Interview with Harrison Telyan, the co-founder of NUMI

Harrison Telyan @ NUMI (YC W20)

Meet the co-founder of NUMI, a full-suite design department for startups, on subscription. In today’s interview, we’re exploring an innovative company that's reshaping how startups approach design. Founded on the principle of making high-quality design accessible, this company introduces a unique subscription-based model tailored for founders and startup leaders.

We documented NUMI's story with Harrison Telyan, discussed the importance of design for early-stage startups, why design is more than just a logo and brand identity, and talked about AI and the future of work.


Could you introduce yourself and your startup?

My name is Harrison Telyan. I'm the co-founder of NUMI, which serves as your startup's design department. We offer a Design as a Service subscription model that pairs you with world-class designers focusing on everything from marketing designs, sales motions, banners, and ads to product design, thought leadership, information architecture, animation, and motion.

Could you tell us more about how you found your idea and how you validated it?

Just for some context, both my co-founder and I come from a design background. He is technical, whereas I was classically trained in industrial design. I'm a Y Combinator alum, a Rhode Island School of Design (RISD) alum, and have 10 years of experience across many different industries.

Lots of failures, but with failures - lessons. Most notably, I was the founding product designer at Imgur, where we scaled to about 400 people within a year or so, and then reached 330 million monthly active users.

That was my first job out of school. What I noticed was how chaotic the design process was, even as a designer, I thought that was chaotic. There's always been this notion within me to try to put it on rails. How cool would it be if you could take a chaotic system and then put it on rails, but not lose any of the value that that chaos brings. That's always been fascinating to me. It's also been fascinating for my co-founder as well. We've shared that love even before we met each other.

After Imgur, I joined a mortgage tech company called Sindeo, aiming to revolutionize the mortgage industry—from purchasing homes to understanding the risks and potentials.

Sindeo folded, unfortunately, there was an interaction of elements that kind of like brought them to their knees. I then moved to a similar company, Clara, which tackled the same issue but focused more on the internal aspects.

That's where I met my co-founder, Agree. We sat across the table from each other, and we became friends. I had a design background, and he had a technical one. I know that it might sound ‘corporatey’, but that really allowed us to build synergy, not only were we becoming friends, but we also understood one another's ability to see the importance of urgency, shipping, thinking through things together.

During our lunch walks, we'd discuss products, and our conversations were ever-expansive. One day, we just got it in our heads that Africa was almost like final frontier, where you could basically build anything due to the lack of legacy products. He inspired me to learn more about the continent, specifically Kenya and Nairobi, where he had spent an internship during college.

He described the notion of setting up software so that small businesses could run more effectively. When we arrived, it was a bit different.

Fast forward, we spent the first two years pivoting from making software for the wholesale medicine industry, which didn't take off.

Then we pivoted to another company idea, B2B wholesale for restaurants. In the third year, we noticed that restaurant owners were buying not only flour and tons of sugar but also items like tampons and butter, smaller personal objects.

This almost became a one-stop shop, very similar to Instacart. Fast forward to our fourth year after many pivots, we found ourselves becoming the Instacart of Africa. At its peak, we had displaced Jumia, the Amazon of the continent, becoming five times bigger in their grocery delivery game. However, the unit economics did not work out. When COVID began to spread, especially in places like Kenya, Nigeria, and Egypt, the continent, no stranger to outbreaks, adopted masks and precautions early.

Everyone wanted to get online to prepare for this incoming wave of corona. Boda Boda drivers, the couriers, raised their prices, as everyone wanted to get online in anticipation of the coronavirus. So our unit economics did not work out very long. We scaled fast without considering monetization or profitability and eventually closed down.

It was a tremendous experience. We learned a lot and I feel like we would be totally different people without that experience, but sprinkled in between there, we had gotten into Y Combinator. It seemed like the right time to pause and introspect. My father passed away from cancer during this period, which derailed our plans.

