How to Run a Subscription Design Business: Vadim Carazan on WeGrow, Nine Days and AI

‍Ian: Today I'm really excited to be joined by Vadim Carazan, who I've been following for a while on X. He's such an awesome designer, someone worth following, sharing great advice, and more recently doing some really interesting stuff with AI and the new tools out there. It's really shown me that these are things that can enhance what we're doing and open up opportunities. Let's start with your origin story. How did you get into logo design and branding?

How Vadim got into logo design and branding

Vadim: Thanks for having me, Ian. It started fourteen years ago, by accident. I was a freshman studying oil engineering in Romania. I was born in Moldova and moved to Romania for high school. A friend of mine was freelancing, doing some logo work on the side, and I thought I could make some extra money too. That night I opened Illustrator for the first time, found a logo contest on 99designs, and just started. I didn't know Illustrator, I'd only played with Photoshop as a kid, but I got obsessed.

I didn't win that first contest, but I kept entering for about four months until I won one, a small fifty dollar job. I remember going to the ATM to take out the money and it felt unreal, like I'd been playing on the internet and here was real money. From there I stayed on that freelance platform for about three years, building a portfolio and client reviews, learning what clients actually respond to. Then I discovered Dribbble, and by the following spring I was on the popular page and picking up my first proper clients from there.

Ian: You clearly have a natural talent for it. I want to flag that platforms like 99designs get a lot of bad press, and rightly so in many cases, because they can exploit designers. But in your case it sounds like you used it as a way to practise, get fictional briefs, and build a portfolio, which is different. My own advice for anyone starting out is usually to find a local business that needs help and offer your time for free in exchange for a testimonial. There's no risk to the client, and if it goes well you get a real case study and word of mouth. But I get why the contest model worked for you, you needed that reward to stay motivated.

Vadim: Exactly. There are different ways to learn design. For me, learning through real cases with real money on the line was the only thing that kept me engaged. Watching tutorials never held my attention the same way. I wouldn't necessarily encourage contest sites, but for a beginner figuring out how to spend their time, I think it can work.

Building an audience, then losing it overnight

Ian: You've gone on to build a proper business with a team, WeGrow, plus your newer product Nine Days. How did you get from a hobby on Dribbble to where you are now?

Vadim: At my third year of university I dropped out, still one of the best decisions I've made. From that point design had to work, there was no plan B. I freelanced on that first platform for three years, then spent seven years on Dribbble, posting daily and building up leads. Then in 2022, the leads just dried up completely. I realised I'd made a mistake, I'd put all my eggs in one basket. I had around 80,000 followers on Dribbble but no website of my own and barely any presence anywhere else.

So I moved to X and set myself the 36 Days of Type challenge. Near the end of it, on the letter W, I dug into my archive and found an old W built from overlapping negative space circles. I tidied it up, posted it, and named it We Grow after an old unused project idea. It got a good response, and I started thinking about turning it into a passion project. Around the same time I was talking to Victor, who became my co-founder, about starting a studio together. We'd already picked a different name, but at the last minute I suggested using the We Grow logo instead, since we both liked it and the story behind it. Design subscriptions were the big conversation on X at the time, and traditional studio models were struggling, so the choice felt obvious.

Ian: There are a few real lessons in that story. Posting daily, whether on Dribbble or on X, is something I did too, and it becomes a routine you build your day around. Having a child disrupted mine, but for a long time it was one of the first things I did each morning. The second big lesson is exactly what you said, putting all your audience in one place you don't own. I had a version of that with X myself. If you want to make a living as a designer, having your own website is close to compulsory, it's the one space nobody else can pull the rug out from under you. And the third thing, which keeps coming up in these conversations, is the value of setting yourself a challenge. Yours was 36 Days of Type. Mine was posting daily. A friend of mine, Liam Jackson, is running a challenge where he shares everything he learns from side projects, and it's clearly helping him find his voice. I think any designer would benefit from picking a challenge like that.

