## 6734Re: [webanalytics] Re: basic math question - percentages

Expand Messages
• Jun 2, 2006
Maureen,

What an excellent point!!!
Eric would be nodding his head as well at this point as I know this his
biggest pet peeve. You **have** to and must look at Targets (and in Eric
Peterson's land "threshold values"). The idea is to benchmark or come up
with this "threshold value"and measure your percentage value's against it.
Now, you can curb a premature jubilation :)
Net, net, net, have another column called threshold (based on your targets
and projections) and measure the percent of a percent value against it.

Good day,

MM

On 6/2/06, matpflum <matpflum@...> wrote:
>
> Stephane,
> I am glad you brought that point up. Besides the math logic and
> terminology, that was actually my point in asking the question. You
> could see conversions going from 1.0% to 1.2% and proudly announce
> that there was a 20% increase in conversions. But what if the goal
> is 1.5% - the 20% good news has to be taken with a grain of salt. It
> drives me crazy when the marketing types take only the 20% increase
> (or whatever the biggest number is), spin that to upper management,
> and before you know it, somebody is wildly happy about a 20%
> conversion rate. Because I have gotten burned by this in the past
> (and it seems you have also), I favor reporting on "differences" in
> these instances and not percentage increases.
> Maureen
>
> --- In webanalytics@yahoogroups.com, "S.Hamel" <shamel67_news@...>
> wrote:
> >
> > Ho boy do I love Web analytics and this newsgroup!
> >
> > I have seen so many times people running around with a single
> number and
> should be
> > printed in all Web analytics manuals. What's important is the
> interpretation
> > of the numbers, not the numbers themselves. This exemlpe also
> highlight how
> > important it is to look at the numbers from various angles. In your
> example,
> > the average sales price could be the same, say 25\$, but the
> revenues for the
> > month would be 40% lower! (10*25\$ vs 6*25\$)
> >
> > That's why a web analytics solution doesn't do any magic, it's the
> people
> > using them that makes all the difference! :)
> >
> > Cheers,
> > Stephane
> >
> > ----- Original Message -----
> > From: "Peter Michael Sopp" <peter.sopp@...>
> > To: <webanalytics@yahoogroups.com>
> > Sent: Thursday, June 01, 2006 4:56 AM
> > Subject: Re: [webanalytics] Re: basic math question - percentages
> >
> >
> > > I'm just a poor sociologist not a mathematician. Nevertheless I'm
> hoping
> > > I could clarify the points already mentioned by others.
> > > The main point is to understand that the result of a substraction
> is a
> > > difference and not a percentage.
> > > The difference is April - March, so 1% - 1.2% = 0.2%-points. The
> result
> > > of a substraction is always a absolute number and never a
> percentage!
> > > To calculate the percentage growth or percentage increase you
> need a base:
> > > (April-March)/March = simple growth rate. Now you can multiply it
> by 100
> > > and you get 20% (= growth rate in percent: The difference of the
> > > percentages is related to the percentage of March).
> > > Let's give an interesting example:
> > > a1t1 = absolute number of visitors in March
> > > a2t1 = absolute number of buyers in March
> > > pt1 = percentage buyers March (buyers related to visitors:
> > > (a2t1/a1t1)*100)
> > > a1t2 = absolute number of visitors in April
> > > a2t2 = absolute number of buyers in April
> > > pt2 = percentage buyers April (buyers related to visitors:
> > > (a2t1/a1t1)*100)
> > > Now the following could happens: The number of visitors as well
> as the
> > > number of buyers decrease. Nevertheless the growth rate is
> positive
> > > because relativly more visitors has converted to buyers in April
> as in
> > > March.
> > > For example: In March you have a simple conversion rate of
> > > (10/1000)*100 = 1% and in April (6/500)*100=1.2%.
> > > The conclusion is - well that's up to you!
> > > Hope this was helpful
> > > Peter
> > >
> > > matpflum schrieb:
> > >> I have always struggled with the logic behind this seemingly
> simple 0
> > >> rate was 1% in March and is 1.2% in April, that the rate
> increased
> > >> by .2% and not 20%? So it's a simple subtraction. I'm not a
> > >> mathematician, but I think that it is wrong to compute a
> percentage
> > >> of a percentage because the percentages were calculated off of
> > >> different bases. Are there any mathematicians who can explain
> which
> > >> way is correct? Thanks.
> > >> Maureen
> > >>
> > >>
> > >>
> > >
> > >
> > >
> > >
> > >
> > > ---------------------------------------
> > > The Web Analytics Forum
> > > Founded by Eric T. Peterson (www.webanalyticsdemystified.com)
> > > Moderated by the Web Analytics Association
> > > (www.webanalyticsassociation.org)
> > > Email moderators at: webanalytics-moderators@yahoogroups.com
> > > Yahoo! Groups Links
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
>
>
>
>
>
>
>
>
>
>
> ---------------------------------------
> The Web Analytics Forum
> Founded by Eric T. Peterson (www.webanalyticsdemystified.com)
> Moderated by the Web Analytics Association (
> www.webanalyticsassociation.org)
> Email moderators at: webanalytics-moderators@yahoogroups.com