The Challenge of Continuous Improvement ... The Integrated Data Source Approach to Market Research

THE CHALLENGE OF CONTINUOUS IMPROVEMENT...
THE INTEGRATED DATA SOURCE APPROACH TO MARKET RESEARCH

Roger E. Godbeer, Global Media Director, Colgate-Palmolive, USA and
Richard S. Wilson, Managing Director, Wilson Market Research Pty. Ltd., Australia

The paper is divided into two parts. The first part discusses the factors, which will lead advertisers in general, and one advertiser specifically into complementary data sources to the traditional data banks preferred by marketers, or those preferred by the broadcast media. The goal will be to significantly improve target group identification and the evaluation and selection of media

The second part presents the information concept, "Unity of Data", from an integrated data source, which enables advertisers to identity markets and consumers in terms of media habits. A specific category will be presented to demonstrate the complete process. The presenter will also establish the close relationship of this data source to the traditional services employed by advertisers and media. It is not the intent to replace these services that best meet the demands and interests of their respective constituents. The concept of "Unity of Data" is seen as a marketing enhancement to make advertising more accountable.

Today, all advertisers have to be most concerned with the efficiency and effectiveness of the media investment. It is not enough to achieve the current year's marketing goals. We have to make that media investment work harder; driving for incremental volume. While we valiantly attempt to establish media value, the reality is that we are failing vis-à-vis other marketing activities.

In the United States, and we believe that this holds true for most countries, media's share of the marketing dollar continues to decline. This is occurring despite major commitment to advertising priority.

The primary medium for many of us, television, is becoming less cost effective. Ongoing liberalisation of channels coupled with cable and satellite expansion may have given the viewer expanded choice, but he or she is generally not spending more time with television.

The consequences are not healthy for television or advertisers. We can expect, and are in fact, seeing higher cost efficiency when the lower ratings are combined with rate increases. This is pushing us short term to be more aggressive in demanding multimedia plans. But replacing television with other media may not address more fundamental issues…do we know who are our prime prospects and are we effectively reaching them?

Concurrently, the industry in Europe is committed to the principle of a harmonised demographic database. While we are an active supporter of this, we will have to await the results of the first contract renewals in Ireland and Germany. Even with a success in Europe, there is the balance of the world to be addressed.

While such may be the ideal reference point or currency for media, demographics are proving both inadequate in establishing the target consumer in terms of value or return to the brand and limiting in assessing media's reach against our various subgroups.

At the 1992 ARF/ESOMAR Research Symposium we presented the premise that there would be limited to no future in expecting advertisers and media to collaborate on a core single source service. Advertisers demand reliable data on retail volume, share, display, co-op, feature pricings and store type distribution. Media, on the other hand, fully expect reliable data on ratings, share, audience profile, and in the future, information on audiences not currently covered in the current services. To maximise their respective interest, we see limited to no hope either party being willing to compromise their real or perceived self-interest. This is not a condemnation but a realistic appraisal.

As a result, we are fully committed to actively exploring complementary services that have the potential to combine the consumer profile, attitudes and psychographics with product purchases and media habits.

The example that we'll share with you today, although not for a Colgate brand, most effectively demonstrates a new and absolutely necessary new direction. We believe so strongly in the benefits that at Colgate, funding is coming off the top of the working media budget. As such, we at Colgate Media, Marketing and our Brand Agencies have to demonstrate week-in, week-out, a return on this incremental research investment.

This is no idle challenge. Bill Shanaban, President and Chief Operating Officer, has laid it on the line: "Does this add value and deliver incremental sales?" We are convinced that we can.

At this moment in time you are probably wondering why we are sharing this business opportunity with you and potentially our competitors. It is simple. We believe that the initiatives developed in Australia by Wilson Market Research have application to other markets. We see a real opportunity for the current omnibus style studies to improve to meet the advertiser's needs of the future.

Proprietary data is expensive. We can live with syndicated data as we believe most strongly that it is not what you have but how you use it. We plan to stay ahead of the pack.

We don't want to advertise to people who are not in the market for our product. We do not want to advertise to people who buy only store brands. We do not want to reach people who for the past three years have been committed to other brands.

We do need to more effectively identify our real prospects and to fully understand them in the context of their shopping and media habits. We need a unity of data that parallels the results achieved by the self or special interest of media and advertisers. We have no choice if the media investment is to grow.

PART 2. UNITY OF DATA —
A FULLY INTEGRATED DATA SOURCE APPROACH TO SOLVING MARKETING PROBLEMS

Today there are many excellent syndicated information services on offer to the marketer, which can be used to establish the status of his brand. Each service usually concentrates on one particular aspect of the market and provides accurate data. You need only to look at the sophistication of scanning services, the detail provided by Millward Brown advertising measures and the new technology on offer for electronic media measurement to understand what I mean. In short, we have a lot of hi-tech numbers rolling down the information highway.

What is missing from these data is not reliability but unity. Slide 13 lists some of the major types of syndicated data on offer around the world and the range of information usually contained therein.

