In the first post in this series looking at D&B's global supplier data strategy, I shared the high-level methods D&B is using to enable its global data expansion and to achieve 350 million commercial entity records by the end of 2014, including a focus on China and other emerging markets. One of the hurdles D&B faces in each emerging market region -- not to mention others targeting global supplier information and ratings -- is the challenge of data quality, given the sources of available information. D&B is attempting to overcome this challenge in a number of ways, including leveraging government information, insight and particulars from its own large customer base, and data and web-mining techniques. D&B's David Clarke agrees that more should and can be done to obtain global company information. That is why D&B is also exploring established and emerging technologies such as Wiki, sentiment analysis, semantic disambiguation, and more to build out its arsenal of global data.
One of the fundamental underlying challenges of building insight into global supplier information in non-Western languages is that of translation and transliteration. I've actually done quite a bit of work in this area in an unrelated software field. The fundamental issue of name transliteration is that a single name in Arabic, Chinese, etc. could translate in potentially dozens of ways into a Western script -- and each of these translations could be appropriate! The same can be said for the Western concept of naming which is vastly different from how names are expressed in other parts of the world. This is one of the major hurdles we face in the war on terror, and why terror suspects have been allowed on airplanes despite their presence on watch lists. On a less dangerous level, the same can apply to suppliers as well (or key individuals in the case of court cases, fines, etc.).
D&B fully recognizes this challenge and shared with me on our call that they recently received a patent for "ideogrammatic matching" in content, allowing them to match words in both local language and in a translation/transliterative context as well. I'm not in a position to judge how solid their approach is in this area as I have not seen the results or worked with D&B supplier data sets, but I do hope to have additional feedback and commentary on it in the future, especially as I get reference feedback from D&B users I know.
Outside of its transliteration approaches, D&B follows another model which, in their words, "emulates what we do in the West" when it comes to leveraging a combination of public and private sector information on companies. Payment information remains key in this context on a global basis. In D&B's words, the intersection of timely payment information is a "predictor" of how a company will pay in the future (which is not necessarily the same as whether or not a supplier will remain solvent, mind you). Still, looking at "cash, bank receipts, how a company manages its money and bill payment records" can help D&B -- and by extension other credit/supplier information providers and their customers -- to triangulate toward risk information. The secret sauce, of course, is obtaining accurate underlying company information, which in emerging markets, is easier said than done!
Consider some of the nuances of China and India that I learned from my call with D&B's David Clarke. According to David, each country plays by its own set of rules (which we all must follow). In India, for example, The Information Act precludes the use of data for certain reasons. Moreover, the availability of inexpensive labor in countries like India and China has led to significant use of manual processes around information collection and usage in the past, which can create its own unique set of challenges (I suppose fuller employment and poorer underlying data quality is something developing markets are willing to tolerate in place of more consistent, but costly automated approaches). This challenge requires D&B to do additional research on companies even after they have government or company-derived information.
People don't just exist in massive numbers in emerging markets. High-level individuals also matter in boardrooms and governments, and in some ways are even more important than they are in the West. In China, for example, D&B has found that the way key "individuals" are dealt with can make the difference between success and failure. D&B has learned that getting information from China depends as much on the execution of relationship building strategies as it does on corporate objectives. Just as it is in manufacturing in China, D&B has found that "the relationship" matters most in the content world as well.
What other approaches is D&B taking in China and other markets to gain access to business -- including supplier -- information? Stay tuned.