Well done! Nested CTEs are very handy when you want to compute complex statistics in one query. For example, you can use nested CTEs to compute the average of some averages. Look at the query below:

WITH salesman_sold_items AS (
  SELECT 
    salesman_id,
    city_id,
    SUM(items_sold) AS total_items_sold
  FROM daily_sales d
  JOIN salesman s
    ON d.salesman_id = s.id
  GROUP BY salesman_id, city_id
),

city_average AS (
  SELECT 
    city_id,
    AVG(total_items_sold) AS city_avg
  FROM salesman_sold_items
  GROUP BY city_id
)

SELECT AVG(city_avg)
FROM city_average

The first CTE, salesman_sold_items, computes the total number of items sold by each salesman. The second CTE, city_average, computes the city-level average. It's a very important performance metric. Finally, the outer query computes the overall average value for all city averages.

Notice that computing the overall average per salesman would render a completely different result.