Connecting consumer understanding to marketing ~ consumer panel data utilization
Marketing research is sometimes defined as “the function of connecting marketers with consumers, customers, and the general public through information” (Ref.American Marketing Association). The fact this definition says “through information” instead of “through data” can be interpreted as being indicative of the premise data utilized in marketing activities will be informatized. Data utilization can thus be considered deriving information from data.
At the same time, marketing activities require consumer information to considerand evaluate measures with. This article will delineate methods for gaining information on consumers from data, and how to leverage this in marketing.
The term data actually encapsulates a variety of different varieties of data. Broadly speaking, data can be divided into ① open data, ② in-house data, and ③ third-party purchased data, as seen in Fig 1.
Fig 1
Both ① open data and ② in-house data can be obtained either for free or at little expense. The first, ① open data is characterized as delivering macro information, but as being hard to obtain micro information from. Conversely, the second, ② in-house data, is characterized as delivering down to micro information, but as being hard to obtain big picture information from.
On the other hand, while ③ third-party purchased data must be purchased, it delivers information at an granularity, from macro to micro.
Fig 2
Third-party purchased data is an effective way to obtain the information we need with brand marketing. As presented in Diagram 2 as a merit, it offers users to bird’s eye view of the market, delivering a clear grasp on their position in market. One form of third-party purchased data offering a wide range of information on consumers is “consumer panel” data, which captures consumers’ state of purchase.
Consumer panel data is a database constructed using a methodology called “panel surveys” which collects and accumulates the same information from specific samples over a long period of time. The typical way this is done is by having monitor panelists – consumers – record the JAN code, place of purchase, purchase value, time/place of purchase etc. for each product they purchase.
Linking this data with each monitor panelist’s demographic information, consciousness data regarding daily life, and product information data via JAN codes delivers highly granularized information.
Let’s think once again about what consumer panel data delivers. The market is created through the accumulation of individual consumers acts of “paying money”. (Fig 3) Whether or not consumers “pay money” for a firm’s products determines that firm’s sales levels as well as the firm’s market share. In order to affect change in a business, consumer behavior needs to be changed. Consumer panels can also be considered databases recording the results of consumer purchase behavior on a daily basis, serving as a starting point for consumer understanding.
Fig 3
2. How is consumer panel data utilized?
From here, we will use actual data to explain what sort of information consumer panel data utilization offers, and what sorts of evaluations and decision-making this can inform. As a case study, let’s conduct a comparative evaluation from the consumer’s perspective of cosmetics brand A and competitor cosmetics brand B , sold in the same price range.
Let’s firstly examine the characteristics of purchasers of brand A to do xxx with. Fig 4 uses data from INTAGE’s consumer panel SCI to score* their state of consciousness and behavior in areas where values related to beauty are apparent, including “beauty and dietary lifestyle care”, “health foods and supplements”, and the “state of their teeth”. We compared purchasers of skin lotion A and the total average, and extracted the attributes that characterize them. * Consciousness attributes are asked about in terms of degree of correspondence on a 5-point scale, with their T2B composition ratio used as their scores.
Fig 4
Next, Fig 5 depicts the scores of skin lotion A purchasers and skin lotion B purchasers, as well as the difference between them on the same consciousness attributes as Fig 4.
Skin lotion A and skin lotion B are both products in similar price ranges with average prices of 1,600 yen. However, skin lotion A purchasers’ purchasers dietary lifestyle care and health foods and supplements-related consciousness scores were over 10pts higher than those of skin lotion B purchasers, indicating that they are highly conscious not only of beauty, but health through diet as well.
Fig 5
Does this difference in users’ consciousness between brands affect their brand loyalty? Let’s take a look at this using an indicator called share in purchasing household. Share in purchasing household is defined as the average “percentage the purchase value for a certain brand comprises of the total purchase value for a certain category” for purchasers of a certain brand. *INTAGE’s panel data provision system, iCanvas uses the indicator name “Share in Purchaser”.
For example, if Ms. A who purchases skin lotion A purchases 3,000 yen of skin lotion A in a year, and Ms. A purchases 30,000 yen of skin lotion from all the brands of skin lotion she uses including brands other than skin lotion A, then the proportion skin lotion A will comprise of Ms. A’s skin lotion purchase value will be 3,000÷30,000 = 10%.
Similarly, if Ms. B who purchases skin lotion A purchases 3,000 yen of skin lotion A in a year, and doesn’t purchase any skin lotion other than skin lotion A in a year, then the proportion skin lotion A will comprise of Ms. B’s skin lotion purchase value will be 3,000÷3,000=100%.
While Ms. A and Ms. B’s skin lotion A purchase values are the same, from the standpoint of purchase behavior, Ms. B can be interpreted as displaying higher loyalty than Ms. A.
On interpreting these results from a brand perspective, brands with a high share in purchasing household can be considered to be winning higher loyalty. Fig 6 depicts a comparison of skin lotion A and skin lotion B’s share in purchasing household over the past five years.
While skin lotion A is deemed to have higher loyalty than skin lotion B in all years, its average growth rate on loyalty over the past 5 years (CAGR) has been 1.02 each year, which can be interpreted as meaning there has been no fluctuation in this degree of change.
The ability to analyze chronological change like this is another characteristic of panel data, which “continuously collects data”.
