Calculate retained customers over time

[FearDerBeard]

I have a transactional table that records a customer record over time.  Customers come in and out over time and I'd like to calculate how many are kept from some starting point.

For example, if you arrive in period 1, did you return in period 2 or subsequent periods.

Table:

CustomerID	PeriodID
1234	1
1345	1
2567	1
3678	1
4890	1
5789	1
6890	1
5421	1
1234	2
1345	2
2567	2
1111	2
2222	2
6890	2
3678	3
1111	3

Outcome:

First Period	Period	Retained Customers	% Retained
1	1	8	100%
1	2	3	38%
1	3	1	13%

I tried a bi-directional join through a table of all unique customerid's and a copy of the original table but the results were incorrect.  Hope this makes sense.  Thanks.

[v-yulgu-msft · Accepted solution]

Hi @FearDerBeard,

In your sceario, do you only need to count the retained customers start from PeriodId 1?

If so, please refer to below DAX formula to see whether it works for you.

Column1 =
LOOKUPVALUE (
    'retained over time'[CustomerID],
    'retained over time'[CustomerID], 'retained over time'[CustomerID],
    'retained over time'[PeriodID], 'retained over time'[PeriodID] - 1
)

Column2 =
IF (
    'retained over time'[PeriodID] = 1,
    CALCULATE (
        DISTINCTCOUNT ( 'retained over time'[CustomerID] ),
        FILTER ( 'retained over time', 'retained over time'[PeriodID] = 1 )
    ),
    CALCULATE (
        COUNT ( 'retained over time'[Column1] ),
        ALLEXCEPT ( 'retained over time', 'retained over time'[PeriodID] )
    )
)

Regards,
Yuliana Gu

[Greg_Deckler]

Perhaps something along the lines of this:

Customers Retained = IF(SUM(Customers[PeriodID])=1,COUNT([PeriodID]),COUNT([PeriodID]) - CALCULATE(COUNT([PeriodID]),ALL(Customers[CustomerID]),FILTER(Customers,[PeriodID]=[PeriodID]-1)))

That isn't the entire solution because it doesn't match customer id's but may get you close.