Measurement

Written by Laurie Kirschner, UK and Pan-European Research Manager, Yahoo!

• Display measurement
• Behavioural response measurement
• IAB Brand Engagement

Campaign measurement is essential to any medium and none more so than the internet. Also like other media, it’s important to determine what you will be measuring upfront. This chapter will discuss different methods of measuring display advertising to help you on your way.

Display measurement

For researchers and measurement specialists alike, the luxury of interactive advertising can be found in its name. By definition, interactive advertising is advertising with which consumers can directly interact. Consequently, these behaviours can be directly measured and impact directly quantified, whether by behavioural metrics (such as clickthrough rates and interaction rates) or brand metrics (such as awareness and purchase intent).

Unlike our offline media counterparts, who can only identify advertising exposure implicitly, exposure to interactive advertising can be measured and tracked explicitly. We know which consumers have been exposed to the ads – their demographics, their behaviours, and so on.

As a result, we can use this knowledge to set up a near experimental design whereby a matched control group of consumers who have not been exposed to the ads can be established, and their resulting behaviour and/or changes in attitudes tracked and compared to that of the test group of consumers who have been exposed to the ad. If the two groups are matched well enough, no confounding variables will be present – the only difference will be the exposure to the ad.

Thus, any differences in behaviour or changes in attitudes can be tied back to this exposure. This methodology can be used to measure the impact of all items discussed in previous chapters of this guide, including the differing impact of various ad formats, the effect of behavioural targeting (over and above contextually targeted or untargeted ads) and the use of ad networks.

Behavioural response measurement

• Click-through rate (CTR) – the number of users who clicked on an interactive ad divided by the number of times the ad was delivered, expressed as a percentage – is the most basic of all behavioural response metrics. It quickly allows marketers to understand the degree to which consumers responded to their ads. Since it is simple to calculate, CTR is often used to measure the success of an interactive campaign. While it is a powerful metric – particularly for campaigns that have an immediate call-to-action – it is important to use it within a suite of other response metrics in order to understand the full story. This is because CTR does not include people who failed to click on the ad, but arrived at the site later, as a result of seeing the ad previously. This latter metric is known as ‘view-throughs’ and has often been shown to account for higher conversion rates than ‘click-throughs’ on interactive advertising.

• Ad engagement metrics – a set of metrics used to show the number or percentage of users who have interacted with a particular rich-media ad – can be used to further measure the success of an interactive campaign. These metrics are often used to round out the picture of campaign impact by showing the additional number of consumers whose interest in the advertised brand was piqued to such an extent that they involved themselves with the ad. The suite of ad engagement metrics can be found to the left.

By using these additional behavioural metrics, the marketer is able to see the total number of actual conversion opportunities since those who interact with the ad are as valuable as the (often smaller number of people) who click through to the brand website.

• Keyword search volume – the change in number of searches on particular branded keywords – is a useful metric used to show further engagement with the advertised brand. Using the experiment design methodology outlined above, marketers can measure and compare the search behaviour of those exposed to the advertisements online, to that of a matched control sample who have not been exposed. The analysis is conducted on branded keywords related to the campaign. If the two groups are well matched and the exposed group shows an increase in keyword search volume over and above that of the control group, we can say that exposure to the interactive campaign led to a change in behaviour and an increase in engagement with the advertised brand.

Although historically plagued by scepticism due to its survey recruitment methodology, online brand impact measurement has evolved considerably over time to become a fundamental part of post-campaign reporting.

When marketers measure the brand impact of a campaign, they can test the following:

• Unaided brand awareness – measures users’ recall of the test brand without prompting.

• Aided brand awareness – measures users’ recall of the test brand from a short list provided.

• Ad recall / online ad awareness – measures users’ recollection of an ad for the test brand.

• Message association – measures users’ recollection of the campaign key messages.

• Brand favourability – measures the percentage of users that view the brand positively.

• Purchase intent – measures the percentage of users thatintend to buy the test brand.

Additionally, marketers can measure the change in brand perceptions as a result of exposure to the interactive ad campaign.

Again, to measure any or all of these factors, marketers would employ the control/exposed methodology summarised above. To determine those that fall in the exposed group, a javascript tag is placed on the ad and this piece of code will drop a cookie (a file that saves information to help website services) on the user’s browser. Those users who do not have this cookie on their browser will be placed in the control group, indicating that they have not been exposed to the ad in question. Individuals in both groups will be recruited to partake in a survey which asks a variety of questions used to get at the above metrics. Finally, answers to the questions by the exposed group are compared to those of the control group and any significant differences are attributed to the interactive ad campaign.

IAB Brand Engagement

Since 2006 the IAB has appointed Carat Insight, now known as ævolve, to examine the effectiveness of online at driving brand engagement, alongside other media, as well as wider influences such as ownership, reading reviews, heritage or word of mouth. The intention was to demonstrate the role that online advertising can play in client’s marketing communications strategies, beyond the click-through, and that online advertising can very successfully drive brand engagement.

These studies set out to investigate people’s attitudes towards products
within the categories tested, and how these attitudes inter-relate to deliver brand engagement, and we so far have results for the automotive, haircare and soft drinks categories. The research in each study used the same 3 components for the methodology and consisted of qualitative, quantitative and multivariate modelling to test brand engagement.

Automotive

The first of these studies carried out in 2006 in the small car market, targeted ABC1C2 women with children under the age of 16. This audience was selected because of their considerable influence over household spending makes them particularly important to a wide range of advertisers.

The topline results showed that communications drove 15% of total brand engagement, and of this, that online advertising drove more brand engagement than any other individual media channel - delivering 39% of the marketing communications effect on brand engagement. This is a really impressive share for online and further convinced us to test different categories as results will invariably differ depending on many different factors related to the type of sector that the brands fall into.

Haircare

The second study researched the shampoo and conditioner category – 1,000 women ABC1C2 and aged 20-60 were questioned. All had to have agreed with 4 out of a possible 7 statements presented to them, such as ‘I only buy exotic bath products’, ‘I only buy the best makes/labels in life’ and ‘I regularly visit a beauty salon’. Therefore, they were a sample of women who felt that personal grooming was very important to them.

This time around we found that online advertising and webpages combined delivered 35% of the marketing communications effect on brand engagement. Further to this, online advertising was 2.4 times more effective than its share of spend suggested. In addition, women seeing a brand’s online ads are 3.5 times more likely to see its webpage and so there is a strong correlation here in this particular category. The internet was uniquely able to combine brand advertising messages that could readily lead women (who were a highly involved audience in this sector) to personally relevant content that helped them to make decisions about their hair products.

Soft Drinks

We carried out the third study in the soft drinks sector and researched 1,000 ABC1 men aged 20-45, with no children and all in full time employment. Topline results in this sector found that, across 4 brands, on average a soft drink brand’s communications increase brand engagement by 5%. Of all the communications channels, internet advertising delivered 24% of the overall marketing communication effect on brand engagement. We also found that online advertising for soft drinks is three times as effective per £8.5 of spend as the all
media average.

Overall these studies demonstrate how online can be really influential in driving the level of engagement that people have with brands. It is particularly interesting to see this in the two FMCG categories considering how much room for growth there still is in this sector for really great branding campaigns. We hope that all three studies have helped to educate advertisers on using online display advertising as a branding medium as well as a direct response tool.


[back to top]