MVT Doesn't Work for Marketing

By Mario Fantoni, President, OfferMaxima, Inc.

Multivariate Testing (MVT) doesn't work for marketing. It needs to be radically
modified in order to optimize marketing campaigns. How do we know? Because
in more than a 200 MVT marketing optimization projects - and nobody can
claim to have our track record - we verified this over and over again. Here are
the facts:

• From June of 2003 to December of 2008, my company,OfferMaxima, Inc.,
  performed 200 optimizations for 60 clients.

• In the first three optimizations we used pure MVT and we achieved
  increases of response rate of 1%, 6% and -5% (that is right, a decrease of 5%)

• After these three learning projects, we made an important discovery we used
  in all optimizations following the first three. We used a modified version of MVT
  and achieved an average response rate 30-50% higher than the response rate
  of a control piece tested simultaneously with the optimized one.

Of course, the success of MVT applied to the manufacturing industry has been
verified thousands of times. Innumerable cars, television sets and many types
of hard products have been designed and optimized with MVT. Why then didn't anybody, apply MVT to marketing until 2004, the year we claimed that we had
found a way to dramatically increase advertising response using MVT? Because
MVT experts didn't have the marketing knowledge to do it. It is as simple as that.
To make MVT work for marketing it takes the mind of both an engineer and a marketer. And the luck to discover the secret that we found back in 2004.

So here is the bottom line: don't use MVT for marketing. It doesn't work.
If someone claims it does, they are wrong.

If you want to know more about the secret that makes MVT work for marketing,
contact us.

Back