Welcome to the forum.

I see from your title that you suspect a normal distribution. In life data rarely conforms exactly to a convenient mathematical distribution; and your first job will have to be to find out what your data 'looks like' when graphed. My graphs below show what might happen.

You need to collect lots of actual spend values. These are your x values.

Group them and count how many in each group. Those frequencies are your y values. It's best to make each group the same width along the x axis eg 0-20; 20-40; 40-60 ..... rather than 0-20; 20-30; 30-35 ......

You'll also have to decide which group to put a value of exactly 20 in. Does it go in 0-20 or 20-40. In practice it shouldn't matter much as long as you only count it once. If you use the cents as well as the $ that should push nearly all borderline cases into the higher group, so you might as well make that the rule. The proper mathematical way to describe this is:

The more data you collect, the better your model will be but, obviously, at some stage you have to say "Right, I think I've got enough evidence; now I'll try the graph."

Once you have a graph post it back here to for an opinion about the resulting distribution. There are limits on image posting for Novices so you may have to post the data instead. Or tell us a link to find it.

If the data does conform approximately to a 'Normal' then there is lots of analysis that can be done, but this won't be valid until we know the distribution. So start collecting data

Bob

]]>Welcome to the forum. It would help to see typical data. You can make it up as long as it has the same flavor as the real data you have.

]]>Thanks in advance for any guidance!

]]>