Why should half of your marketing fail

The need for a high failure rate is in direct contradiction to many of my conversations. It is also against most people’s approach of trying to do everything right the first time. I was reviewing one of Donald Reinertsen older books, Managing the Design Factory. Reinertsen is simply a great author that takes what I call Geek information and converts to a level of understanding that I even get. I used material from the aforementioned book re-writing it for the purpose of marketing versus software development.

Why should 50% of your marketing fail?

Using the Information Theory: the more probable the event, the less information that is needed. Why is that true? If you receive information that you expect, it contains little information. For example, if you have a targeted message to one person, that information will have a better chance to succeed. If you send the same message to 2 people, you have introduced more risk and less chance to succeed. So rather than try to drive failures out of the process or become more efficient we must introduce large amounts of information and as a result more risk. In fact, that magic number for efficiency is 50%.

To generate that 50% number lets define the 2 extremes. If we want 100%, the information theory states the lower amount of information needed. That means if we “do it right” the first time we have driven all the information out of the process except for a very select audience. If we look at 0% that means that we provide all the information to a very large audience. An analogy that I use for 0% may be a Super Bowl Ad. I am pretty sure at this point that anyone reading this is not contemplating a super bowl ad next year. My failure rate at this point may be high but it is not at either extreme. At least it is at a starting point.

How do we generate this information efficiently (50%)?

Distinguish between failures to generate useful information, which are new failures and those that generate information that we already have, which are the old failures.
Providing tracking information or checklists especially from past experiences. Good accounting of your failures is really more valuable than the description of the most successful work.
The early you test the better.
Use the smallest batch size possible.

In our discussions with small batch size strategies we can think of the process is producing potential errors at a certain rate. If we can test early, we choose to receive these errors when the costs of reacting to them are low. The striking advantages of the small batch size are that information arrived early and our total population of errors remains small because it arrives in more manageable chucks. Of course, the more batch sizes you have the more you waste design resource each iteration incurs extra costs and of course extra cost or path.

These two areas are always in direct conflict and one of the things that need to be weighed is the costs of the trials tested. When the cost of the trial is high fewer iterations will be performed and vice versa. However, frequent iterations can actually be much more valuable than people suspect.

I think this is a very interesting concept and deserves further study. I use this theory in developing Facebook and Google Ads on a regular basis. Seldom are my ads stagnant. They are constantly evolving and change as success rates change. For an example, if you have a campaign that has three or four ads in it, you can constantly evolve these ads to increase hit rates or conversions…but only try to be 50% successful!