But through that, a handful of our founder friends reached out asking for design work, engineering work. So, we started leaning on people we could contract out to satisfy those engineering and design needs.

Suddenly, this side company that had loosely started in Africa and gotten us through Corona was making money. So, we were like, "Whoa, go figure when you lean into your core competencies, you can monetize something."

Screenshot of website

So, press fast forward again, we're about almost two years into it at this point with our company, but we've dropped engineering so that we can devote all of our focus to the design discipline just because it can go so deep.

We just want to have one North Star. But yeah, that's kind of the medium-short version. That's how we arrived at where we are today.

At this point, can we say that your startup is operating in the productized service business model?

I mean, we're a marketplace at this point, but we run things just a little bit differently than our competitors. You know, what's interesting is a lot of our competitors don't start off from a design background.

They start off with a marketing background. And oftentimes, companies, startups will go to some of these competitors of ours wanting product design, but they'll just get a flashy new website that can't translate into the product.

It just ends up as bullshit, right? Or like vaporware. And then they're stuck maybe with tens of thousands of dollars, hundreds of thousand dollars in some cases, holding the bag of this shitty brand identity that's actually not a design system that's not scalable, that doesn't actually capture anything about the ethos of the company.

We wanted to take this different approach where it's a company started by designers for designers. So we focus religiously on just making the designer the happiest person alive. Because if our designers are happy, that means that our customers and users are happy.

We wanted to kind of take this different approach where it's a company started by designers for designers. So we focus religiously on just making the designer the happiest person alive. Because if our designers are happy, that means that our customers and users are happy.

If our customers and users are happy, then Numi's customers are happy. And if Numi's customers are happy, Numi's happy. It just creates this flywheel of support and happiness, which in turn creates great products.

Can I ask you to elaborate more in detail about the financial part of the company? I mean, how does it financially make sense? How do you manage to offer a better price than, let's say, agencies or in-house designers without the hassle of hiring freelancers?

Yeah, I mean, so just to be clear, we do have freelancers. That's what makes up a marketplace. But I think what separates us from the rest is, there's some secret sauce in there that I'm not able to share, but I will say this, being a designer myself definitely affords us a certain type of unfair advantage in terms of assessing talent and what you might price that talent at.

But you talk to any of our designers, we never ever tell a designer to reduce their cost or suggest a price or something. They view us, our guild views us as design advocates first. We're always on their side from the jump.

We have tons of data points to suggest based on your skill level, your proficiency in certain tools, your discipline around what type of price would keep you competitive and what type of price might really resonate given a certain package or a certain skill set.

I think that's probably as much as I'll speak to it because our competitors take a different tactic where they will hire full-time or they will just pressurize these designers to lower their prices.

And it puts them in a position where they feel commoditized. And I don't know about you, but whenever you're doing your SEO work and a client's pressuring you, do you feel very inspired to give your best work? Probably not.

So again, it just kind of goes back to this ethos, this guiding light of how can we make the designers the happiest people on the planet? And we very much subscribe to that ethos of how Airbnb puts its host on a pedestal. We very much subscribe to putting our designers on a pedestal.

How scalable is your financial model? Do you think you could scale indefinitely?

Yeah, I think like, you know, the sky's the limit, right? And the sky goes on for a while. Yeah, I think like probably the biggest roadblocks that we might encounter might be TAM-related, you know, how big is the TAM for startups that need design work?

I think on the surface, it's really easy to dismiss it as small or maybe even medium, depending on the day, but I actually think it's quite large. When you come from a design background, like my co-founder and I do, the design tasks can get incredibly deep.

There is so much depth in the design world, including product design, user research, film, video, motion, animation, illustration, lighting, sound, voice, AR, VR, and graphics. There are so many disciplines.

Now, what we are seeing, which we love but are also not paying too much mind to, is our competitors trying to branch out into other disciplines like engineering, HR, writing, or more marketing-focused areas. Have at it; we are going to continue on in design and get really deep into this discipline.

