Why a Product Master Now? ~Mr. Kobayashi, who once worked on the product master at a major convenience store, talks about what the ideal product master looks like~

In this article, I would like to explain the importance and possibilities of what I call the “product master,” based on my own experience and with various observations. I hope it will be of some help to your business.

1. Introduction

To begin with, let me talk about my background. I say this because I believe that the experiences through which my feelings toward the “product master” were formed are one of the key points.

To get straight to it, I joined a major convenience store chain in 1997 and worked there until 2023. After that, I moved to a different industry and joined a company that operates a SaaS business. During my convenience-store days, I experienced store operations, launching new businesses at headquarters, data analysis, and building data-analysis infrastructure. In particular, I inevitably became deeply involved with the product master in work related to data analysis. I think many of you reading this article also have some kind of connection to the “product master,” and I will talk about my views on the product master from a retail perspective, shaped by those experiences, as well as my outlook going forward.

2. What is the product master?

Simply put, a product master in retail can be understood as data that compiles the essential information about each individual product needed as it is shipped from manufacturers, placed on shelves, and sold. For example, it includes information such as JAN codes used for barcodes, product names, the name displayed on receipts (sometimes abbreviated due to limited print space), the number of units in a lot when products are delivered to stores, product size, delivery temperature range (whether refrigerated, etc.), product classification, product launch date, manufacturer name, and so on.

3. The product master so far

As explained in the previous section, the “product master to date” (hereafter referred to as “Product Master 1.0”) has evolved by gradually adding the information needed to sell products. However, its main purpose has been optimized to carry out operations such as delivering products to stores, placing them on shelves, and processing them at POS registers. At the same time, Product Master 1.0 has been localized by each retailer and has undergone its own unique evolution. For example, product classification—specifically, potato chips would be categorized in a hierarchy such as “Confectionery > Snacks > Potato Chips”—differs greatly between supermarkets and convenience stores. This is because the breadth and depth of the assortment differ. For instance, a convenience store may carry only around three types of natto, whereas a food supermarket may have 15 or more. In this case, for convenience stores, simply having “natto” as the classification is sufficient, but in supermarkets

Natto (3-pack)
Natto (hikiwari)
Natto (small-grain)
Natto (large-grain)
Natto (cup)
Natto (domestically grown soybeans)

and so on. In other words, this reflects the policy for product assortment. For that reason, product classifications at retail companies are often treated as confidential—because they are essentially the overall business policy.

4. Why did the issues with Product Master 1.0 become apparent?

I believe there are two major reasons why the issues with Product Master 1.0 became apparent. The first is the evolution from POS data to ID-POS data. The second is that environments for analyzing big data have become better established. In the era of POS data, it was enough to aggregate by product category to see which products sold to what extent... or rather, it would be fair to say that, including computer resources, that was all that could be done. However, as retailers introduced membership cards and began to treat customers as individuals, or IDs, and as cloud computing and AI made big-data analysis easier to perform, the shortcomings of Product Master 1.0 became apparent. The aggregation and analysis axes needed to capture customers were overwhelmingly lacking in Product Master 1.0.

The basic aim of introducing membership cards is to realize CRM by treating customers as individuals. At this point, it is common to capture (that is, analyze) a customer using basic member attributes (gender, age, place of residence, family composition, etc.) and purchase history (ID-POS data). What can be visualized using this general attribute information is “who (gender/age) bought what (product), when and where (store), and how many times (repeat rate).” However, what we really want to know is missing here. That is, “why did they buy it?”

There are several ways to answer the question, “Why did they buy it?” The most classic approach is to ask customers directly through a questionnaire about their reasons for purchase. However, surveys take time, effort, and money. The method I eventually arrived at was to ask, “What kind of values did the customer who bought it have?” For example, if a customer who values not spending time on housework buys a bathtub cleaner that removes dirt without scrubbing, it would make perfect sense. When I tried to analyze using this value-based approach, the problems with Product Master 1.0 became apparent. That’s right—Product Master 1.0 contains nothing related to values such as “not wanting to spend time on housework.”

5. The product master of the future

Now that the environment for analyzing big data is beginning to come together, building a “product master” that also includes analysis as a purpose (hereafter referred to as “Product Master 2.0”) has become an urgent task. In building this Product Master 2.0, a common approach is to add product DNA.

Product DNA is data that attributes product characteristics based on consumers’ lifestyles and values—such as attitudes toward health, food, and housework (diet-oriented, safety-oriented about food, etc.), sensitivity to information (liking new products, not wanting to be behind trends, etc.), and values regarding shopping (sensitive to sales, okay with paying a little more if it is a good product, etc.). By using this product DNA as aggregation and analysis axes, we infer the customer’s values.

However, assigning product DNA to each product comes at a very high cost. For example, in the convenience-store industry, about 200 new products are launched every week; imagine the work of buyers assigning several dozen product-DNA items to each and every one of those products. Of course, because everything must be assigned in bulk at the time of introduction, tens of thousands of products are involved, making it an overwhelming task. It goes without saying that it would be even more difficult for supermarkets and home-improvement stores, which carry a wider variety of products. Incidentally, I have experienced this overwhelming task twice. The first time, several of us, including myself, carried it out manually, and it was truly a mind-numbing job. The second time, fortunately, we implemented it with some structure in cooperation with an external partner, so it was not quite that level of work, though it still incurred considerable financial cost.

6. The ideal product master

I believe there are two points to watch out for when adding product DNA. The first is whether the items set as product DNA are appropriate. For example, suppose there is a product-DNA item called “environmentally conscious.” Nowadays, SDGs are widely accepted and generally recognized, so it would make sense to include it as DNA. But what about 15 years ago? Product DNA needs to change its items in line with the social climate and consumer awareness of the time.

The second is that the product DNA of each product continues to change. The value a product has can change depending on the presence of competing products and the passage of time. For example, a new product later becomes an existing product, and in the case of Megmilk Snow Brand’s “Mainichi Honebuto,” when it was first launched in 1993, its main target was children in their growth years, but later the target shifted to age groups that need measures against osteoporosis... in other words, the value provided by the product = product DNA changes. To use a product master that includes product DNA built and introduced for analytical purposes, constant updating is necessary. Considering that, Lazuli’s “Lazuli PDP” automatically builds and continuously updates various product-related information using product names and JAN codes as keys, and can solve the two issues mentioned above. Isn’t it perhaps the service that is currently closest to the ideal for building the desired product master—what you might call Product Master 3.0?

7. Closing

As a side note, I first learned about Lazuli when the company was just starting out. I was in the process of building Product Master 2.0 myself, and when I saw the online news, I learned that they were aiming to provide a service related to the ideal product master I had in mind. I was so excited that I couldn’t sit still and contacted them directly. I should also add that this connection is what led me to write this article.

At Lazuli, based on siloed product masters, we develop and provide “Lazuli PDP,” a solution that generates and processes product DNA and other product data necessary for corporate data utilization, such as what we call “categories” and “feature tags.” As the amount of product data needed for initiatives such as providing information to customers, launching EC, and data analysis continues to grow, if you are looking to centralize and automate those tasks, please contact us from product data.

About Lazuli PDP: https://corporate.lazuli.ninja/feature/