From Bus Tickets to Billboards

Traditional Tracking Models Fail to Account for the Effects of Newer Forms of Communication. This Article Shows How Millward Brown's New Approach Closes the Gap

David Chantrey

Most of us have given very little thought, when considering how our brands should be researched, to the increasingly large number of ways in which a brand owner can choose to communicate with consumers. Much has been said about media fragmentation more channels, stations and publications meaning either:

  • a brave new world in which individual consumers can be specifically targeted according to their lifestyle, interests and behaviour, or;

  • abject confusion;

  • or somewhere in between.

But this is only half the story. In addition to fragmentation, new avenues of communication are continually appearing. Postcards, mobile internet, till receipts, takeaway lids, comet cursors, fuel pumps, private cars any space, real or virtual, is fair game. Then there are 'media' that are not new, but have been ignored by tracking research in the past. What price hoardings around sports fields? Or pointofsale material? Or product placement in TV or movies?

These are familiar to marketers, but traditional tracking was never designed for such an eclectic media mix. The fundamental idea still holds that brandlinked ad awareness is a key indicator of likely sales effects1. But the questions used to measure this were conceived when the media landscape was very different. The traditional methodology asks people whether they have seen a brand advertised on TV, in the press, on posters, etc2. To get the total effect, these measures are simply added together.

For many brand owners, this will still be a perfectly acceptable way to track their advertising. Until recently, 80% of Unilever's $3.3 billion media budget went on TV. In this situation, tracking TV communication plus a few extra media as backup is probably doing the job. However, this is rapidly becoming less common. As Alan Rutherford, Unilever's worldwide media director, has said: 'Unilever needs to change both its structure and approach to communicating. In media, we need to look at the whole range of channels that exist, not just traditional media which has major implications for how we and our agencies operate. We need to have the mechanisms in place to reevaluate those different channels'3.

In doing this, Unilever and others are thinking about 'advertising' in its broadest sense: 'Any paidfor communication intended to inform and/or influence one or more people'4. Tracking research needs to take a similarly broad view.

What should this encompass? What questions should we be asking, and how? What can this approach tell us about how multichannel communications can work? The more communication channels proliferate, the more difficult it becomes to get an accurate read of total brand communication using the traditional question. 'Have you seen Brand X advertised on bus tickets recently? What about in the back of taxis? And have you seen people holding up signs in the crowd at sporting events with Brand X written on them?' It's just not practical.


Millward Brown conducts over 700 tracking studies worldwide. To try to answer these questions, we selected a number of these studies for experimental work. We picked studies from markets furthest down the multi-channel communication route and set up parallel tracking programmes alongside them.

We tracked the same brands in the same markets across both sets of studies. In one we asked separate questions about advertising awareness in each main medium. In the experimental studies we first asked an initial, general communication awareness question. If consumers claimed to be aware of any communication for a brand, we followed up to determine the source and obtain diagnostic information on executions.

We experimented with several versions of the initial question, to maximise the range of communication considered by respondents. We proved that our final wording led people to think of all traditional media, and also of less obvious 'advertising' such as direct mail, promotions, pointofsale and sponsorship.

The idea is the same as traditional tracking it is just more inclusive. Most brand communication influences people not only at the time of exposure, but also in a delayed way over a long period. If it is sufficiently new and different it can sometimes have a major, immediate impact, but this tends to be shortlived 'new news' rapidly becomes old news. Econometric modelling shows that, to pay for itself, most communication must have a more longlived effect5. This makes intuitive sense. People don't run out to buy soap the moment they see a commercial (or taxi side, or banner ad) they buy when they run out. So communication needs to last long enough to carry through to the purchase occasion, and should come to mind in the act of choice. By prompting respondents with the brand for memories of recent communication, the new question is designed to uncover these mental links.


New questions are only useful if they add to learning. So what does this give us that the traditional method does not?

A simple example shows this. Assessing communication awareness for two telecoms brands using traditional mediumbymedium questioning showed One2One with a clear lead over its nearest competitor. However, by asking the question from the 'top down', the gap is closed significantly (Exhibit 1). The reasons become clear from the diagnostics. Vodafone has its own stores, and this, plus sponsorship and direct mail, increases its presence relative to One2One (Exhibit 2). Traditional questioning would have missed this.

Throughout the work since our initial experiments, pointofsale shows a significant influence on overall communication awareness for many brands. The next example, for an oil brand (Exhibit 3), illustrates just how important it can be.

The data come from a multicountry study. In Brazil, the brand has had no recent TV advertising and, as a result, its overall presence is relatively low compared to South Africa. However, it is still making some impact on consumers and pointofsale is the most significant contributor to this. Exhibit 3 shows the importance of pointofsale in the absence of recent TV advertising. In South Africa, the brand has had a recent TV campaign, which has driven up its overall communication awareness. But pointofsale is still the second biggest driver, outperforming print advertising and sponsorship. Again, traditional questions would have missed this.

Both examples are static snapshots of communication awareness, and its source. However, a key benefit of tracking is its ability to monitor how awareness develops over time and in response to marketing inputs. How does this approach work with continuous data?

The first continuous studies using the new questions have been running for some six months. We have limited continuous data as yet, but early signs are encouraging. Awareness, medium by medium, does not vary that significantly from the old to the new method: a TV ad awareness line from an old study remains at roughly the same level with the new method it is total communication awareness that increases. What the new method allows us to do better is to track total communication awareness and deconstruct it into its components. Exhibit 4 shows a good example from a tracking study for a food brand.