Clearly not all of the services detailed offer as many features as are noted in our slide and perhaps a couple offer more. Those that offer the most information are usually not continuous and therefore incapable of reporting trends. Although the integrated data source approach may not be able to guarantee the level of precision of a subject specific research study, this approach is demonstrated to deliver at least 85% accuracy while also providing the most important element of all – a unity of data.

We have validated market share data against industry sources for many clients over the course of the last three to four years and answered questions on form and size of our self-completion questionnaire. We'd be pleased to provide further details if required, however, Slide 14 compares industry brand share data with volume shares derived from our integrated data source for leading brands in a range of markets. In some instances we are comparing value share with volume share and sometimes the measurement periods vary slightly. By and large it is evident that the integrated data source approach provides market share estimates, which are reasonably consistent with industry sources.

The fully integrated data source approach lacks the sophistication of the “people-meter” method of television audience measurement (the industry currency) as it is based on an individual self-completion diary technique with interviewer assistance. Slide 15 shows for a similar measurement period the differences in program ratings between people meters and the integrated data source approach. Results are based on a four-month average for the national metropolitan market.

Clearly, despite methodological differences, Slide 15 demonstrates there are many similarities in ratings with one exception – namely the 6:00-6:30 pm time slot weekday when the news is telecast. After 6:30 pm the numbers appear to converge. Equally, at 5:30 pm the numbers are very similar. However between 6:00 and 6:30 when most folks are coming home from work there appears to be some variance. At another time it might be interesting to discuss this phenomenon but today I would draw your attention to the similarities rather than the differences.

Prior to discussing the type of information generated from a fully integrated data source approach, I will provide some background on the methodology we use.

A total of 45,000 interviews are conducted each year. This equates to 4,500 interviews per month, over 10 months, February to November. Reports are provided monthly to subscribers in hard copy and on diskette.

Door to door recruitment is used in all but remote areas of Australia where telephone is used. At the door interview demographics and advertising awareness are sought from the respondent. The diary completion method is also explained and demonstrated.

Respondents participate for just one week and diaries are collected the following weekend.

Diaries from remote areas are returned by mail.

Mid-week reminder calls, thank you cards, introduction cards and interviewer IDs are used to ensure a high cooperation rate.

Collection rates are good. An average collection rate of 81% is achieved on diaries personally placed whereas 60% is obtained for surveys recruited by telephone and returned by mail.

A validation level of 25% is maintained.

A random sample of Census Collectors Districts (CCD's) drawn at start of each year is the basis for the year’s assignments. Starting points are selected within each CCD. A stratified sample is required for each assignment of 12 interviews, which must be achieved from 25 calls.

One interview per household only is required – and callbacks to target households are made for up to three weeks. The respondent reports for his or her own behaviour – not the household

Our first step in this analysis was to determine where market volume was coming from by way of a regional analysis of consumption. This was done for company sales regions and all major cities. This provided the advertiser with a CDI (Category Development Index) ranking of overall market importance.

A similar analysis was then carried out for brand volume to show regional differences in brand performance using a BDI (Brand Development Index) ranking.

Having determined market geography our client asked the following questions of our integrated data source:

Can we increase the retail price of our commercial brand,

How can we more efficiently combat the market leader in the premium segment,

How can we more effectively reach with increased frequency our target market, and

Is our advertising working?

To answer the first question we decided to analyse heavy sparkling wine drinkers to find out their importance in the market after which we would look at the image of our client’s commercial brand.

Slide 18 (not shown) indicates that heavy drinkers (defined modestly as 9 glasses or more per month) are just 19% of all drinkers but account for 45% of volume. Slide 19 compares the image of our client’s commercial brand with two other leading commercial brands. Slide 20 compares our client's commercial brand with a leading premium brand (approximately twice the cost of the commercial product).

With the exception of a couple of image attributes our client’s commercial brand sits favourably up against a leading premium competitor. Further inspection of slides 19 & 20 allows us to agree that given the image similarities between the premium competitor and our client’s commercial brand differences with other leading commercial brands on a variety of core dimensions e.g. quality, our client could justify shifting the price of his commercial brand closer to the premium end of the market. This he did with no adverse sales reaction. Our client's commercial brand was clearly in a category of its own in the mind of the sparkling wine drinker.

We carried out a cross consumption analysis to determine which brands of sparkling wine hung together in a consumption repertoire and found a fascinating result.

We identified three market segments based on consumption patterns. Without explaining in detail the nature of each, it was the third of these that was most relevant to the current goals of our client. Slide 21 shows Segment 3 accounts for 18% of drinkers and 26% of market volume and fertile ground for developing both products of interest.

In Segment 3 our client’s existing premium product is performing poorly but his commercial brand performs extremely well. Scope clearly existed to develop the client’s premium brand in this segment although how that was done is of course confidential. However, easier in the near future would be to increase the presence of the client’s commercial brand in the segment.

In earmarking Segment 3 as a real marketing opportunity we proceeded to explore the image of our client’s commercial brand against commercial competitors to determine where each brand figured in the mind of the consumer. Given that on all but two variables our client’s commercial brand performed extremely well against the competitor’s premium brand (market leader) and completely devastated other leading commercial products, we again demonstrated to our client that he was justified in seeking a higher price point with his base line product within Segment 3 as well as in the market overall as previously shown.