Inside the retainer and subscription model at We Grow

Ian: Can you explain the retainer or subscription model for anyone unfamiliar with it?

Vadim: For us it's really just another way to charge clients. There's a separate idea of "productising" design, treating requests like a queue you work through mechanically, and I don't really believe in that. What we're trying to do is act like an in house design team for our clients. We stay closely connected, sometimes with daily fifteen minute calls just to check on progress. It's always been roughly fifty fifty for us between retainer clients and project based work. Some agencies running subscriptions manage twenty or more retainer clients at once, but our number has always stayed low, around eight maximum, which keeps it manageable.

Ian: There's a well known framework for this kind of model, using a tool like Trello to give each client a queue where they can only pull one job into progress at a time. Do you manage it in a similar way?

Vadim: We've got two types of retainer clients, ones who stay with us for two or three years, and shorter term clients who treat it like a cheaper version of project based work, sometimes just one or two months. We're fine with either. Long term retainers give us the balance we need. If a client doesn't have ongoing work, they'll pause or cancel, and we replace them with a new client. We do let clients pause their subscription for weeks or even months and pick it back up later. We've also had clients pay for six months without sending a single brief, because they were really paying for guaranteed availability whenever something urgent came up.

Ian: That's the dream scenario, being paid to simply be available. What happens at the other end, if several retainer clients all need something at once?

Vadim: With a smaller number of retainers, like our eight, the odds of everyone needing something on the same day are low, and it does happen occasionally, but we have a team to absorb it. Normally we quote two to three days to deliver work. If something urgent needs to go to print the next day, we'll prioritise it, and everyone else waits their normal turn. Communication matters a lot here too. If we're sick or something takes longer than promised, we'll add those extra days onto the client's subscription so they never feel like they're losing money for time we didn't deliver.

Ian: Having a predictable, repeatable income is one of the biggest advantages of this model. My own offering is more productised, logo design mainly, which means I'm constantly reliant on new business. AI is changing that too, since a lot of founders are now generating a rough logo themselves in ChatGPT before they ever think about hiring a designer.

Nine Days: strategy, identity and a landing page in one sprint

Ian: Tell me about Nine Days, your newer product.

Vadim: The idea itself isn't new, branding sprints have existed for a while, and most studios focus on just one piece, either branding or web design. Our idea was to combine everything into one launch package, so a client walks away with a finished landing page ready to launch. It sat on the shelf for about a year before we finally launched it at the end of May. We're currently working on our first client project under it. One thing we've built in that not many subscription agencies do is strategy and positioning work, alongside naming, logo design and product UI, all delivered under the retainer model.

The process starts with strategy and positioning, then moves into logo and visual identity in a single direction rather than multiple concepts, because a clear strategy makes that direction obvious. We present that to the client, take feedback, then move into the landing page.

Ian: How does the nine day timeline actually work in practice?

Vadim: We like to start midweek, running a strategy workshop on a Thursday, working Friday, then picking back up on Monday. The weekend isn't counted, but I still think the brain keeps processing ideas even when you're not actively working, so it's not wasted time. Nine days is really the ideal case scenario. In reality there's almost always a round of revisions, and how long that takes depends on how quickly the client gives us feedback, that time isn't counted against the nine days either.

Ian: So in reality it's a sales pitch and a framing device as much as a strict deadline, the real product is a design sprint with clear deliverables, and nine days is the best case outcome. I think that's a really good lesson for anyone building a design business, spotting a model and building a distinct offer and story around it, rather than just doing design work reactively. It's very much the "work on your business, not just in it" idea from books like The E-Myth Revisited.

Is AI a threat to designers?

Ian: Do you think AI is a threat to designers?

Vadim: I wouldn't call it a threat or an enemy, but if you're asking whether it will eventually take our jobs, yes, eventually it will. That can be a good thing for some designers and a difficult thing for others. If you're purely moving pixels, that job is already gone. If you're a thinker, someone who creates experience and works through ideas, I think it's a great tool to work alongside.

