Can you actually track ROI across markets? (a founder's honest attempt)

Кейсы и истории успеха
1

We’re a tech startup with really strong roots in Russia, and we just launched into three new European markets. Naturally, one of my first questions was: “Okay, how do I actually measure whether these influencer partnerships are working across all these different places?”

Short answer: it’s way harder than it sounds. Long answer: keep reading.

The challenge is that we have totally different unit economics across markets. A customer in Russia might cost us $12 to acquire via influencer outreach. That same customer in Germany costs us $34 because the market is more competitive and CPA is higher. So when we run a campaign across both, simple ROI math breaks down.

We also realized early on that the time to conversion varies wildly. A Russian customer might convert in 3 days; a German customer might spend 3 weeks considering. So if we just measure “conversions in the first 7 days,” we’re drastically undervaluing German campaigns.

Here’s what we ended up building:

Step 1: Establish market-specific baselines.

Before jumping into influencer campaigns, we tracked how much it cost us to acquire a customer through “normal” channels (paid ads, organic, direct traffic) in each market. That became our benchmark.

Russia: $12 CAC (customer acquisition cost)
Germany: $34 CAC
France: $28 CAC

Influencer cost varies by creator size, but in each market, we’re paying creators between $500-$3,000 per post.

Step 2: Track cohorts, not campaigns.

Instead of just asking “did this post convert?” we built a cohort tracker. Every click from an influencer post is tagged with the creator’s name, market, and date. Then we follow that cohort for 30 days and see who converts, at what cost, and when.

This is where things got interesting. Some creators drove high-intent clicks (converted at 8-12%), others drove high-volume clicks (converted at 2-3%). Both were valuable, but in different ways.

Step 3: Account for attribution complexity.

Here’s where I’ll be honest: we don’t have perfect attribution. A customer might click an influencer post, then see an ad, then finally buy via Google search. Who gets credit?

We use a multi-touch model where the influencer gets 40% credit (they drove initial awareness), the retargeting ad gets 40%, and the final click gets 20%. It’s not perfect, but it’s way more realistic than first-click or last-click attribution.

Step 4: Calculate market-adjusted ROI.

Once we have cohort data and attribution, the math becomes:

(Revenue from cohort - Cost of influencer post - Cost of all downstream ads) / Total investment = ROI

Then we compare that to our baseline CAC in each market. If influencer campaigns are delivering CAC below our baseline, they’re working. If above, they’re not.

What we found:

  • Russia: Influencer campaigns averaging $9 CAC (27% better than baseline). Worth scaling.
  • Germany: $38 CAC (12% worse than baseline). We needed to try different creators or product positioning.
  • France: $22 CAC (21% better than baseline). Sweet spot.

The hard part? Patience.

Some of these insights took 45-60 days to fully emerge because of the time lag in conversions. We almost cut the Germany campaign after 14 days, but we stuck with it and eventually realized the issue wasn’t the influencers—it was that German customers just need more consideration time. Once we accounted for that, the ROI looked different.

Also: external noise is huge. A competitor launching a sale in one market, a currency fluctuation, platform algorithm changes—all of this impacts performance. So we obsessed over removing as many variables as possible and tracking relative performance (vs. baseline) instead of absolute performance.

What we didn’t expect: the best insight wasn’t a number. It was realizing that German influencers underperform US campaigns not because they’re worse at their job, but because German consumers are more skeptical of influencer recommendations. So we shifted the brief from “soft sell” to “transparent review,” and suddenly performance improved.

If you’re tracking ROI across markets, how are you solving for attribution and time-to-conversion differences? And have you found any external factors that completely blindsided you?

2

Спасибо за такой детальный разбор реальной проблемы! Я работаю с брендами, которые расширяются в новые рынки, и вижу эти вопросы постоянно. Ваш фреймворк с market-specific baselines это именно то, что бренды должны делать перед запуском кампаний, а не после.

Мне особенно нравится то, что вы обратили внимание на культурные различия—например, что немецкие потребители более скептичны к инфлюенсерам. Это такой важный инсайт! Я должна обязательно делиться этим с клиентами и криэйторами, с которыми работаю.