Maybe they just don't know how deep it goes, or maybe they do, and it's uninteresting, but to us, it's very interesting because it's our bread and butter. It's what we love to do. So, we're very grateful for our designers to be on this adventure with us.

Do you think that in this market, this business model in general, it's possible to see consolidation where a few big players dominate the whole market in the future? Is there any advantage to scale in this business model?

I think that’s the case for any industry, like for tech, it's basically Apple, Microsoft, Google, OpenAI, right? For oil, it's Exxon, BP, Shell. Any industry is going to have major dominant players. They didn't get to where they are though without a certain type of allegiance to a belief. And so it's hard for me to speculate around what the other players in our design space are doing. But I can tell you this, it certainly helps coming from a design background. And I do know that we are quite anomalous as founders who are also designers. Typically, I find that designers don't really go into a business track or are business-minded.

Now, that typical nature, hopefully, what I'm seeing is now changing where you're seeing a lot more solopreneurs, you're seeing a lot more companies that are design-led like Pinterest or most notably like Airbnb.

But, you know, I wish that there were more. And I always tell our designers as well, wouldn't it be wonderful if maybe one day we went into competition together? The only person that's going to benefit is us because it forces us to innovate on behalf of the customer.

And I love thinking about that. We constantly get a comparison to SNL, right? All these famous comedians went through SNL like Adam Sandler, Eddie Murphy, Tina Fey, Amy Poehler, and now they're all really successful in their own careers.

And so if we could be, that alone would make us happy. We're already really happy. So, you know, if we just keep on doing what we're doing, it's just not going to stop. We have no intentions of stopping.

Solopreneurs are trying to leverage AI to launch one-person companies. So, they just outsource design or SEO or copyright to other third-party services, which opens new markets.

Sam Altman, I forgot where he was talking, but he was saying that he and a handful of his tech founder friends are in this group chat, basically placing wagers on when the next billion-dollar single-person company is going to happen.

And at its root, it's made possible by AI. Now, internally, there are a bunch of conversations that we're having with our designers, with ourselves, on how AI is going to be blended into the design process.

The top question is, are designers going to get replaced? And I really don't think that's going to happen for a while. I think there's a lot of context building and inputs.

And there are just so many human things that need to happen in order for something as seismic as a designer getting replaced by a computer script. But, you know, talk to me in five, seven years when AGI breaches the surface; that will be interesting to see.

But even then, Sam Altman said something really interesting as well, which is people don't really care what tech can do. People care what other people can do, right? And I think that's going to be true for a very long time.

I think that's just biologically baked into the human experience.

That might be true with designers, less with SEOs. Because we mainly rely on unstructured data. That's pretty much replaceable.

I mean, how are you seeing it in the SEO world?

Well, at this point, AI content writers suck. How do you define what the good writing is? AI tools they can optimize for keywords, for structured data, for metadata, for all sorts of technical parameters. But how do you define great writing? Why does AI content feel like it's written by a robot? At this point, AI is pretty much able to all the technical tasks, like the keyword optimization, link building, and so on and so forth, but it’s still writes robotic content. 

I don't know too much about that world. So it's always interesting to talk to an expert.

The second problem with AI content is the more AI tools appear in the market, the more AI content is on the web. So at some point, AI tools themselves will be trained on the content they generate, because most of the content will be AI generated.

It's almost like it's learning from the poor quality that it put out maybe two years ago. I guess that's why the New York Times is suing OpenAI right now. Because OpenAI's original model was trained on all this really sophisticated data, like the New York Times, maybe the Wall Street Journal, and all that jazz.

So yeah, it will be interesting to see how the quality bar is maintained because it's already pretty low. It will also be interesting to see how companies like Jasper maintain their relevance and significance in a world where, basically, if you get good at prompting, you can get away with $20 a month for your ChatGPT subscription.

