I'm not sure how to ask this question, so please bear with me. I'm not even sure if it's possible with just this information. This is the only information that we received.
I'm using this equation for rankings:
$$\text{CONSLOSS}+{\text{LOSS}} - {\frac {\text{IN}/\text{OUT}} {\text{IN}/\text{OUT} + 1}}$$
CONSLOSS LOSS MONTHS IN OUT Rank formula
Row1 2 25 25 30 10 26.25
Row2 1 1 25 8000 1000 1.111
Row3 0 0 25 100 90 -0.526315789
Based on the result, Row1 is worse off, which is correct mathematically-speaking.
But for payroll, Row2 is worse than Row1 because in 25 months, it only last 1 month, and it's the last month of the sample. They're adding more weight to that one loss and the fact that their IN & OUT churn is much greater than the other two rows.
I guess my question is... is there a way I can add something to this formula so that Row2 ends up having a higher value because a component was given more weight?
You can experiment with two constants $A$ and $B$. Modify your formula so that you multiply LOSS by $A$ and the fraction by $B$.
Play around until you have values for $A$ and $B$ that make the output look right for your purpose.
These are easy experiments to do in Excel.