June 40 sign ups July 50 sign ups

to estimate August I divide June / July (40/50) = .8

I subtract .8 from 1 giving me a growth rate of 20% (.2).

I multiple 50 * 1.2 to get an expected sign up of 60

I understand the above..now look at this

September i had 881 signups October i had 47 signups

If I want to estimate November signups this is what I'm doing

881 / 47 = 18.74 1 - 18.74 = -17.74

do i multiple 47 * -17.74? i don't think its right...can someone help me with my math here.


migrated from stackoverflow.com Nov 2 '11 at 21:33

This question came from our site for professional and enthusiast programmers.

  • $\begingroup$ this is a software/programming site $\endgroup$ – leonbloy Nov 2 '11 at 21:27

More normally, given just the two months June and July, if you want to assume a percentage growth rate you would say August=July (July/June)=50(1.25)=62.5. You could also assume a linear growth rate and say August=July+(July-June)=50+10=60. Now that you have four months of data you can do various things. If I look at $$\begin {array}{r r} June&40\\July&50\\August&881\\September&47 \end {array}$$ I wonder if the data is right with August so high and I give up trying to forecast. You can do various kinds of fits (and Excel will provide an answer), but I don't think any will be meaningful.


Actually to compute august you've done: July * rate = July * (July/June)
You do exactly the same if you want the same linear progression as between the last two months:
November = October * (October / September) = 47.0 * 47.0 / 881.0 = 2.5
Obviously this is very very basic prediction...

  • $\begingroup$ That's not linear extrapolation but exponential $\endgroup$ – Henry Nov 2 '11 at 21:43
  • $\begingroup$ arf you're totally right... shame on me $\endgroup$ – PierrOz Nov 9 '11 at 10:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy