Nowadays, calculating campaign effects is becoming more and more complex. This can be explained by the rapid increase in media channels and their diversity (i.e. outdoor, internet, radio, television) as well as the effects on reach such as second screen use (smartphone and tablet). Applying classic methods like OTS (Opportunity To See) to measure online and offline campaign effectiveness has become outdated.
Why we cannot rely on OTS
OTS research measures through (online) surveys the chance a person came into contact with a specific advertisement. These surveys consist of questions like: “When did you watch television?”, “Which channel?” and “At what time?”. In 2005, the association of market researchers, MOA, already criticized the OTS method (MOA yearbook) on the following two main complications:
The memory problem
One the one hand, people are often convinced they see certain things that actually aren’t there, whilst on the other hand they don’t see things that do occur. So in fact, the respondent’s memory is being tested and not the desired effect. Alfred Levi (2005) studied the behaviour of consumers during television commercials. His results suggest merely 19% of the people attentively watches a commercial, 18% leaves the room during a commercial break, 43% stays in front of the television but doesn’t watch it and the other 20% zaps to a different channel. His conclusion: 81% of the people hasn’t seen or didn’t pay attention to the commercial.
The motivational problem
Besides this memory problem, a motivational problem arises: people tend to see themselves better than they actually are and most of the time they are unaware of it. For example, if we highly value intellectuality, we think we read quality newspaper De Volkskrant more attentively than showbiz magazine Privé (in Psychology, this phenomena is called a self-serving bias). However, this is not the case. Overall, both the respondent’s memory and motivational problem lead to inaccurate results in this type of research, and thus are reasons for criticism.
Moreover, from our experience we have learned that e.g. brand awareness can increase with 5% and decrease the next week with 6% without any clear reason. The OTS model makes these direct campaign effects insufficiently insightful. But then, which method does have the ability to objectively measure the effect of online and offline campaigns?
A new type of research is born
Through the internet, online analytics have become available to provide insights into web visitor data. This objective consumer data is of great importance to a lot of advertisers since one of their goals is to use offline campaigns to attract online web visitors. For advertisers, web visitor data is a crucial measure for campaign effectiveness, even when online sales aren’t their priorities. For example, in the automotive industry, less than 2% of all new cars are sold online. However, online traffic is an essential link in the sales process. For many car manufacturers, an increase in web visitors equals an increase in sales. Measuring offline campaign effects by using online analytics is called the method of Attribution Modelling (ATM).
How does ATM work?
In contrast to classical methods, Attribution Modelling measures only those consumers who have been in contact with the campaign. By using econometrical models and algorithms, the direct impact of the campaign can be measured by applying the following steps (also displayed in the figure below):
- To apply Attribution Modelling, the influences on the number of web visitors need to be explained. The norm is being determined: this represents the web visitors that are not caused by a campaign. This is done by indicating all external impacts on web visits, such as weather, seasonal and holiday influences.
- With data of media schedules, external influences and web visitors, the daily impact of the campaign on web visits is calculated using statistics. The accuracy of this model finds itself in adding the calculated influences on web visits (temperature, holidays, seasons and campaigns) and comparing this with the true web visits.
- Now that the web visits due to the campaign are calculated, Attribution Modelling is applied to determine the effectiveness of different channels. Using this model means: the more data available, the more insights in the campaign effects. The effect is measured by putting the spent media budget into perspective with the effective reach. Measuring the effect in equal units makes different channels comparable. Subsequently, it is possible to zoom in on specific channels to see its effectiveness on channel- and hour level
Attribution Modelling makes offline media effects just as measurable as online media
With the ATM method, an advertiser can directly see that the most cost efficient timeslot for broadcasting a TV commercial on channel A is between 11:00 AM and 2:00 PM (see figure above). During that time, the commercial has the lowest costs and people are most likely to take action (go to the shop, visit the website, buy a product). Online marketers might be familiar with this: in the online marketing world this is called the Cost per Click (CPC). These are the average costs to get a consumer to the website. Attribution Modelling therefore makes the effect of offline media just as insightful as online media, something that is impossible with research methods like OTS. Where OTS fails due to unpredictable respondent behaviour, Attribution Modelling succeeds to find the effects of the deployed media channels with high accuracy.
This blog is published on external websites: