Marta Jarzynska
After 10 years of working in digital marketing and user acquisition (UA), I’ve seen the job evolve significantly. Back in the day, it was all about control. You could manually adjust bids, pick specific audiences, and set up campaigns in a way that felt like pulling the right levers to drive growth.
And for a while, that worked, until it didn’t.
It started with Google UAC. Suddenly, many of the levers we relied on were taken away. Meta followed the same path. When I joined Too Good To Go in early 2019, we were still running city-by-city ad sets to match local supply and demand. However, that approach soon had to change. We were encouraged to expand our reach, targeting entire countries and letting the algorithm “figure it out.”
As if losing control wasn’t enough, we also started losing visibility. In 2021, Apple introduced SKAN (SKAdNetwork) as part of its efforts to prioritize user privacy. It became even more difficult to determine what was working. Attribution windows shrank, user-level data disappeared, and the feedback loops we relied on became slower and less effective.
Now, not only were we flying blind in a black box, but we were doing it with half the dashboard lights turned off. The gap between campaign inputs and actual business outcomes widened, and UA teams were left trying to connect the dots with missing pieces.
Once automation took over, creative became the most important lever we still controlled. And yes, strong creative can make a big difference. But it doesn’t last long. People scroll fast, and ad fatigue sets in quickly.
We doubled down on creative assets, building systems, testing, and utilizing user-generated content (UGC) to boost engagement. Sometimes it worked. Other times, we felt like we were just guessing. Creative is important, but it can’t fix a campaign if the underlying targeting is off.
Then it clicked: if creative is the message, signal is the guide that helps platforms deliver it to the right people consistently, over time. In the past, we relied on lookalike audiences. However, with short attribution windows, privacy limitations affecting platform match rates, inconsistent user journeys, and, in many cases, no reliable live signal, that’s no longer enough.
We needed to optimize not just for how cheap a user was to acquire, but also for how valuable they’d be in the long run, and do it early enough to guide the algorithm.
That realization led me to Churney. After years of managing UA from the inside, I wanted to address the issue that kept getting in the way: the disconnect between campaign performance and actual business outcomes.
Churney solves that by helping UA teams focus on real user value. It uses machine learning to predict long-term LTV based on early actions and sends that signal back into Meta, Google, and other platforms. Instead of just getting users who install the app, you get users who stick around, spend money, and bring value.
For the first time, I’m seeing what it looks like when UA and business goals are finally aligned. No more chasing low CPIs that don’t lead to anything. No more guessing what’s working behind the black box.
I believe the future of UA belongs to teams that build smart systems - combining strong creative with quality signals, learning fast, and staying focused on long-term impact. Being on the advertiser side taught me how frustrating and limiting this space can be. Now, helping build solutions and supporting multiple advertisers, I’m more motivated than ever to help others break through those roadblocks.
And full automation is just around the corner. Meta plans to run entire campaigns by 2026 using only a product image and a budget. With creative soon in AI’s hands, UA teams will move from managing campaigns to guiding strategy, ensuring data and signal quality, and interpreting AI-driven insights.
UA isn’t just getting smarter - it’s becoming fully automated. The challenge ahead is adapting to this new landscape while still driving meaningful results - exactly what Churney is built to support.
Your data warehouse has incredible value. Our causal AI helps unlock it.