So another thing on this, and then we can move on, obviously, but it's interesting to talk about where I'm not sure unless OpenAI is cooking up something we don't know about. I'm not sure whether companies like Jasper are going to lose their footing for a while.

A lot of what makes them great is they know their ICP. They know their ideal customer. They have spent years refining these models and taking out those keywords or making it short or refining that right tone, like you were saying, versus something like myself just trying to write up a prompt, not being an expert in that field at all. They do a lot of fast-forwarding for me.

People are selling AI as a tool to enhance human productivity. So instead of making you 20% more productive, you're basically made 70% more productive. But there are humans in the loop. AI allows you to sell workforce, not human labor enhanced by AI. That's something many people miss. And I think that will eventually happen in the future.

Yeah, that's a good point. I hadn't thought about it like that.

How important do you think design is to early-stage companies, early-stage startups, and not companies?

You know, I think design thinking and design as an output is less important than design thinking and user research. These are the core tenets of the discipline. I talk to maybe like eight to ten founders a day, and the common thread that I see between all of them is they're asking how they should be thinking about design.

How can they leverage design for profitability, for growth, for finding PMF, and I always tell them more or less the same thing, which is to go do it yourself, go hop into Figma yourself and start making prototypes that you can show to your customers.

Even before that, though, you should be trying your very hardest to talk to like ten customers a week, and that will continue until the company closes or you retire. Obviously, it's smaller for B2B companies or like enterprise, but you really ought not to start sentences with 'I think.'

You want to get in the habit of starting sentences with 'It should be because,' and then you're pointing to evidence such as your Mixpanel data or analytics, industry insight, customer interview, or testimonial. The less that you can start sentences off with 'I think' and switch it over to 'It should be because,' and then pointing to evidence, the more objective you're going to be and the less subjective you're going to be.

We also see this with early designers as well, if they're starting off sentences with 'I think,' they're typically maybe more on the art side. That's not really design. Design is objective; it's made for someone else; it's not for you, and you have to constantly be sourcing the root of where that solution came from in the first place.

It's incredibly important in these early stages. I think, though, that founders really genuinely struggle with knowing how design can be leveraged in those early days because I think they reduce the discipline down to, unintentionally, "I need a logo, I need a brand identity, I need to make some wires, and I need to learn Figma." Sure, that's maybe the implementation side, but I always say this as well: Figma, these design tools, they're just an implementation detail.

If you find yourself wanting to hop into Figma right off the bat, you are not doing it right. You need to go talk to your users first, and the only time you should be hopping into Figma or doing some of that other stuff is after you've spoken to your users and you know that you're done with a cohort of research, after you can start anticipating what those users are going to say.

The way that you run it, I tell this all the time as well, is go talk to your users and give them a prompt, and maybe it's like, "Hey, I want you to sign up," right? Using this prototype or this, you know, and you should always prototype before you build because engineers are really expensive, why not use Figma prototyping?

So you give them a prompt, "Hey, go sign up," and then they might turn to you and say, "How do I do that?" Well, you turn to them and say, "How might you do that?" And they're going to go through that flow.

During that flow, you're going to watch them put their hand on their head in frustration, or you might see some retinal vibration, which indicates confusion. You might see some cursor shake.

You might see them sighing or something like that. Those are great places to jot down and come up with problems they're experiencing. From those problems, you come up with "How might we solve this frustration at this point of the sign-up process?"

When you look at the "How might we" and the solutions, then you prioritize them. If those don't have a one-to-one mapping of your feature product roadmap, you're in trouble. That actually means you don't know your user.

Your product roadmap is wrong because that means you're inferring. You're starting off sentences with "I think" versus "It should be because" and pointing to the customer interviews you just had.

And it's like, "Well, how do you get really rich data outside of observing?" You ask good questions at the end. Like, "If you had a magic wand, what would you change about that experience?"

Was there anything in there that you were expecting to see that you didn't? Or simply, "What was frustrating about that experience? What would you do differently?" These open-ended questions and giving them the reassurance that critical feedback is welcome with open arms.