3

Это как раз то сообщество должно обсуждать! Слишком много успешных кейсов, недостаточно честных историй про то, что не сработало и почему. Спасибо, что поделились реальными числами и реальными проблемами.

4

Отличный кейс, и я ценю честность по поводу несовершенного атрибуцион. Но давайте обсудим вашу модель подробнее.

Вы используете 40/40/20 split для multi-touch: influencer/retargeting/final touch. Как вы обосновали эти веса? Вы A/B тестировали разные модели, чтобы увидеть, какая наиболее прогностична? Или это было интуитивным выбором?

Также: когда вы говорите о 30-дневном окне для отслеживания конверсий, это означает, что любая конверсия позже 30 дней не учитывается? Или вы всё ещё отслеживаете, но обрезаете на 30 днях для анализа?

И последний вопрос: как вы обрабатываете repeat customers? Если customer из России конвертировалась через influencer кампанию в месяц 1, а потом вернулась в месяц 3, её LTV это что-то совсем другое. Вы смотрите только на первую покупку или на LTV-adjusted ROI?

5

Это ровно с чем я борюсь прямо сейчас. Спасибо за шаг за шагом разбор, потому что я был совсем потерян. Я не знал, что делать с тем фактом, что время конверсии в России vs Европе совсем разное. Я хотел просто измерить всё в один временной полосе и все казалось неправильным.

Один вопрос: когда вы говорите 45-60 дней, чтобы получить полную картину—это значит, что я должен ждать 60 дней перед тем, как масштабировать что-то успешное? Или вы смотрите на промежуточные сигналы и итерируете быстрее?

6

This is exactly the rigor that should separate tier-one agencies from everyone else. Most of my agency competitors still present influencer ROI as simple ROAS, which is completely misleading when you account for time lag and multi-touch dynamics.

Full transparency though: implementing this level of tracking is expensive and time-consuming. You need solid analytics infrastructure, clean data, and discipline. Most small to mid-size agencies frankly can’t afford it. How did you build this infrastructure? Did you use an off-the-shelf platform, custom build, or hybrid?

7

The insight about German consumers requiring a different creative brief is so valuable. This is the kind of learning that’s worth way more than a 10% improvement in CAC. It’s the difference between stumbling onto short-term results and building a sustainable, scalable system.

8

Quick question for when you scale to more markets: how do you keep from drowning in complexity? Because right now you’re managing Russia, Germany, and France. If you go to 10 markets, do you build more automation, hire more analysts, or do you choose your battles and focus on the top 3-4 markets where you’re really investing?

9

I’m reading this as a creator, and honestly, this makes me feel so much better about my job. You’re saying that some creators are “high-intent drivers” and others are “high-volume,” and both are valuable. That’s validation that I don’t need to transform into something I’m not.

I also love that you tracked when you’re paying creators (and how much) vs. the actual results. So many brands are vague about budgets and expectations, then blame creators when results don’t materialize. If every brand operated with this level of clarity, negotiations would be way less painful.

10

The part about cultural differences in how audiences respond to influencer recommendations is huge. It explains why some of my content performs better in certain markets even with the same strategy. Thanks for helping me understand that geography matters just as much as the content itself.

11

This is rigorous, pragmatic work. The multi-touch attribution model you built is solid, and the insight to compare against market-specific baselines rather than a universal benchmark is exactly right.

But here’s my challenge to you: how are you validating that your 40/40/20 split is actually optimal? Like, what if the true contribution is actually 50/30/20, and you’re systematically over-crediting influencers? This matters because it could lead you to over-invest in influencer channels relative to retargeting and organic.

Have you run incrementality tests (holdout groups) to validate this, or are you comfortable with the current model despite the potential for error?

12

The patience element you mentioned is critical. Most founders and CMOs want results in 14-21 days, which is why they make bad decisions about channel performance. You stuck it out for 45-60 days and that’s exactly when the real signal emerges. That discipline is what separates successful expansion from failed attempts.