The Attribution Problem: How B2C Apps Track Users Who Don't Click Links

You've just spent thousands of dollars on an influencer campaign. The video goes live, views roll in, and new users start signing up. Success, right?

Not so fast. Where did those users actually come from? Was it the influencer you paid? Organic search? A Facebook ad they saw last week? For B2C apps, especially those running multiple marketing channels simultaneously, attribution is one of the hardest problems to solve.

On the Levels Podcast, Barak Glanz, CMO of Coddy, shared how his team built a multi-layered tracking system to solve this challenge. As a web-based code-learning platform that scaled to $1M ARR primarily through influencer marketing, getting attribution right wasn't optional—it was essential for knowing which channels to scale and which to cut.

The Core Problem: Users Don't Click Links

Here's what makes attribution so challenging for B2C apps: people don't behave the way your tracking systems expect them to.

When someone watches a TikTok video featuring your app, the ideal scenario is they click the link in bio, land on your site with UTM parameters intact, and sign up. Clean data, perfect attribution.

Reality is messier. Much messier.

"A lot of our users don't click a link. So oftentimes they don't have like a UTM or hyperparameters, nothing saved in their cookies."

Users see your app mentioned in a video, close TikTok, open their browser, and Google your app name. Or they hear about it, make a mental note, and sign up three days later after seeing another mention somewhere else. The click-through path you've carefully architected? They've bypassed it entirely.

This behavior creates what Barak calls a "big challenge" for understanding channel effectiveness. When you're spending money across multiple influencers and performance marketing channels, not knowing where users come from makes it nearly impossible to optimize spend.

The Attribution Graph: A Three-Layered System

Rather than relying on a single tracking method, Coddy built what Barak describes as an "attribution graph"—multiple overlapping systems that work together to capture user sources.

The three layers are:

Layer 1: UTM Parameters The standard approach. When users do click through from an influencer link or ad, UTM parameters track the source in their cookies. This works great for the users who follow the expected path, but that's a minority.

Layer 2: Onboarding Survey During the signup process, Coddy asks new users a simple question: "How did you hear about us?"

"When they register during the onboarding poll, I ask them, how did you hear about us? And then they click on Instagram and then I say, okay, which channel on Instagram?"

This two-step approach is critical. First identifying the platform (Instagram, TikTok, YouTube), then drilling down to the specific creator. It captures users who bypassed link tracking entirely.

Layer 3: Coupon Codes At checkout, when users subscribe to Coddy's premium plans, they can enter coupon codes. Influencers often share unique codes, creating another attribution touchpoint that survives even if all other tracking fails.

By layering these three methods, Coddy captures attribution data across multiple user journey variations. Someone might not have UTM parameters but answers the survey question. Another might skip the survey but uses an influencer's coupon code at checkout.

The 20-30% Gray Zone

Even with a sophisticated attribution system, perfect tracking remains impossible. There's always a cohort of users who can't be definitively attributed to any channel.

"I guess like 20, I would say like 20-30% you don't know who they are."

For this gray zone, Coddy takes a statistical approach: distributing unattributed users proportionally across known channels based on the patterns they see from successfully attributed users. If 40% of attributed users came from Instagram influencers and 30% from Facebook ads, they apply those same ratios to the unknown cohort.

It's not perfect, but it's reasonable. The key is recognizing that chasing 100% attribution accuracy is a losing battle—better to build a system that's 70-80% accurate and use statistical modeling for the gaps.

Why Web Apps Face Extra Attribution Challenges

Barak pointed out something that mobile app founders might not fully appreciate: web-based apps face even harder attribution problems than native mobile apps.

"Coddy is a web app. We don't even have a mobile app yet... so we can't use AppsFlyer or anything like that. So attribution is even harder. On mobile app it's like, it's still hard but it's easier."

Mobile attribution platforms like AppsFlyer provide sophisticated device-level tracking that's simply not available in web environments. Browser cookies get cleared, users switch devices, and privacy protections limit what can be tracked.

For B2C SaaS companies operating primarily on the web, this makes building custom attribution systems not just helpful but essential. The off-the-shelf tools designed for mobile won't solve your problems.

The Business Impact of Better Attribution

Why does all this matter? Because attribution accuracy directly impacts your ability to scale profitably.

Barak shared that in influencer marketing, the typical pattern is that 9 out of 10 campaigns lose money, while the 10th one covers all the losses and generates profit. But if you can't accurately attribute which influencer drove which users, you can't identify your winners and losers.

With Coddy's attribution system feeding into their CoolScript 2 algorithm, they could compare predicted ROI against actual performance. This closed feedback loop let them continuously improve their influencer selection criteria and negotiation strategies.

The attribution system also enabled smarter retargeting. When users hit Coddy's energy limit (their freemium restriction) and churned, knowing their source channel helped inform which platforms to use for bringing them back.

Building Your Own Attribution System

If you're running a B2C app with multiple acquisition channels, here are the practical steps to build a similar system:

Start with the onboarding survey. It's the lowest-effort, highest-return addition you can make. A simple "How did you hear about us?" dropdown with a follow-up for specifics captures data that UTM parameters miss.

Layer in coupon code tracking at conversion points. Whether it's checkout, subscription upgrades, or premium feature purchases, unique codes provide a secondary attribution signal that's hard to game or lose.

Accept the gray zone exists. Rather than trying to track every single user perfectly, build a system that captures 70-80% accurately and use statistical distribution for the rest. The marginal effort to move from 80% to 90% attribution accuracy rarely justifies the engineering investment.

For web-based B2C companies specifically, Barak's experience shows that custom-built solutions often outperform off-the-shelf tools designed primarily for mobile apps. The constraints are different, and the tracking mechanisms need to account for browser-based behaviors.

Key Takeaways

• Most B2C users don't click tracking links—they Google your app name after seeing it mentioned, breaking standard attribution flows
• Build an "attribution graph" with multiple overlapping tracking methods: UTM parameters, onboarding surveys, and coupon codes
• Accept that 20-30% of users will remain unattributed and use statistical modeling to distribute them across known channels
• Web-based B2C apps face harder attribution challenges than mobile apps since tools like AppsFlyer don't work in browser environments
• Better attribution enables you to identify which influencers and channels actually drive profitable growth, turning qualitative guesses into quantitative decisions
• Onboarding surveys provide the highest ROI for capturing attribution data that traditional tracking methods miss

Listen to the full conversation with Barak Glanz on the Levels Podcast to hear more about how Coddy scaled to 2 million users and $1M ARR.