This is what separates million-dollar companies from billion-dollar companies. They are constantly at it, constantly asking the user questions, constantly observing their behaviors. That's where design starts for startups: talking to users and making sure they can understand them so well that they're anticipating what they're going to say before they even say it.

To be honest with you, before this interview, I was guilty of the same thinking about design: basically, a logo, a landing page optimized for conversion.

You know, most startups don't even need a website to begin with. They don't need a logo; they don't need a brand identity. I always remind them of the fact that Craigslist built a $10 billion company off the backs of blue URL links.

If you were to go to Craigslist right now, it's still as unattractive as it was 10 years ago. But guess what happens when you apply design to it? You've got a company like Airbnb that's doing the same thing, but they are very prescriptive with that user experience because they understand exactly where it's coming from.

It's coming from a place of knowledge. It's coming from a place of Joe Gebia and Brian Chesky staying and running Airbnbs themselves routinely. And obviously, that culture trickles down to your entire product team, your engineering team, your marketing team, legal, whatever.

So now you have a culture of design thinking, and it's ubiquitous across the entire company. When you have something like that, I mean, that is seriously potent. And I'm getting goosebumps just talking about it because it is exciting.

It is exciting not only for them because I use their products, but it's also exciting for NUMI because that's the culture we're trying to create as well. You know, another thing is like, design has a fun side. Design's fun side is like the soft underbelly of design: marketing design, sales motion, brand, and logo, all that fun stuff.

But most founders don't earn the right to work on that stuff. You know, most founders are really eager to hop right into that even before they've validated an idea or asked a user, "How much would you pay for this if this prototype was real?" You want to put yourself in a position when you're a founder of earning the right to do fun things. So we at NUMI, for example, did not earn the right to think about our brand identity until we made our first dollar.

A lot of founders today have been spoiled by the "move fast, break things" era of Facebook where funding was free. And it was a zero-interest-rate environment, and growth at all costs was the name of the game.

Now I think we're being brought down to earth again. "Hey, have high growth or have no growth at all." You may want to figure that out. And you may want to start monetizing as soon as possible so that if you do go out to fundraise again, you're coming from a position of power, you're not standing on one leg. Investors, more or less, as my co-founder agrees, are just loan sharks in khakis. So, you know, they're just looking for a payday.

But you're looking at your company as your child. They're two very different outlooks. So if you can walk into those chats with monetization under your belt, maybe even just one-fifth of your customers as an experiment, you've already put yourself a league above others, and you get there by understanding your customer, talking to them, observing and knowing what is valuable to them so you can charge.

Do you think that, let's say at some point in the future, AI is able to automate or maybe replace some of the workflows that perform the human design currently, do you think companies like yours that rely on high-quality human labor could adapt somehow or change the business model?

Yeah, I mean, we've already deployed; we use AI at NUMI all the time. Right, like we're constantly, I mean, right now we have a couple of projects that are being integrated into our workflows to automate some stuff, to bring.

It all just goes back to pointing to the designer. If we can build tools for the designer and we can also integrate AI into our workflows or our back office ops to free us up, like that's the win, that's the W, right?

LWe're not really focused on creating vaporware or some of these like GPT or OpenAI wrappers. We're also not rushing because we want to be really thoughtful about, hey, we don't want to just interact with AI just for the sake of us being a tech company and interacting with the latest technology.

We want to actually be thoughtful about what we're creating. We want to be really considerate of the time we spend on something versus the value it delivers. So we have a few things that maybe our competitors won't be too happy about giving away for a low cost or basically for free soon.

Thank you for your time and your willingness to talk about your experience and about what you think about design in general and the future. We’ve had a good discussion.

Yeah, likewise. Thanks for having me on the show. Appreciate it.

If you enjoyed the interview, subscribe to Founders Project

Join the conversation

or to participate.