Things get even murkier if you're in a B2B or "considered purchase"
environment when it comes to calculating ROI. Personally, whenever a
marketer starts throwing around "ROI," I get nervous. It's a simple
concept, but it's a lot trickier to pull off in Marketing than it is in
some other business areas. Even in Dave's simple example, there are a
lot of factors to take into account. The reality is that, in many, many
cases, to calculate ROI, you have to make various assumptions/estimates.
The more rigorous you are on that front..the more complex and harder to
understand the "model" is.

It's a vicious situation:

1. If you keep the calculation simple, then it's easy to understand, but
may not be all that meaningful, and a lot of people will fault it
because it doesn't take into account x, y, and z.

2. If you take into account x, y, and z, then, as you present the model,
folks eyes will start to glaze over...AND they may start questioning the
assumptions/estimates. You'll wind up debating the calculation rather
than discussing the results

If you step back and ask why someone says, "we need to show ROI," it
usually comes down to one or both of the following:

1. "We want to make people accountable." ROI isn't necessarily a silver
bullet there. Actually, the Eisenberg's and Brian Carroll had a pretty
interesting exchange on that front last year; I wrote about it and
linked to the pertinent articles: http://tinyurl.com/34lqp8

2. "We want to determine what works and what doesn't." In that case,
it's usually better to get down to a level below "revenue" and hone in
on the specific outcomes you were expecting to drive with a given tactic
or strategy. And then measure that. It's a softer deal, though -- it
requires stating up front what you are going to accomplish and getting
agreement that, if you accomplish that, then, ultimately, revenue will
be positively impacted. This goes back to my B2B comment, but it also
plays into any company that's starting to fiddle with social media --
"revenue" results from the successful execution of multiple tactics when
it comes to converting a single prospect into revenue. Truly focusing on
ROI can lead to myopia. It's akin to "If you can't measure it, don't do
it," which sounds good, but is overly simplistic
(http://tinyurl.com/2kzq7e).

I'm not a negative person. Really!

Tim

Tim Wilson

http://www.gilliganondata.com <http://www.gilliganondata.com>
http://www.bulldogsolutions.com <http://www.bulldogsolutions.com/>



________________________________

From: webanalytics@yahoogroups.com [mailto:webanalytics@yahoogroups.com]
On Behalf Of "Wandering" Dave Rhee
Sent: Friday, March 07, 2008 4:21 AM
To: webanalytics@yahoogroups.com
Subject: Re: [webanalytics] ROI Models - High Level Thinking



Hi, Ty,

This is fairly standard MBA stuff -- was there a particular issue
you've already had difficulty with that you were seeking clarification
for?

In general, you want to separate out the fixed / sunk costs (like
hosting a website) from the variable costs (developing additional
pages for a site that's already built). Fixed costs (like buying a
server) might be amortized over time, but headcount (salaries and
related benefits) is a little trickier.

Remember to apply a discount rate to future cash flow projections if
you're calculating net present value -- make sure all dollars, both
expenses and revenues, are compared to each other in the same time
frame (e.g., "today's money").

As far as calcuating return on ad spend, I always find this to be
grossly misleading. For example, saying that "return" on my $1 spent
on a search engine got me $12 "return". This implies a profit of $12,
when in reality, it's only a gross revenue of $12, net revenue of $11.
Worse, if the goods cost me $8 to produce, then that's only $3
profit. On a $1 expenditure, that's still 300% return, but very
different from 1200% return. (And even less as total margin, which
here would be $3 / $9, or 33%.) Further, if the ad that drew people
in was a free shipping offer worth $2, then the cost of this promotion
should also be factored in to the overall sale. $12 revenue, less $8
cost of goods sold, $1 cost of advertising, $2 shipping, leaves only
$1 profit.

That's far from a rigorous example, but I usually see calculations
that don't even take into account these things. (I didn't even
mention variable agency costs, say 15% of ad spend, on top of all that
-- not just for the ad placements and bid management, but also the
creative development fees, etc.) It's like asking how much money you
made on your stock portfolio just by watching the market price, but
neglecting management fees, broker's commissions, taxes, etc. True
only in some not-very-useful-for-daily-decision-making-purposes sort
of way.

WDave Rhee, Moderator
Analytics Country Manager, Germany, OX2 / LBi

On Thu, Mar 6, 2008 at 7:09 PM, Tyler Snouffer <tsnouffer@...
<mailto:tsnouffer%40resource.com> > wrote:
>
>
>
>
>
>
> Can anyone suggest material that speaks to developing ROI models?
> Whitepapers, books, blog entries, etc.
>
> It doesn't necessarily have to be web-related ROI either.
>
> ROI has always been a pretty tough nut for our firm to crack and I'd
> like to do some learnin' before I try to develop any specific model.
>
> If you're interested in sharing how you calculate ROI for your
> projects feel free. I've got the basic calculation, but really
> applying it can be much more complicated than the basic formula.
>
> Thanks
>
> -Ty
>
>





[Non-text portions of this message have been removed]