Goal: Create a summary table to evaluate marketing campaign performance
Output: data/marketing_summary_by_date_channel_campaign.csv
Preview:
| date | media_source | campaign | installs | revenue | costs | impressions | clicks |
|---|---|---|---|---|---|---|---|
| 2024-03-01 | tiktokglobal_int | tt_campaign_1 | 254 | 0.00 | 373.68 | 55620 | 2201 |
| 2024-03-01 | tiktokglobal_int | tt_campaign_4 | 93 | 0.00 | 130.91 | 27650 | 675 |
| 2024-03-02 | tiktokglobal_int | tt_campaign_1 | 218 | 11.82 | 370.28 | 85610 | 1470 |
Goal: Identify which campaigns on the tiktokglobal_int channel are effective based on the criterion ROAS_7 > 0.18
Output: data/tiktok campaign eval.csv
Preview:
| campaign | costs | total_revenue_7 | roas_7 | status |
|---|---|---|---|---|
| tt_campaign_4 | 892.36 | 400.75 | 0.4491 | эффективная |
| tt_campaign_3 | 4893.33 | 1316.01 | 0.2689 | эффективная |
| tt_campaign_1 | 9331.25 | 2191.31 | 0.2348 | эффективная |
| tt_campaign_2 | 2879.35 | 440.45 | 0.1530 | неэффективная |
Goal: Check the correctness of user attribution for the googleadwords_int channel.
Results:
- Total users: 29,767
- With attribution: 20,711
- Without attribution: 9,056 (30.42%)
- With googleadwords_int: 1,457 (7.03% of attributed users)
Conclusions:
- 30% of users lack attribution, which may indicate an issue in install tracking or attribution setup.
- The googleadwords_int share is only 7%, which seems unusually low.
- Potential data loss or incorrect processing for googleadwords_int users — further investigation recommended.
- Python, pandas
- SQLite (sqlite3)
- Jupyter Notebook