Brown-Forman Corporation – the company behind famous global spirits brands such as Jack Daniel’s bourbon, Chambord liqueur, and El Jimador Tequila – is raising a toast after adopting a tool that integrates and unifies global sales data for the first time.
Founded in 1870 and still headquartered in the same town of Louisville, Kentucky, Brown-Forman employs 5,700 staff and is the largest American-owned spirits company in the market.
The brand has enormous global success, with sales of its 40-plus beverage brands spanning 170 global markets. Last year, that enviable reach drove sales of more than $4.2 billion. Despite this scale, until recently, leadership was unable to see how Jack Daniel’s Tennessee Honey whiskey was faring compared to Woodford Reserve in a given geography.
In fact, Brown-Forman’s siloed data meant that the central product team lost visibility of 80% of products once shipping, meaning it couldn’t truly see how those products performed in the global market.
Even worse, potentially helpful data that was bought in from consumer behaviour analysis agencies was also siloed, often describing the same drink with different item labels for different countries. To take just one example, the record for ‘Jack Daniel’s’ could be described using any of 2,000 slightly different tags across the dataset.
Mike Homer, Senior Manager, Master Data Management, explains:
Around 2017, I realized we have all these different data sets that we’re being asked to combine, but we also need to buy information from different data sources, and we lose visibility for products once they are shipped out.
Many of our markets are distributor markets, where we sell to a distributor who then sells to their customers. This means we lose visibility of those products once they arrive at the distributor. We then buy data back to get that information, along with competitor data sets, to understand how we are performing against competitors.
It’s easy to combine data sources if we’re talking about two…but if you add in a third or a fourth. Also, a product might be global in name only – so what they call Jack Daniel’s in the United States isn’t what they call it Canada, Poland, or Germany.
That means a seemingly simple question like, ‘How is Jack Daniel’s doing across the world?’ is almost impossible to answer, without a tool to bring those data sets together. In fact, the business hasn’t been able to ask that question before; it simply wasn’t answerable with what we had.
A partner at the right serving size
Brown-Forman had excellent, in-depth visibility of local, in-market product performance. The challenge was how to collate all of that data to provide a single version of the truth.
Homer’s aim was to clear up the issues with international sales and provide a solid reporting basis so that Brown-Forman’s line of business managers could pinpoint the best and worst performing sectors and have more effective localized pricing and promotion.
The solution, he says, has been to implement Master Data Management software from French-headquartered unified data management specialist Semarchy:
Homer’s jumping-off point for trying to identify who could help was the Gartner and Forrester top lists of vendors in the Master Data Management space:
I knew tools existed out there to do that, which would be a better fit for us instead of trying to do this ourselves. The other criteria was a company right sized for our organization and which didn’t have a solution that’s going to be a lot of overhead for us or something that would have to be deeply ingrained in our entire technology stack, but something that was just fit for what we were trying to do.
Taking a shot at unified data
Semarchy came online with no hassles in the midst of the COVID pandemic four years ago and now provides a database of 500,000 ‘harmonized’ products, which are sold all over the world. Homer jokes:
We’re in the spirits industry and it can take four years to make some of our best products. Time was a little on our side so we went at a pace where we felt we could do the work well – learning the tool ourselves and not relying on a third party to do the development. That learning curve cost us a little time, too
Some 105 data inputs are now fully ‘harmonized,’, he adds:
All of the product information and all of our distributor information across all of our markets is flowing into the environment, and we’ve come up with a library of 25,000 products that have a hierarchy of manufacturer information, brand family, and brand information we’ve defined. We’re mapping all of these external data sources back to those definitions.
To help, Brown-Forman has also implemented a fuzzy match logic against the definitions of the words that comprise descriptions of its products to understand if they have been seen before:
As you can imagine, the first couple of data sources uncovered Brown-Forman products we’d never ‘seen’ before. Now, as we bring in a new data source, like from South Africa, we’ll have a data owner from South Africa on point to harmonize those products with our records.
As to how this benefits the business, he says:
Bringing these data sets together means we’re able to analyze any source of information that provides product data to us. For the first time ever, we’re able to look at all of Europe and see how our business is doing against competitors. We’re also looking at things like how consumer takeaway data compares to what our distributors are seeing. Do our shipments match up with what we’re seeing in the environment from a consumer takeaway standpoint?
We can also look at how our own P&L stacks up against consumer takeaway information, or any of the other sources we try to look at. It’s also enabled us to do more strategic pricing based on environment, across an entire region.
Now the business has got these data sets available to them, they’re always finding new and creative ways to use the harmonized data sets.
Next steps
Finally, Homer and his team are actively looking at ‘standing up’ an AI assistant to take harmonization and Master Data Management to the next stage at the company:
Because we’re setting up each of these products individually, it takes a little time to do the research to understand who the owner is and what the relationships are.
Right now, it takes us anywhere from two to 29 minutes to create what we call ‘p-codes’; if we can offset that to an AI, it buys us a bunch of resources so that we can free up time to do other work.
Summing up his experience of Master Data Management, for Homer the conclusion can best be served neat:
Through strategic implementation of master data management, the company that makes Jack has improved. Its entire bar of global business insights.
link