I recently prepared a web report summarizing a few visitor trends when I realized (in spite of my liberal arts education) that a straight up average of all the stats was not going to do the job. Some referring sites were sending a ton of visits my way with the bottom-feeders sending over only a few dozen.

How could I weight those stats for the sites who were driving the most traffic? The answer: A weighted average (or weighted mean).

More practically, how could I accomplish this computation in Excel? A quick Google search revealed the following formula for Excel:
= SUMPRODUCT(B2:B15, C2:C15) / SUM(B2:B15)

I was looking for trends in pages/visit, % of new visitors and % bounce rate, among many other metrics.

Example (using fake web stats):

So, to find the weighted average for pages/visit, you would compute:
(400×5.3) + (35×2.1) + (230×1.5) …etc. / (400+35+230 …etc.)

Related: Simpson’s Paradox

You almost had me Disney Rewards Visa Card from Chase with your Little Mermaid sticker. Almost.

These guys must have a pretty good profile of me in their system (also, if you’re reading, I like Easy Bake Oven). Now I’m torn between the “Tinker Bell” and “Pixie Dust” card. I may never be able to decide.

I’m not sure how I passed up 0% Fixed Intro APR until May 2009
(Drew Geraets is a compensated endorser).

But, if for some reason, you’d rather not receive very adult things (credit card offers) in the mail (who wouldn’t?) with very childish things on it (hey, I like Mickey and I can rent a car) there is a place you can go (yeah, go you fun-killer!).