Consumers' memories of the brand's communication are driven mainly by radio. However, this has been augmented by sampling, and recently TV has provided additional support. Even though memories of radio are starting to decay, TV and sampling are helping sustain the brand in consumers' minds.

Although it can be complicated (and sometimes made impossible) by concurrent spend across media, the data let us begin to establish the relative contribution of each medium. By modelling each media line against spend, we can determine ROI for each activity. Where campaigns overlap, we need to seek out cases that let us reduce the number of variables we are trying to find (for example, a subsequent burst with the same executions, but where weights of the various media are different).

Awareness is only the first step. More awareness means more opportunity to communicate a message, and to build brand equity. This is why we see a strong correlation between an ad's ability to build awareness and its ability to generate sales. But it is only a start, and any good tracking study will also measure other variables on the road to success.

Table 1 shows why this matters. This brand had spent a significant sum on a relaunch via TV and print, and had seen an impressive rise in awareness of its communication. However, the relaunch came at a bad time the industry faced significant problems and many brands in the sector suffered bad PR. This PR effect was evident from the source of awareness, and also in responses to a question about whether the communication left respondents feeling more positive or negative towards the brand.


Total communication awareness:
Pre Post
% %
31 45
Feelings towards brand created by all communication:
Pre Post
% %
More positive 37 30
More negative 5 21*
*Relaunch coincided with industry problems, with negative effect on brand

It became clear that the communications had made consumers feel more negative towards the brand. We were able to pin this down directly to the bad PR, and show that the TV and print had partly offset this effect. Nevertheless, overall, the brand finished worse off than it began, and if we had only looked at the response to the ads we would have missed a critical part of the picture.


So it is important and possible to measure the net effect of a brand's communication, and adopting a 'top down' approach helps pick up some supplementary media that traditional tracking might have missed. What have we learned about the roles of individual media within a campaign? Is it all about return on investment, or can different media bring different things within the total communication mix?

This warrants a paper in itself (it has6). We recently carried out a major qualitative project into how consumers' impressions of a brand's communication are shaped by the medium through which they are received. This shows that different media can impart their own character to a message. A message that is trustworthy and credible in one medium might be viewed with wariness in another even for the same brand. The extent of this media 'spin' varies by country and the status of each medium within that country. For example, in most markets the internet is suspect (and this does not seem to lessen where the internet is more established). This can be overcome to an extent by the credentials of a large brand, but has implications when planning a campaign involving the internet. It may require another channel such as TV or outdoor to lend credibility to the message.

If you are putting together a multimedia campaign, what are the likely interactions between ads across different media? Should the work be synergistic and, if so, what are the likely benefits? This is a field where we are still building case histories. Intuitively, synergy across media has benefits. Seeing a message from a brand in one medium might make an impression, but seeing the same message slightly differently in different channels might be even better.

Can we quantify this? The number of variables makes it extraordinarily difficult we are faced with media effects, and the quality of each individual ad throws in a different set of unknowns. We have to rely on particular cases to help us to understand the possibilities cases where something has been tried one way in one place, then slightly differently somewhere else.

Here is an example of how synergy can work: a major launch in Europe and the us. The strategy was to surround the consumer with communication at launch. TV was the lead medium, plus a variety of other media bussides, radio, outdoor, print, etc. The TV campaign was common across markets, but the weight and nature of the secondary media varied by country, and this gives us our case study. Table 2 shows the proportion of the budget spent on secondary media. TV dominated the campaign in the US, but in Europe greater weight was given to other channels.


US France Germany UK
% spend at launch* on secondary media 8 40 24 28
*% in terms of dollar investment

This variation in campaign structure was seen in tracking. The client had pretested the TV campaign and knew it had a highly efficient, powerful launch vehicle. In all markets, tracking and sales results validated the pretest but in Europe the communication cut through much more efficiently than expected. This was more marked in Germany than France, and was most marked in the UK. So why the differences?

The explanation lies in the nature and timing of the secondary campaigns. In Germany and the UK, all the creative for secondary media was synergistic with TV, and the secondary media were timed in the UK to surround consumers without oversaturating them. By a combination of synergy and careful planning, the client had constructed a campaign where the whole was demonstrably greater than the sum of its parts.


Tracking must evolve. As media choices proliferate, we can better understand how communications work by adopting a holistic approach to accessing consumers' memories. This maximises the chance of capturing all the effects, discovering what is and isn't working, and improving campaigns for the future.

Thanks to Sally Beswick of Millward Brown for her help in preparing this article.

1N Hollis: 'The link between TV ad awareness and sales effects new evidence from sales response modelling', Journal of the Market Research Society, 36(1) January 1994.

2S Colman and G Brown: 'Advertising Tracking Studies and Sales Effects', Market Research Society Conference Papers, 1983.

3S Bidlake: 'Think Different', Advertising Age International, 8 February 1999.

4J Bullmore: Behind the Scenes in Advertising. NTC Publications Ltd, 1991.

5P Dyson: 'Justifying the Advertising Budget', Admap Conference on Advertising Effectiveness, 1998.

6A Branthwaite, K Wood and MCM Schilling: 'The Medium is Part of the Message', ESOMAR/ARF Conference, November 2000.



David Chantrey

David Chantrey

    David Chantrey is a group account director at Millward Brown. He specialises in brand equity and tracking research, and is currently the agency's global account director for the Levi Strauss & Co. account.

From Bus Tickets to Billboards