Translator Role Aligns Business and Tech Teams, Saves Time and Money
George Bernard Shaw once described Britain and America as “two countries separated by a common language.” The same could be said of the business and technical teams charged with delivering a new data-driven project that’s expected to produce significant benefits for the business.
​
Deep technical expertise is essential, but it’s not enough. Sometimes you need a translator too.
Branch Point is a Canada-based retail and distribution business serving customers in the manufacturing industry.
The company needed much better performance from its inventory management and forecasting system. The challenge was to warehouse and integrate data that was siloed across various SaaS systems, making it easier to understand existing inventory, significantly improve the accuracy of demand forecasts, and automatically generate high quality inventory recommendations.
“The previous set-up made us very reactive,” says Andrew Calderon, Operations Manager at Branch Point. “It was difficult to predict where a problem was going to occur before it happened.”
The business engaged a team of developers to build the new system. Branch Point decided they needed to add an additional set of skills to keep the project aligned with the evolving business needs, and maximize the business value delivered.
“We needed an independent, no-skin-in-the-game point of view, a hybrid role of fractional project lead and deep subject matter expert,” Andrew says. The person would function as a bridge and “translator” between the business and the technical teams. He engaged Analytics Strategies to fill the role.
Over the course of the project, Analytics Strategies helped build a bridge solution while the data engineering team slowly built the infrastructure. We built the supply chain models and helped manage the data engineering team.
Along the way there were a variety of technical needs to help the project move along. The initial prototype supply chain models were built in PowerBI and focused only on the recommended shipments to Amazon (which represented a significant majority of total sales).
Once the initial data was loaded into the warehouse by the data engineering team, Analytics Strategies began migrating the supply chain models into the new platform. The work involved both validating the data from the source systems and building models to forecast sales across the distribution network.
A key challenge was managing products which were sold as a kit instead of singles. If a single product was sold both as part of kits and as an individual unit, it was time consuming to do the analysis and see if there were enough on hand to meet all demand. To create a system which would be easy to maintain we developed an effective rationing procedure that would ensure most stockouts were avoided. If there were enough raw materials to meet the normal shipment quantities, the recommendations instead followed the regular models.
The project has significantly reduced the time to produce forecasts, and linked those forecasts to analytics-based inventory management, while cutting the number and duration of out-of-stock items. “Now we can do in three hours what used to take 60,” Andrew reports.
He expects to cut even those times in half as the project matures, and he credits the proprietary models that we built for much of the drastic reduction.
The “translator” aspect of our work was particularly valuable, he says adding that it takes a rare combination of skills. “In that role, there are very few people who can confidently and competently communicate between the developers and the business,” Andrew tells us.