Online measurement: Connecting all the dots Matthew Dodd Admap March 2010
Online measurement: Connecting all the dots Matthew Dodd Admap March 2010 Connecting all the dots Matthew Dodd Online branding spend has suffered from a lack of satisfactory metrics but a new approach could give brand owners the data on ROI they need At long last, major brand advertisers are starting to invest in online display. However, how much of this is driven by economic necessity, rather than a leap of faith, is clearly evident when we look at the overall investment levels, which are only 1% of total media expenditure and significantly less than TV, the mainstay of branding communications. So, why the reluctance to move more branding money online, when it is a highly accountable medium? Is it because there is still a question mark over the metrics? As the new kid on the block, the internet was keen to demonstrate quickly its value to advertisers, so the click-through rate (CTR) became the metric of choice. This was fine when the medium was in its infancy and advertisers enjoyed CTR rates of 510%. But as the medium flourished and attracted more advertisers, response rates fell dramatically and now average 0.18% (source: AdTech 2007). At the same time, it pigeonholed the internet as a DR medium and created a vicious circle from which it is only just beginning to escape. However, the industry has not helped itself by putting up more ad slots in the hope that it would improve the opportunity for someone to click on the ad. This has created poor and cluttered environments, which have further hindered the advertisers' ability to build their brand messages. Furthermore, the ad delivery measurement metric the ad impression gives little or no guidance on online's ability to deliver audiences or to compare this performance with other media. Add to this the industry's acceptance that cookie deletion and multiple device and location use lead to an over-inflation in (browser) audiences and an underestimation of frequency the very measure that brand managers are keen to use to communicate their brand messages. By 2000, the internet really started to gain momentum and the industry realised that the click-through metric was limiting its ability to sell to brand advertisers. A new effectiveness metric was born in the form of brand effectiveness studies delivered via a pop-up survey, often served as someone left the site on which the ad ran. Unfortunately, this has done very little to remove the scepticism around evaluation metrics, since the campaign is measured at the peak of consumption and, again, this over-inflates online's ability to build a meaningful brand dialogue. 2
Also, with the growing emergence of social media, none of these approaches takes into account the holistic nature of online and its impact, both in terms of managed (display, search) and unmanaged (social media, blogs, boards etc) online media, in driving both positive and negative brand perceptions. We need to move the internet effectiveness measurement debate forward and provide a better assessment of its impact and role in building audience reach and changing brand perceptions and behaviours. There are three things that need to be done to improve the industry's ability to demonstrate the effectiveness of branding online. The first is to develop an online gross rating point (GRP) to isolate digital's ability to deliver audience reach versus, and in conjunction with, other media. The second is to provide a fairer assessment of the impact of online display, rather than the click-through metric, and provide a better attribution of its role in driving conversions. The third is to align online brand effectiveness surveys with traditional media effectiveness approaches to provide a more realistic measure of long-term branding effects. POSITIVE CHANGES To address these issues in Europe, Nielsen has developed a new approach, which addresses these challenges head-on and answers the fundamental question: did the campaign positively change attitudes and behaviours? passively capture these exposures at a respondent level on our desktop metered panel. Our metered panel is fully representative and weighted to the country's online population. It also takes into account cookie deletion, as well as filtering out non-resident based exposures to the advertising. This enables us to determine who, and how many, actually saw the ad, and provide an estimation of audience reached by the campaign and the frequency of exposure. We can also see whether it reached the right people demographically and if it elicited the desired online behaviours, such as: did people go on to visit the site post-exposure? How long did they engage with the brand? Did they conduct a related brand search, or did they conduct a quotation or transact online? We can also see how the constituent campaign components, such as creative, ad format, sites and placements, performed against the objectives. Figure 1 This makes it possible to provide a more insightful, comprehensive and accurate picture of the overall performance of the campaign. To determine the campaign's ability to build brand engagement, we 'pair-wise' match the control (unexposed) 3
respondents to the test (exposed) group on a range of demographic and behavioural variables. To determine the impact of online on attitudes, we tag the Research Now (RN) panel to use those panellists that have agreed to have the Nielsen cookie placed on their PCs (400,000 panellists). The process of dropping the cookie is a simple image call to our Nielsen server; it's the same process that occurs all over the web daily. Once the respondent has successfully joined the programme, RN only needs to ensure the cookie is reset and constant. RN employs a variety of methods to do this, including regular email campaigns and placing the same code on high traffic pages on its site. This allows cookies to be matched when the panellists are exposed to tagged ad units, enabling Nielsen to identify respondents who have been exposed to the advertising and emailing them to participate in the survey. This eliminates the need for pop-up survey invitations on the site where the advertising is being conducted. Survey ad effectiveness research is fielded on a 24-hour rotating schedule. We identify a sample for inclusion in the research and send survey invitations every 24 hours, although some clients have widened the time period between exposure and survey to determine the latency effect of