<img src="static/logo.png" alt="drawing" width="500"/>
Our system ingests search term data from Google Ads API into a PostgreSQL database, via an AWS S3 Data Lake.
Once ingested we score each search term with its Return On Ad Spend (ROAS).
ROAS = conversion value / cost
-
Some CSVs have been given (campaigns.csv, adgroups.csv and search_terms.csv), load them into a database.
-
Create some private end points to return the Top 10 Search Terms by ROAS for a campaign
structure_valueor adgroupalias.
There are tools that can help you load CSVs into a database table easily.
The Top 10 Search Terms should be based on aggregated cost and conversion value data e.g:
| structure_value | search_term | total_cost | total_conversion_value | roas |
|---|---|---|---|---|
| night watch | black mens bike | 5 | 100 | 20 |
| ranger | women's high rider mountain bike with lock | 10 | 105 | 10 |
| alias | search_term | total_cost | total_conversion_value | roas |
|---|---|---|---|---|
| Shift - Long - Name 1 | iphone refurbished | 6 | 340 | 56.7 |
| Shift - Long - Name 2 | google pixel 2 brand new black | 13 | 105 | 8.1 |
We really value neatness and things being put in place to aid local development and continuous integration.
Needless to say, we like tests.
Please fork this repo to complete the challenge, once done email back link to your repo.
Good luck we are rooting for you!