Cohort Analysis by First Product

I have tried to tackle this myself, but I'm hitting my usual noob-ish problem that I can't seem to conceptualise how I'd break this problem down to solve it in SQL (BigQuery).

Effectively, I am trying to ascertain how customers proceed to purchase products, depending on their initial product — but I'd like to understand both what the n+1 matrix is as well as the n+x drilldown is for a particular starting product.

For example, if I have the following data:

Customer   Month    Product     Revenue
--------   -----    -------     -------
001        2019-01  Product-A   1.00
001        2019-02  Product-B   5.00
001        2019-02  Product-A   2.50
001        2019-03  Product-C   10.00
002        2019-01  Product-B   3.00
002        2019-02  Product-A   4.00
003        2019-02  Product-A   1.50
003        2019-05  Product-B   2.10

I'd expect the n+1 matrix (i.e., after product-x, how many customers bought product-y next?) to look like (x there to represent NULL):

           Product-A  Product-B  Product-C
Product-A     x          2          0 
Product-B     1          x          1
Product-C     0          0          x

...but I'd also want to drill into a particular product to understand how different customers started from Product-X. For example, if we were looking at the Product-A cohorts (i.e., after using Product-A, in what order did they use subsequent products?):

Customer  Product-B   Product-C
001       2           3
003       2

I have a feeling that I need to create the latter query to aggregate to create the former, and I want to say some clever use of PARTITIONing and using ROW_NUMBER would be my starting point, but I just cannot wrap my head around how I'd start to tackle this.

Any help or pointers would be super-appreciated!

Answering your first question only...

Your intuition for PARTITION BY and ROW_NUMBER is good!

with 
data as (
  select *,row_number() over (partition by Customer order by Month ASC) as rn 
  from <dataset>.<table>
),
first_product as (
  select * from data where rn = 1
),
second_product as (
  select * from data where rn = 2
),
joined as (
  select first_product.Product, second_product.Product, count(*) as ct
  from first_product
  left join second_product using(Customer)
  group by 1,2
)
select * from joined

Note, the above query does not pivot (that is a different question imo, and can usually be solved easier in a spreadsheet or BI application, especially as we are not aware of how many Products you might have).

Also note, I hope you have a more specific date type than 'month' so you can actually determine the true order (in your example, Customer 001 has 2 products in Feb, making it hard to tell which is actually 2nd).

Final thought: I am curious as to why your sample output has x's. If a customer purchases A, and then purchases A again, are you looking for the next non-A purchase to answer your question? Just something to think about for your business case.