### Presentation Techniques Can Confuse!

Bookassist's Des O'Mahony points out that some graphs are not quite like others...

Sometimes you see marketing performance presentations with what look like very impressive results! Until you look more closely.

Suppose someone is measuring marketing performance for your hotel in terms of conversion of lookers to bookers, i.e. the number of visitors who come to your site (green below) compared to those who eventually book (red below). If the following two graphs of this were shown to you, which of them looks the more impressive performance at first glance?

The one on the right looks like a much more impressive result!

*But in fact the information in both graphs is exactly the same*- there are 1150 visits (or lookers) and 11 bookers shown in each example graph, a conversion performance of 0.96%.

If it's not obvious to you that the graphs above actually show the same information, then an explanation is needed. It goes like this.

The graph that looks "worse", on the left, has a

**linear**scale. This is the usual sort of scale you see in graphs, a scale that goes up in even steps 1,2,3 etc. Each step is the same size, in this case steps of 250. This sort of scale reflects the natural way that people think, so its the "true" or natural scale to use to compare the size of two things. 11 bookers looks small on this, 1150 visitors looks way bigger, but that's what you would expect.

The graph that looks "better", on the right, has a

**logarithmic**or log scale. Each step goes up as a

*power of ten*, so instead of going upwards as 1,2,3 like the linear scale, the scale goes up as 10

*to the power of*1, 10

*to the power of*2, ten

*to the power of*3 etc - 10^1 = 10, 10^2 = 100, 10^3 = 1000. This means the steps are 1, 10, 100, 1000 etc. The steps are not at all the same size and get bigger by a factor of ten each step upwards you take. On this log scale, 11 bookers is just over 10, the first step, while 1150 is just slightly above 1000, the third step, making the bookers line look almost like a third of the lookers line at first glance.

Log scales have the effect of amplifying very small numbers, while reducing the apparent size of very large numbers.

Log scales are very useful and important in mathematics and science. But when used in marketing, they can sometimes mask the true result and make poor results seem far better than they are. Beware of how you interpret results that are shown to you in graphical form. When you have an option, always make sure you view data on a linear scale to get a clear interpretation.

Labels: marketing, statistics, strategy

## 0 Comments:

Post a Comment

Subscribe to Post Comments [Atom]

<< Home