How Coddy Hit $1M ARR in One Year Using Influencer Marketing

Hitting $1M in annual recurring revenue is a milestone most startups dream about. Doing it in just one year? That's the kind of growth that makes founders stop and ask: what did they figure out that we haven't?

On a recent episode of the Levels Podcast, we sat down with Barak Glanz, co-founder and CMO of Coddy, a gamified code-learning platform that's reached nearly 2 million users. What makes Barak's story particularly interesting is how his team achieved that $1M ARR milestone primarily through one channel that many startups struggle with: influencer marketing.

The Channel That Worked When Nothing Else Did

When Coddy first started experimenting with growth channels after raising their initial funding, they tried the usual suspects: Facebook ads, Google ads, and partnerships. But according to Barak, only one channel delivered meaningful results in those early days.

"At the very beginning only influencer marketing worked which is kind of weird I think."

This wasn't just about throwing money at creators and hoping for the best. The reality of influencer marketing is brutal: most campaigns fail, tracking is a nightmare, and it's easy to burn through budget with nothing to show for it. Barak experienced this firsthand, working with around 20 influencers initially with disappointing results.

What changed everything was taking a systematic, data-driven approach to a channel that most treat as more art than science.

Building CoolScript 2: An Algorithm for Influencer ROI

After those first 20 failed experiments, Barak did something unconventional. Instead of giving up on the channel or continuing to guess, he built an algorithm to predict which influencers would actually deliver returns.

He called it CoolScript 2 (version 2 because the first version, predictably, didn't work well enough).

"You give it a few parameters on an influencer channel and it estimates how much money you're going to get back from an ad. And this thing gave me so much leverage in the negotiations that I can estimate how much revenue I'm going to see back from an influencer."

The algorithm worked by analyzing multiple factors beyond just follower count. Barak tested it against historical data from previous influencer campaigns, comparing predicted revenue against actual attributed users and their conversion rates. Through iterations, CoolScript 2 became accurate enough to guide negotiation decisions and, crucially, to know when to walk away.

What Actually Matters in Influencer Selection

One of the most valuable insights Barak shared was learning which metrics actually predict success. Spoiler: it's not what most people think.

Follower count? Overrated. Average views per video? Critical.

"Number of followers is not that important. But the average number of views is very important. There is some correlation between the two, but it's not like 100%."

Channel size matters too, but not in the way you'd expect. Barak discovered there's a "sweet spot" for every company:

"If the channel is too small, it often wouldn't be worth it because it's too much hustle to work with them and even if you will see the investment back, it wouldn't be too much."

On the flip side, channels that are too large present a different problem. Big creators often have inflated expectations about their value, demanding payments that don't justify the returns.

For Coddy, the focus became short-form video content on Instagram Reels, TikTok, and YouTube Shorts. These platforms aligned with where their target users were already spending time, and the format worked well for demonstrating the product's gamified learning experience.

The Attribution Challenge Nobody Talks About

Here's a problem that's specific to B2C apps but affects nearly everyone in the space: when someone sees your product in a TikTok video, they often don't click the link. They just Google your app name.

Barak explained the challenge:

"People watch a video on TikTok or Instagram and they hear about Cody... they not necessarily click on a link in a video, they just Google Cody. And then we lose tracking and it's hard to monitor where the users are coming from."

This creates a massive attribution problem. You're spending money on influencers, but you can't definitively prove which ones are driving results because users are entering through organic search.

Coddy solved this with what Barak calls an "attribution graph" – a multi-layered system using UTM parameters, onboarding survey questions, and coupon codes at checkout. During onboarding, they simply ask users how they heard about Coddy, and if they select Instagram, a follow-up asks which specific channel.

It's not perfect. As with any B2C attribution system, there's still about 20-30% of users that can't be definitively attributed. For those users, Coddy distributes them proportionally across known channels based on the patterns they see from attributed users.

YouTube: The Long Game

While short-form content drove Coddy's initial explosive growth, Barak is increasingly excited about long-form YouTube videos as a channel. The reason? They're evergreen.

"When you do an ad on YouTube, it's like an asset that stays there for years gathering more views, more customers. So it's like a longer term investment."

The challenge is that CoolScript 2, which works brilliantly for short-form content, can't accurately predict returns on long-form videos. The payback period is too extended, and the cumulative effect too difficult to model. But Barak sees the value in building these assets anyway, even if they're harder to measure in the short term.

The Negotiation Advantage

About 90% of Coddy's influencer deals are structured as flat fees rather than performance-based payments. This might seem counterintuitive – wouldn't you want to only pay for results?

But flat fees actually work in Coddy's favor when they have accurate ROI predictions:

"Usually we pay only after they publish the video. Not that it matters that much, but it's like, I don't know, helps them be on schedule."

Having CoolScript 2's predictions gives Coddy leverage in negotiations. They know their maximum viable payment before entering discussions, which means they can negotiate confidently and walk away from deals that don't make financial sense.

The system isn't foolproof. Barak acknowledged that roughly 9 out of 10 influencer campaigns don't break even on their own. But the 10th one covers the losses from all the others and generates profit on top. That's why accurate prediction is so critical – it helps maximize the hit rate and minimize losses on the misses.

Key Takeaways

• Influencer marketing can work as a primary growth channel for B2C apps, but it requires a systematic, data-driven approach rather than treating it as experimental spend
• Build prediction systems early: CoolScript 2 gave Coddy negotiation leverage and prevented wasted spend on low-potential partnerships
• Average views matter more than follower count when evaluating influencer channels
• Solving the attribution problem requires multiple tracking methods: UTM parameters, onboarding surveys, and coupon codes all working together
• Long-form YouTube content acts as an evergreen acquisition channel, similar to SEO, continuing to deliver users long after publication
• Most influencer campaigns will lose money, but the winners need to cover the losses and generate profit – which makes accurate prediction critical

Listen to the full conversation with Barak Glanz on the Levels Podcast to hear more about Coddy's journey from students working on a side project to a platform serving 2 million users.