Introduction to Advertising II

As mentioned before our next topic is going to deal with measuring advertising effects. First let’s talk about getting data. We’ll get to the evaluation part later. There are three categories of tests we’re concerned with:

1. Pre-tests
2. Post-tests
3. Tracking

Pre-tests are designed to measure any short-term effect of the ad design BEFORE the start of the campaign, whereas post-tests measure the same effects during the course of the campaign. Tracking refers to frequent measurements (mostly interviews or questionnaires) of long-term effects also during the campaign. Be sure to understand the difference between aided and unaided recall tests and recongition tests. We won’t go further into the details, just a word of warning: Don’t use panels for your tracking data. Remember: you want unbiased information!

You may have noticed that I was very quick to dispose of the subject of getting data. That’s because we have some work to do. We’ll now start talking about advertising response functions.

Response functions are supposed to map a dependency

choice variables effect variables

based on either educated guessing or empirical data.

Recall that the five choice variables in advertising were budget, timing, pressure, ad design and choice of media. Our effect variables could be awareness, attitude, share of users…

In 1985, Petty and Cacioppo studied frequency effects of advertising on attitude and came up with a very straight forward result: The number of expositions being the x-axis, attitude the y-axis, their function can be modelled as a negative parabola shifted a bit to the right. Why? Well. Their idea is that for a contact person to start learning, he/she must be exposed to the ad for a few times. Once the learning begins, there will be (hopefully) an “increase” in the attitude towards the product (not quite sure how to measure that one…). However, a few expositions later we will start to observe a phenomenon called reactance. The attitude starts decreasing until it eventually reaches zero or even a negative value. This, of course, is a very simple model. That needn’t be bad though.

16 years later, Hallemann presented an algebraic approach to pretty much the same question; this time considering TV and print ads. In his proposed equation we can observe an influence of the past period’s value of our effect variable on its current value. This dependency is considered a “carry-over effect”. The idea is rather simple: If we consider months as our time periods, a person watching our TV ads will (again: hopefully) not have forgotten everything he/she has learned from our ads in March just because it’s the 1st of April. Generally, in an analytic function this effect is modelled by adding last period’s value of our target variable weighted by a so-called carry-over coefficient to the term.

Similar to the carry-over effect is yet another dynamic principle, the “lag effect”. Mostly attributed to Simon Broadbent in his “adstock model”, the lag effect can be compared to the half-life of a radioactive material. The two effects are very similar, yet there’s a slight difference in interpretation:

While the carry over effect models the remains of effects from previous periods, the lag effect considers still active pressure from activities in previous periods. Make sure you understand the difference here. Maybe some examples will help: If I see an ad in a magazine in January, I might still be able to recall a bit of information in late June. That was carry over, then. If, however, I pick up an old magazine from January in late June and see the ad for the first time: that is a lag effect. So to this point we have associated these dynamic effects only with memory effects. But are they also applicable to consumer behavior? Well… Yes and no. Let’s start with the good news here. So why would we be able to apply these principles to consumer behavior? The reason for this are the different purchase intervals. Consider a TV set. I don’t need one right now, but I sure like these Bravia ads (Haven’t seen one lately, though). So chances are that the next time I buy a TV it’s going to be a Bravia. That’s a carry-over effect.

Now the bad news: This won’t work for fast moving consumer goods. If we don’t see immediate effects here, we shouldn’t wait for some carry-over effect to appear.

Now, we may primarily want our budget to play the role of the major choice variable. If you do, please consider the following: If you don’t see an effect, it doesn’t mean that there actually is no effect. People tend to forget, and you will need to use some money on making sure that they won’t forget your products. To that end we divide our spendings into “maintenance energy” (making sure they won’t forget) +”shift energy” (increase of whatever). Note that maintenance energy will always come first!

So try to see your budget as an investment and also as a reinvestment. Consider dynamic effects and necessary maintenance energy. Spend money on advertising! Yay!

Ah, and try to be different. There’s this ugly phenomenon called “interference” which does what it says: Similar campaigns stating similar claims about similar products not good.

Next time we’ll talk about our advertising goals. So bring your ambitions. See you then!

[Via http://zornslemon.wordpress.com]