WITH rev AS (

    SELECT log_date,
           utm_source,
           SUM(CASE WHEN name = 'ads_rev' THEN CAST(object_value AS FLOAT) ELSE 0 END) AS revenue,
           COUNT(DISTINCT user_id) AS dau
    FROM events_log
    WHERE log_date BETWEEN '2022-11-30' AND '2022-12-05'
    GROUP BY log_date, utm_source

),
mean_rev AS (

    SELECT utm_source,
           SUM(revenue) AS ads_rev,
           AVG(dau) AS mean_dau,
           AVG(revenue / CAST(dau AS FLOAT)) AS mean_arpdau
    FROM rev
    GROUP BY utm_source

)
SELECT * FROM mean_rev
ORDER BY mean_arpdau DESC