Slide 22 shows that our client’s commercial brand performs strongly in both Segment 1 and Segment 3. It was Segment 3 where strategically our client’s interests were at the time and therefore the next step was to gain an in depth understanding of those drinkers of sparkling wine residing in Segment 3.

Once the above profile was developed, a media profile and schedule was constructed to optimise delivery to Segment 3.

Slide 24 highlights the importance of Regions 2, 3 and 4 in Segment 3 membership with Region 2 the hub of activity and possibly the focus of concentrated electronic as well as media below-the-line activity.

Slides 25 and 26 show Segment 3 to be concentrated in the 25-54 age bracket and predominantly of AB socioeconomic background. This is considerably different from the population and all sparkling drinkers in general.

Base= Drunk sparkling wine in the last 30 days. M-W 25-54 is likely to be the target demographic for media

Slide 27 (not shown) provides some detail as to the wine consumption repertoire of Segment 3 drinkers which can only be described as comprehensive and indicates the extent to which sparkling wine is competitive to other wines in their thirst-quenching repertoire.

In reviewing lifestyle profiles we found that Segment 3 members were more outgoing and active with a genuine penchant for experimentation. This would influence the type of below-the-line activity that our client considered, such as sponsorship of social or sporting events and theme nights at hotels and nightclubs. Television program directors in Australia are now using this type of lifestyle information to develop programs relevant to target groups. From a media planning perspective program environment issues can be weighed against the lifestyle profile of the target consumer to attempt to optimise both delivery and medium environment.

Slide 28 provides an overview of media habits for Segment 3 sparkling wine drinkers in relation to all sparkling drinkers and the target demographic — men and women 25-45.

TV is undeveloped in Segment 3 and with sparkling wine drinkers generally. As a result, careful program selection is needed. A closer inspection of the data indicates that the position is even more critical if non-commercial channels are removed from the composite television index. Newspapers and magazines perform better than the demographic but an important performer for Segment 3 is clearly cinema. Additional data on movie classifications enabled us to pinpoint the types of movies in which our client should place cinema advertising. This outcome had a significant effect on the overall weighting of commercial television in the media mix and the allocation of monies to alternative media opportunities. It was however the ability to handpick the optimum programs for Segment 3 that ensured television’s dominant role in the media plan.

Segment 3 performs strongly with a range of magazines well above a level predicted by sparkling wine drinking generally. In particular, the magazine Australian Geographic outperforms its demographic readership by 250%. Of the conventional women's magazines, New Idea delivers a solid percentage of Segment 3 drinkers. Its major competitors perform far less impressively.

Slide 31 demonstrates that for television advertising a demographically based plan would be a total failure against Segment 3 and only slightly better for sparkling wine drinkers in general. Commercial television under-delivered against our target by close to 50% if based on a demographic plan.

Men and women 25-54 and its delivery in Segment 3, then similar plan based on optimising delivery to Segment 3

Top rating programmes only have been used to construct this plan
Programmes initially ranked on the demographic to achieve 200+ tarps then ranked on the target segment to achieve 200 tarps

An alternative plan based on the actual television consumption of Segment 3 was then instigated to achieve optimum reach in our target.

Slide 32 shows the differences in ratings between the top commercial programs for Segment 3 and for people 25-54 generally. Careful program selection enabled us to provide reasonable delivery despite the poor performance of television overall.

Slide 33 tracks the effect of the campaign for our client’s commercial brand depicting awareness levels in the month prior to launch and in the subsequent two months during activity. Brand registration was clearly on the up with Segment 3 during the period of advertising.

Depicts month prior to commencement of campaign and two months during campaign
Subject: Client Commercial Brand

More important, however, is —did the campaign work? Slide 34 highlights growth in penetration in Segment 3 of our client’s commercial brand during the period the campaign. There appears to be a straight-line correlation here.

With the help of the fully integrated data source approach and the good fortune of having a unity of data we have been able to:

Identify the critical market segments for our client’s brands,

Understand brand usage in detail. Provide justification for a 20% price increase,

Improve media efficiency over a demographic based plan by 50%, and then concurrently

Measure on advertising impact and awareness, and ultimately the

Market outcome.

As marketers demand greater accountability from their advertising investment, they will look beyond the core services that specifically serve the advertiser or the media.

This is an opportunity for the media omnibus studies to refine their product to provide added value, both to the market and media owner. Let us reiterate once again. We do not see the advertiser or the broadcaster being willing to modify or compromise their core market evaluative data source. This is a significant departure in thinking from the concept of single source.

Instead we see a real opportunity for a secondary, parallel data bank that enables both parties to more fully understand the consumer, their purchase and media habits to make the advertising investment more effective.

However, without a fully integrated data source such as MLI, it is questionable in today’s fragmented media market, if an advertiser can clearly demonstrate the return on the advertising investment. The generic media tests have too high a failure rate. We just have to be more focused.