the campaign. As a result of the survey invitation procedure, the time at which the respondent answers the survey versus the time of exposure can be controlled. This eliminates the problems of effectiveness measurement solutions that measure the immediate effect. This approach provides a more realistic measurement of the advertising's long-term effect, and greater validity and credibility to the research results. The analysis presented by Nielsen examines the brand's ability to increase its brand awareness, affinity and consideration by comparing the results of the control cell respondents (unexposed) to the test cell respondents (exposed). Any difference in the responses collected from each of these groups can be attributed to the advertising's effectiveness. In addition to examining how a particular advertisement is able to succeed at each of the levels described above, it is also important to look at the campaign effectiveness from a variety of different analytical viewpoints, such as site level (ad networks versus portals) and format (rich versus standard media). By examining the results of 100 studies, based on a sample of more than 380,000 panel observations and 60,000 survey responses, we now have a much clearer picture of online's ability to impact at all stages of the consumer purchase funnel. We are able to determine that online display is able to instil positive brand memories. By measuring 24 hours' post-exposure, we are able to pinpoint online's ability to instil top-of-mind brand awareness the Holy Grail for brand managers. Spontaneous (first mention) brand awareness increased to 30% among those exposed an uplift of 20% on the control group (Figure 1). Our results also show that online works at the next level of brand awareness, with 31% of those exposed to the display campaign we measured being able to recall the advertised brand a 12% increase on the control group. One in three were able to correctly associate the advertised brand with the campaign message, illustrating either the improving online creative execution and/or greater connectivity with offline media. From a branding communication perspective, online is clearly driving all the key metrics. So what about its role further down the purchase in driving brand engagement and, ultimately, conversions? From our normative database, we were able to determine that online display campaigns can significantly bolster advertiser site engagement levels up to 83% for overall audiences, 33% more pages per person and 23% for time per person on the site (when we compare the exposed group to the 4
control group). ONLINE DELIVERS ROI By visualising all campaign exposures on the metered panel, we are also able to determine the true role of online display in driving people to the advertiser site post-exposure, and not just directly after they clicked on the ad. We were able to report back that online display is, on average, 20 times more powerful than the current CTR metric would have us believe (ie on average, 3.6% went on to visit the advertiser site post-exposure, compared with an average CTR of 0.18%). Perhaps online publishers should be charging similar premiums for their display inventory and now use the 'click to exposure' metric, which is a much fairer and more illustrative measure of how online display actually drives brand engagement and, ultimately, ROI for the advertisers. Our findings are very encouraging and it is clear that online display can deliver strong top-of-mind awareness and brand engagement. It is also clear that the current suite of metrics is insufficient in evaluating campaign delivery and assessing the viral impact a campaign may have on extending and amplifying the advertiser message. Brand advertisers need to reassess the real opportunity of online. They need to move away from the 'last-click-counts' model to look at the bigger picture that online display plays in the customer journey, thus gleaning a richer insight into its ability to distil brand messages in the minds of online consumers, stimulate brand engagement and deliver online ROI. CASE STUDY: EVIAN Evian's roller-dancing babies execution was launched online before airing on TV in the UK and France Online is unlike any other media, given its participatory nature. Evian wanted to tap into online's viral capability to boost campaign reach and 'talkability' for its Live Young campaign. The humorous campaign concept centred on a dozen roller dancing babies who showcase their skills through the vitality of Evian. This wasn't the first time Evian had used babies to communicate the 'vitality' message; in 1998 they featured in a synchronised swimming spectacular. However, this time, the campaign was solely devised for online the ad was subsequently shown on TV in France and the UK (in December 2009) but only after the online campaign had aired. The online branding campaign aired for one day across six markets on YouTube, with a 60-second Flash video being shown on the homepage. Through our new ad effectiveness methodology, we were able to identify not only those who saw the paidfor YouTube homepage ad but also the viral amplification of the campaign. By isolating the viral impact of the campaign in the US, for example, we were able to Determine that this boosted campaign (audience) reach by 84%, online GRPs by 160% and Evian brand (time) engagement by 36%. Assessing all these impacts enabled us to prove that the viral campaign significantly amplified and extended the campaign reach, GRP delivery and Evian brand engagement. Moreover, from a brand building perspective, five to seven days after the homepage takeover activity, the campaign still had a 5
significant impact on all key brand diagnostics among the core audience of regular bottled water drinkers (Figure 2). Figure 2 ABOUT THE AUTHOR Matthew Dodd is vice-president of research and analytics at Nielsen Online, EMEA. He leads the EMEA research and analytics teams. matthew.dodd@nielsen.com Copyright Warc 2010 Warc Ltd. 85 Newman Street, London, United Kingdom, W1T 3EU Tel: +44 (0)20 7467 8100, Fax: +(0)20 7467 8101 www.warc.com All rights reserved including database rights. This electronic file is for the personal use of authorised users based at the subscribing company's office location. It may not be reproduced, posted on intranets, extranets or the internet, e-mailed, archived or shared electronically either within the purchasers organisation or externally without express written permission from Warc. 6