Problem: Manually generated CPO reports and did not see the whole picture of conversions
Posted: Sat Apr 19, 2025 8:47 am
At the beginning of our work with the client, at the end of each week we manually filled out a report in Google Sheets that showed expenses and CPO for Yandex.Direct and Google Ads channels.
To generate such a report, we needed to extract data on advertising expenses from the Yandex.Direct and Google Ads advertising accounts, and then compare it with the conversion data from Yandex.Metrica.
The report had several drawbacks.
1. Didn't show daily dynamics. Dynamics of expenses and orders by forex data days helps to understand on which days customers buy goods better, and on which days - worse. Plus, it allows you to notice in time that sales suddenly start to fall, and take measures in time: make changes to automatic strategies, increase or, conversely, decrease daily limits, add rate adjustments.
2. There was no data broken down by campaigns and products. Filling out such a report manually every week, let alone every day, would be very labor-intensive. Campaign names in many UTM tags might not match the actual campaign names, and the campaign_id parameter — a unique identifier that could be used to easily compare data — was not added everywhere.
Therefore, we conducted deep analytics across campaigns only in those moments when there was a clear decline in orders, and CPO was growing. In all other cases, we focused more on some indirect indicators such as the bounce rate, viewing depth, or conversion in Google Ads.
3. Conversions in Google Ads and Yandex.Metrica are quite different. For example, we saw that some product was sagging in CPO in Google Ads, we turned it off. But it turned out that in Yandex.Metrica, on the contrary, it was converting and had a low CPO. As a result, sales fell, the client was unhappy.
To generate such a report, we needed to extract data on advertising expenses from the Yandex.Direct and Google Ads advertising accounts, and then compare it with the conversion data from Yandex.Metrica.
The report had several drawbacks.
1. Didn't show daily dynamics. Dynamics of expenses and orders by forex data days helps to understand on which days customers buy goods better, and on which days - worse. Plus, it allows you to notice in time that sales suddenly start to fall, and take measures in time: make changes to automatic strategies, increase or, conversely, decrease daily limits, add rate adjustments.
2. There was no data broken down by campaigns and products. Filling out such a report manually every week, let alone every day, would be very labor-intensive. Campaign names in many UTM tags might not match the actual campaign names, and the campaign_id parameter — a unique identifier that could be used to easily compare data — was not added everywhere.
Therefore, we conducted deep analytics across campaigns only in those moments when there was a clear decline in orders, and CPO was growing. In all other cases, we focused more on some indirect indicators such as the bounce rate, viewing depth, or conversion in Google Ads.
3. Conversions in Google Ads and Yandex.Metrica are quite different. For example, we saw that some product was sagging in CPO in Google Ads, we turned it off. But it turned out that in Yandex.Metrica, on the contrary, it was converting and had a low CPO. As a result, sales fell, the client was unhappy.