Today I saw in a news story that Microsoft is testing a new tool that is called ‘Engagement ROI’. This tool, developed by aQuantive’s Atlas unit, is supposed to measure the all the ways in which buyers interact with a company’s online adverts until the point when they eventually buy something. Basically, Microsoft says that focusing on clicks is not the whole story and the fact that website visitors are actually not clicking on ads doesn’t mean that adverts don’t work.

Ok. This is fascinating. Let’s take a step back and think this through. What do adverts actually do?

Let’s say, I am reading a magazine or watching TV. I will see 100s of ads in both instances. However, I only know a few types of ads that will spur me into action directly. Examples might be:

• Special offers/discounts
• Limit period products/promotions
• Job adverts

Essentially, these are all the adverts that indicate that something is only available for a short period of time. The message is: this is a great deal, if you don’t act now, you will miss the window of opportunity. Dell’s computer adverts are a good example for this style of advertising. They constantly give you the feeling that if you don’t buy the computer now, you will loose out.

All other adverts, I ignore. Or do I? Certainly, they don’t spur me into action. But there is one thing that they do: they position a product in my mental map. This means a couple of things. It means that when I am actually looking for something, I might remember this product and consider it in my purchase decision or give it a higher consideration. If this product is positioned well, I might be prepared to pay more for the product than it is worth objectively. Also, once I have bought a more expensive product, I will also enjoy it more than if the product had been cheaper. Here is a good story about how the perception of price changes experience.

So, what does that mean? In the vast number of circumstances, advertising has the potential to:

• Increase probability that user will buy a product
• Increase price users are prepared to pay
• Increases satisfaction with product, leading to more recommendations, and repeat sales of these products.

I guess what Microsoft is saying is that they want to measure these effects. This is, in theory, a very interesting idea.

So, here is the trick. I was in science a little while back. One thing that I remember very clearly is that when you are measuring an experiment, you need two sets of data. The data set of your experiment and a control data set. You draw conclusions based on the comparison of these two data sets. Ideally, you also have a positive control, so you know what the results look like, when the experiment works.

What I don’t understand is how Microsoft wants to establish the control data sets in these circumstances. I guess it is possible to figure out who has actually looked at which ads online. But how do you know who hasn’t? How do you measure overlap with offline advertising?

I am looking forward to seeing the answer later this year.