Is there an efficient way to calculate number of rows since a change in a categorical variable?

Question

I have a table that records a daily state for a user. I would like to calculate, for every row, the number of rows (base 1) the user has been in the same state.

I have a workable solution, but I suspect it is not very performant given the use of three window functions. This script is run over a table of c.50million rows and 10 columns.

Please could someone recommend a more efficient means of doing the following?

Current Solution

-- return the state in the previous day for a user
with lagged_data as (
  select 
    user_id
    ,day
    ,state_end_of_day
    ,lag(state_end_of_day) over (partition by user_id order by day) as state_end_of_day_lag
    ,case when state_end_of_day_lag = state_end_of_day then 0 else 1 end as state_change_flag
  from 
    stg_user_states_daily
  )

-- use cumulative sum of the change flag to assign groups for each period in a state
, state_group_index as (
  select 
    *
    , sum(state_change_flag) over (partition by user_id order by day rows between unbounded preceding and current row) as state_group_index 
  from 
    lagged_data
  )

-- count number of days user has been in current state on each day
, days_in_state as (
  select
    *
    ,sum(1) over (partition by user_id, state_group_index order by day rows between unbounded preceding and current row) as days_in_current_state
  from 
    state_group_index
  )

select
  user_id
  ,day
  ,state_end_of_day
  ,days_in_current_state
from 
  days_in_state