How New Analytic Models Are Improving Government Trade Promotion Efforts

As we know from recent news events, governments are highly engaged in the global trade activities of their national, regional, and local industries. Governments typically offer considerable assistance to companies within their jurisdictions to identify export opportunities, secure needed financing, navigate trade-related bureaucracies and regulations, and promote their goods and services to foreign markets.

As the field of data science advances, we are seeing data play ever-expanding roles in improving program effectiveness while enabling more efficient resource allocation. An exciting recent development in how data is helping to propel global trade can be seen in Canada where government agencies are employing specialized data models to identify companies that are well-positioned to benefit from various trade promotion and assistance programs.

Dun & Bradstreet’s Export Propensity Model: Connecting the Dots, Accelerating Trade

Specifically, Canadian agencies are using Dun & Bradstreet’s Export Propensity Model to identify specific Canadian-based companies that are not yet exporting but probably should be because they possess all the traits of an export-ready enterprise. This information is critical for governments looking to connect trade-ready companies with the right government programs that can help them quickly transition from a purely domestic-focused company to a thriving international vendor.

What we do is we look at the companies that we know are exporting and ask, 'what is unique about those businesses in contrast to their non-exporting peers?'
Steve Munden, Director for Government Solutions, Canada, Dun & Bradstreet

“Canada is finding companies out there that look like they should be exporting but aren’t. Most governments don't have a handle on that,” said Steve Munden, Dun & Bradstreet’s director for government solutions in Canada. “Governments would rather put more of their budget into helping exporters as opposed to spending a lot of money on advertising trying to find those potential exporters. So we built an Export Propensity Model that a government can use and say, ‘we want to find all of the companies in our jurisdiction that have the propensity to export but are not exporting today.’ So our modeling and analytics team filtered through our business and trading data to develop a model that effectively identifies those businesses so governments can contact them directly about their various trade assistance programs. This allows them to divert marketing dollars to more effective purposes so more money ends up going into the programs as opposed to the marketing — and they see better results.”


The model works by compiling data derived from exporting companies — and since Dun & Bradstreet collects data on more than hundreds of millions of companies, that’s a lot of data. "What we do is we look at the companies that we know are exporting and ask, 'what is unique about those businesses in contrast to their non-exporting peers?'" Munden said.

Actionable Insights Made Possible by Advanced Analytics: The Nuts and Bolts

Mark Seiss, a director on Dun & Bradstreet’s advanced analytics services team who works on the Export Propensity Model, explained: “We leverage data from known exporters and use that commercial data and business activity data to create a model that predicts a company’s likelihood to be an exporter.” The model then looks at non-exporting companies to find data profiles that approximate those of companies that export. The model then ranks how many of those data signals appear for a given company and then compiles a rank score to indicate how ready a particular business is to export.

Export propensity scores range from one to five, where a five signifies a company ready to export right away while a lower score suggests a company that may want to export but perhaps lacks certain capacities or knowledge to jump quickly into the global marketplace. For a government agency, such insight is invaluable because it helps identify and segment the business audience for various types of government assistance, Munden said. A company with a lower propensity score, for example, likely will need more educational assistance to better understand and navigate the path to becoming an exporter. They may need help finding markets, weighing shipping options, filling out paperwork, or understanding the advantages that come with being a small business. Conversely, a company with a high propensity score likely will need more tactical help, such as financial assistance to scale production or trade promotion to penetrate specific markets.

Mission-Based Analytics Result in More Effective Government Programs

The Export Propensity Model is a great illustration of how data and data analytics are increasingly empowering governments at all levels to become far more precise and proactive in applying their services and resources to achieve better economic development results. This applies to programs designed to expand tax bases, develop high-quality talent pools, and better position local economies to take advantage of emerging new industries. “It is really critical that governments avoid looking at their businesses in a vacuum,” Munden said. “Only by looking at their businesses in depth over time can governments truly anticipate changes in those businesses and the larger ecosystems in which they operate. With such insights, they can bring a great deal more precision and efficiency to their development efforts for better outcomes.”