Fig 6
Consumer panel data also enables us to understand brand loyalty as consumer purchasing behavior. The purchase value of a certain brand by consumers can be broken down to: purchase value = no. of purchasers × no. of purchases per purchaser, purchase value per purchase
Fig 7 is a breakdown of skin lotion A’s and skin lotion B’s respective purchase values based on the outlook above. It is apparent that while the purchase value per 100 people is higher for skin lotion A than skin lotion B, that skin lotion B has a higher number of purchasers.
skin lotion A has a higher no. of purchases per purchaser (purchase frequency) than skin lotion B, and a greater purchase value per purchase (spend per customer), and thus has a higher purchase value than skin lotion B. skin lotion A purchasers’ purchasing behavior may be related to the higher loyalty for skin lotion A than for skin lotion B.
Fig 7
Based on the results of the analysis to this point, it can be concluded that skin lotion A is purchased by consumers who are more conscious of not only beauty by health through diet as well, and possesses higher loyalty than skin lotion B. This high loyalty may also be linked with purchase frequency and spend per customer. If there are large numbers of people who don’t use skin lotion A who are “highly conscious of not only beauty but health through diet as well”, targeting these sorts of people would be a potential direction to ensure business growth with.
3. Elucidating the strategy to take from consumer panel data
From this point onwards, I would like to introduce a way use the share in purchasing household indicator mentioned above to analyze the performance of existing brands and to identify issues with user composition.
Fig 8 was produced through modeling changes in the number of purchasers and share in purchasing household when sales of an existing brand increase.
When sales of existing brands increase, the number of new purchase is first deemed to increase, as depicted in the diagram to left in Fig 8. New purchasers’ share in purchasing household of the brand is generally not as high as that of existing purchasers, so when sales initially start to increase, the brand total’s share in purchasing household average will decrease to below its level before sales increased (herein referred to as its current state).
Once new purchasers start re-purchasing, sales increase further, and as depicted in the diagram to right in Fig 8, new purchasers’ share in purchasing household is deemed to increase. If this happens, the brand total’s share in purchasing household average will also increase.
Fig 8
Fig 9 models the changes in numbers of purchasers and share in purchasing household when existing brands’ sales decrease. When existing brands’ sales decrease, as depicted in the diagram to left in Fig 9, its purchasers with low share in purchasing household, or to rephrase, purchasers deemed to have little loyalty to the brand, are deemed first to leave. Relatively high share in purchase household purchasers stick with the brand among its current purchasers, so the brand total’s share in purchasing household average is deemed to increase to over its current state. Even if they don’t leave the brand, if their number of purchases continues to decline, sales will further decline, and as depicted in the diagram to right in Fig 9, share in purchasing household is also deemed to decrease. If this happens, the brand total’s share in purchasing household average will also decrease.
Fig 9
Fig 10 models the current state as a starting point, and combines changes in share in purchasing household and changes in sales values. Cases on par with the current state are depicted in gray, those that are increases above the current state are in blue, and those that are decreases below the current state are in red. The outlook here is to interpret brand performance through a total of 9 combinations of 3 patterns for each indicator, with ① depicting the most positive state, and ⑨ the most negative state.
Fig 10
Let’s take a look at an actual brand performance evaluation using this outlook. Fig 11 depicts year-on-year changes in share in purchasing household and sales value for brands A through E in the chocolate market for a certain period.
Fig 11
In this diagram, 95% ~ 105% versus the previous year are interpreted as remaining unchanged, with other scores either an increase or decrease from the previous year. Brands A and B are ranked around ⑤ ~ ⑦, so are deemed at a stage where low share in purchasing household customers are leaving the brands, so winning over new customers is the next action to conceivably take. Brand E is ranked ⑧, with relatively high share in purchasing household customers are leaving, so winning over these lapsing customers again is the next action to conceivably take. On the other hand, brands C and D are ranked ① ~ ②, so can be interpreted as making progress with winning over new purchasers and repeat purchases.
The next action conceivable would be promoting further repeat purchases and setting next new purchaser targets.
In this way, organizing the relationships between data obtained from consumers provides a grasp on the state brands are currently in, and elucidates the next actions to take.
Summary
In this article, we explained how consumer information can be obtained from consumer panel data and leveraged in brand evaluations, as well as how to connect this with next actions to take. In order to affect change in a business, consumer behavior needs to be changed, and without consumer understanding, this will be difficult to achieve. We would be delighted for you to add consumer panel data utilization as a means to understanding consumers.
◆This report is copyrighted by INTAGE Inc. Please check the following prohibitions and precautions, and specify the source when reproducing or quoting this report. “Source: INTAGE “Knowledge Gallery” article published MM/DD/YYYY article”
◆The following are prohibited: ・Alteration of this article in full or in part ・Sale or publication of this article in full or in part ・Uses that are against public order and morality, and uses connected with illegal activities ・Reproducing or quoting panel data* for the purpose of advertising or promoting companies, products, or services *Panel data include: SRI+, SCI, SLI, Kitchen Diary, Car-kit, MAT-kit, Media Gauge, i-SSP, etc.
◆Other precautions: ・INTAGE Inc. shall not be liable for any trouble, loss, or damage caused by the use of this report ・These usage rules do not restrict the use of quotations or other uses permitted under the Copyright Act of Japan
◆For inquiries about reproduction and quotation, click here
