without adding complexity. What’s been the most valuable part for you so far (accuracy, workflow speed, or something else)?
Impressive journey, Romàn! It's incredible to see how you turned a successful exit into a new product that leverages the very tactics that made your previous venture successful. The transition from doing manual lead generation to automating it with Gojiberry AI is a perfect example of solving your own pain point, and the results speak for themselves.
I love how you validated the idea through manual services first before jumping into product development. It's a smart approach that minimizes risks and ensures there's real demand. The fact that you didn’t write any code for the MVP and instead focused on customer validation is a great lesson for any aspiring SaaS entrepreneur.
Your growth strategy is another gem. Leveraging multiple channels, from Reddit and LinkedIn to cold email and influencer partnerships, shows how versatile and scalable a multi-dimensional marketing approach can be. The key takeaway for me is that there’s no silver bullet, just consistent, compounded effort in the right places.
Also, focusing on a narrow ICP and use case is super valuable advice. It’s easy to get distracted by the idea of serving everyone, but honing in on the right audience makes all the difference. I’m excited to see what’s next for Gojiberry AI as you expand and refine the platform!
Do you have any advices for who are trying similar approaches but failing. For example I am trying the same approach for Reddit for whenasked[dot]com platform. But couldn't find anyone to use it for feedback.
Turning the exact growth loop that led to a 7-figure exit with CocoAI into a focused product like Gojiberry AI feels like the ultimate “don’t reinvent, refine” move. Also love the reminder that intent beats volume every time, and that narrowing the ICP makes everything easier (product, messaging, sanity).
Really interesting breakdown.
In a crowded market like this, what took longer than expected to figure out —
the gap itself or the distribution strategy?
This resonates. I underestimated how painful maintaining scrapers would be long-term. In hindsight, data reliability mattered more than speed of iteration. The manual approach forced me to understand what data points were actually critical vs. nice-to-have before building infrastructure around them.
This is a great example of something people say they believe, but rarely execute: services as validation, not a fallback. ~
It’s not that you got to $27k MRR, but that you delayed product on purpose until demand was undeniable. Many founders feel the need to ship software. You Used Manual Effort To Buy Clarity.
Some worthy points to note.
Real constraints were imposed by manual work.
Enabling access to services revealed the true value of intent, timing and context, not where it was easiest to automate.
Sales dictate product decisions before any software change.
The product roadmap wasn’t theoretical as customers were already paying Every feature of friction was experienced by you.
Distribution was baked into our strategy from the beginning.
LinkedIn was not a “channel” you found out about later it was where the pain already lived.
Many people think services don’t scale so it’s wasting effort. But it’s software that is unvalidated that’s wasting effort. Services reduce feedback cycles, turning guessing into observation.
What’s interesting isn’t that you automated outreach it’s that you first proved people want these outcomes badly enough that they’ll tolerate you delivering it manually. The product lowered the cost, speeded up and made repeatable.
The right kind of boring is how real businesses usually get started.
This approach highlights something important.
Services force you to sit with real context — real conversations, real constraints.
That signal is hard to get once you jump straight into building.
Validation isn’t about speed. It’s about clarity.
The manual validation approach really stood out to me.
Starting with services and getting people to pay before building the product is such a strong signal of real demand. It’s a great reminder that effective validation doesn’t have to be complicated.
How do you go about getting customers?
This is a great example of letting customers shape the product instead of the other way around. Starting with services feels like a smart way to reduce risk, especially because you’re getting paid while learning what actually matters to users.
I also like how this approach naturally forces you to focus on outcomes rather than features. When people are willing to pay for a service, it becomes much clearer which problems are urgent and which are just nice to have. By the time you transition to a product, a lot of the hard prioritization work is already done.
It’s encouraging to see that this path can lead to real revenue without needing a massive launch or upfront build. Definitely a good reminder that momentum often comes from practical execution, not perfect ideas.
Loving the articles James! Thanks for writing! Be great to share more how we pivoted from Dolphin AI to Pretty Prompt (Grammarly for prompting) and hit 20k users in 6 months with users from LinkedIn, Upwork, Chili Piper, Wix 🙌
This resonates a lot. Doing things manually first feels inefficient, but it forces you to see what users actually value versus what just sounds good in theory. I’ve noticed the biggest signal isn’t usage, it’s when people start asking for the same thing repeatedly without being prompted.
The manual validation bit is the part most people skip because it feels like backwards progress. You're supposed to be building a tech company, not managing VAs in Pakistan for a summer.
But there's something powerful about doing the painful version first. You learn exactly which parts suck the most - and those are the features users will actually pay for. If you'd jumped straight to code, you probably would have built the wrong automation first.
The other underrated benefit: by the time you do build the product, you actually understand your customer's workflow inside out. Hard to replicate that from market research alone.