Insights

Organized Chaos Global Data Harmonization
Article

Organized Chaos Global Data Harmonization

Joe Palermo, V.P. Global Services Product Leadership, The Nielsen Company

SUMMARY: Think it’s impossible to truly harmonize data from multiple countries, given the variations in barcodes, brand names, packages, products and sizes involved? While meeting the challenges of global data differences is daunting, it can be accomplished with a unique combination of local data access, global understanding and a design and implementation plan that delivers action-oriented applications.

The data to make decisions is hardly global in nature…

Major fast-moving consumer goods manufacturers and retailers are making decisions about their businesses on a global scale, but the data they use to make their decisions is hardly global in nature. In fact, consumer-packaged goods data is often derived from multiple sources of information across as many as 100 countries in local languages with disparate product and category descriptions and even secondary desk-top research data included.

Distilling the differences

So what options exist for a global manufacturer or retailer when faced with the reality that data doesn’t match across countries, divisions, or formats? Technology and applications are components of harmonization, but not the drivers. In light of the fact that global product catalogues are rare, coding within company divisions and across countries varies, and product ownership, manufacturer names and even product uses are not universal, data harmonization is a process that requires experts with local access and global experience to follow seven key steps.

A major obstacle to global data harmonization is multiple product codes…

Step One: Establish a detailed set of rules

A major obstacle to global data harmonization is the confusion created by multiple product codes, descriptions, manufacturer names and categorizations. With 85% of the EAN (European Article Number) or UPC (Universal Product Code) codes in local databases unique to their country of origin, it can be next to impossible to report on a product’s sales from market-to-market.

Some brands have different owners in different countries, and manufacturer names can differ across borders. Manufacturers may use terms such as “multi-use” to mean 1- or 2-liter bottles in one market, but 750mL in another, based on the necessities of manufacturing.

Product usage also varies by market, and some markets describe a product as it physically appears—not how it is used. Chocolate sprinkles, for example, are used as a topping for cakes or ice cream in the United States and Belgium, but as a sandwich spread in the Netherlands. Likewise, compote is a dessert in southern European countries and a meal accompaniment in northern Europe. Local coders may miss these distinctions since they do not realize that a product is used differently elsewhere.

The key to overcoming these challenges is establishing a clear set of rules for input coding in order to collect quality information. These rules must be generic and based on what the product actually is allowing for de-culturalization, but also must provide sufficient information to allow flexibility.

Eliminate the ambiguity that results from different interpretations…

Step Two: Avoid ambiguity

The coding structure must be supported with detailed descriptions and definitions—particularly for multi-country coding—in order to eliminate the ambiguity that results from different interpretations. In addition, the coding structure must be in the local language with careful translations that use terminology understood within the local business community. For example, there are three words that could mean “ambient” in Germany: Raumtemperatur, Umgebungstemperatur or Konservierung bei Raum, but the local Nielsen business uses Kamertemperatuur as the best term for what would be identified as ”shelf stable” in the United States.

Mandate only one way of coding…

Step Three: Implement a system

Once the detailed definitions are in place, it’s important to mandate only one way of coding any given item. This enforces the disciplined use of values assigned to product characteristics and ensures the same values are used consistently for the same product characteristics across countries.

Step Four:  Build in quality checking

The coding system must include additional data for enrichment purposes from a variety of sources to address data gaps. Multiple approaches can be used, such as allowing copy facilities from similar products, or enriching centrally by facilitating coding of information which has not been supplied.

Distinguish each product using a set of characteristics and a unique numeric number…

Step Five:  Establish a unique product identification

Integral to the success of the global data harmonization system is the ability to distinguish each product using a set of characteristics and a unique numeric number. Barcodes and GTIN (Global Trade Identification Number) numbers have traditionally served this purpose, but for a number of reasons they’re not entirely reliable. Often, the same product is marked with different codes, or is coded the same as a product in an entirely different category. And in some cases, promotional packs are given the same barcode as a regularly-packaged product. For example, the Coca-Cola 330mL regular can has 30 different codes across Belgium, the United Kingdom and France. To truly harmonize data from multiple sources and locations, a unique identification number must be created and adopted.

Effective data harmonization requires a separate, central decision team…

Step Six: Create a global governance organization

Effective data harmonization requires a separate, central decision team, with a direct link to local teams with experience, product knowledge and cultural understanding. That central governance team should be independent and consistent in order to manage the changes that result from product or category evolution.

When new and unique products enter the market, the governance board determines where they belong. For example, are under-arm absorbency pads deodorants or clothing protectors? What’s the best way to manage products and categories that are unique to one market, such as squash and cordials in the United Kingdom or Cuberdons in Belgium? How should unique packages be handled, such as mayonnaise in plastic bags, offered in Denmark, or pickled eggs in a jar, from the United Kingdom?

Step Seven: Keep it simple and provide centralized support

As databases can quickly fall out of sync without proper maintenance and updates, the data harmonization team must consistently assign global codes for new items in the market and a project manager to perform multiple ongoing tasks, including: ensuring that codes are properly applied and are consistent across countries; supplying new codes when necessary; resolving internal and external issues; clarifying borderlines, communicating results and assessing whether local country changes have impacted the data synchronization as a whole.

Think local, act global

Harmonization is an ongoing process that requires strict governance and constant monitoring. But when the process is in place and is drawing from both local knowledge and global expertise, it provides high-quality information and a solid foundation for an integrated information system.

With extensive, worldwide experience, Nielsen’s global project teams are well versed in the harmonization process and are uniquely qualified to discern the global needs of both manufacturers and retailers, providing consistent views that support individualized client needs. “One size fits all” strategies just don’t work globally. Global harmonization provides the clarity needed to move beyond data differences toward integrated information that drives better decisions on a global scale.