Ian: I'll push back gently there, because a lot of designers are genuinely scared. There was that viral post from a designer who quit after seeing an AI tool design a logo better than he felt he could. I've noticed it in my own town too, AI generated branding on shop banners and leaflets that's obviously AI to a trained eye, but invisible to most people. It reminds me of how quickly AI video went from obviously fake to genuinely convincing. My honest take is that there will always be a market for both, the same way IKEA and handmade furniture, or Primark and Prada, both exist. The easy jobs, the ones that used to go to students and beginners, are the ones most at risk. But people still want a story and a human behind their brand, not just an AI output.

Vadim: It's bigger than design too. A lot of the conversations I'm hearing are about employment generally, not just creative work. Design solves real problems, and it exists because there's a competitive market. That won't disappear in the next few years. I'm not saying it's time to walk away from design, there's still a huge amount of opportunity, especially in design tools and products. A lot of designers are experimenting with AI right now, building and launching their own products. I think we'll see a wave of designer founders, because designers have always had the ideas but were limited by skills like coding. That barrier is coming down.

Ian: That barrier dropping is probably the single biggest shift. I'm currently writing a new book, and I'm using AI to help tidy up my own drafts. Before AI, I'd spend an entire day wrestling a single paragraph into shape. Now I can get a rough draft into something readable far faster, then rewrite it back into my own voice. It's not AI writing the book, it's removing friction from a part of the process I've always struggled with.

Vadim: With AI there are almost no creative limitations left, but that creates its own paradox. When you can generate literally anything, it becomes harder to find the right idea, your mind jumps between options endlessly. I think creativity actually needs constraints. Independent designers used to be limited by their own skills or budget, if they couldn't afford an illustrator, that constraint shaped the work. Now you can generate anything, which is exactly why strategy and positioning matter more than ever, they're what create direction and constraint inside an otherwise unlimited space.

Building Tessera: an AI powered tool built in a day

Ian: You've built some genuinely interesting tools using AI, can you talk through one of them?

Vadim: I built a small app called Tessera that generates a portrait of a designer using their own design work, pulled from their portfolio. It was inspired by a viral tweet around designers celebrating their own work, and I felt the design community reacting to that in real time. The idea itself was simple and took about a day to build.

Ian: That's the same week we originally started talking about doing this interview, and you already had a working tool. Which AI tool did you actually build it with?

Vadim: I used Claude, specifically Claude Fable 5. I found a reference treatment on Pinterest, similar portraits built from smaller product images, and used that as a starting point, then iterated a lot to get the look right. The hardest part wasn't the AI generation, it was reliably pulling the right images from a designer's Dribbble profile without accidentally grabbing unrelated images, Dribbble has a colour filter feature that was getting pulled in by mistake, and I had to work out how to filter that out.

Ian: For anyone who hasn't tried vibe coding yet, this is a great example of just starting with a small, specific idea. I've seen a similar concept from another designer I follow, who built a tool for a client with a flower inspired brand mark that lets the client generate endless unique variations of their bloom symbol using sliders. It reminds me of the old Tate Modern logo, which was designed to shift and change, but in practice the team using it just kept reusing the same static file because generating a new version each time was too much friction. A vibe coded tool removes exactly that friction, and makes a genuinely dynamic identity system practical for the first time.

Vadim: Agreed, and I think that's the most exciting use of these tools for designers specifically, they offer structure and control that you don't get from just prompting Midjourney or Claude directly for a one off image. Studios used to build custom tools like this for clients before, but now anyone with an idea and a real brief can build one. We'll still need people to make the direction human, to decide what's actually good. AI has a huge amount of knowledge but not real experience, it doesn't think the way we do after years of working on real projects. That's still our advantage, at least for now.

Ian: A great note to end on. This has been a genuinely useful conversation with a lot of practical takeaways. Thanks so much for coming on.

Vadim: Thank you, Ian, it's been great.

Ian: Where can people find out more about you?

Vadim: The best place is X, I share my work and my thoughts on AI and design there daily.

Ian: I'll link everything in the show notes. Thanks again for your time.