Over the last few months we’ve had some interesting questions posed on our LinkedIn group. We plan to share them regularly, to help modellers who are trying to apply the FAST standard.
Below are some of the discussions, and the relevant links to the LinkedIn group.
FAST use the I column if you are looking back one period. How would you go about if you want to lookback up to 6 periods, and take let’s say the average, if available?
This is rather surprising because they are at least as controversial as name ranges.
I prefer doing some of the pre-processing input in the model itself and in this way reduce the number of data entry translations after handover and in this way minimise chances of mistakes.
But is it possible to go without pivot tables if this preprocessing needs to summarise the fixed asset register of more than 4000 items? Perhaps simply using SUMIF? And how do you control the timing in the model if you do use Pivot Tables? The standard flags don’t work. I would be interested hearing anyone’s thoughts on this!
Hello, I am currently building an integrated model (bs/pl & cf) my question relates to how to treat sales and cost by customer/bid. Currently I have inserted a sheet (PL) for each customer and make use of “start” and “end” sheets to sum these x sheets. This however increases the number of sheets considerably. How would this differ in a FAST approach?
Do you have any conventions for rolling up smaller time periods into larger?
A model might have one year as months, then 3 more years as quarters, then 10 or 20 more as years. And for reporting they are combined as years. Do you just refer to the ranges of detail columns, although that results in formulas that cannot be easily copied for consistency? An alternative would be indexing and offset, which gains generality but loses wider comprehension.