Not long ago, consumer goods companies’ retail coverage strategies were based on two variables: budgets and store counts. All stores were considered equal. Then, advancements in technology, such as product potential indices and geo-demographic modeling, enabled retail coverage budgets to be allocated based on other variables and the measurable impact of having a merchandiser in the store. Shortly thereafter, strategies were strengthened by on-shelf availability systems that incorporated exception-based performance thresholds defined by anomalies in unit movement and dollar sales.

Now, retail expertise is being bolstered by the power of new technologies — such as machine learning’s speedy, complex calculations applied to big data — to bring new efficiencies to retail services. Atlas Technology Group is testing an intelligence-based retail coverage system, dubbed RORI (Return On Retail Investment), that combines machine-learning measurements with immediate on-site feedback to optimize retail-related investments. The system, now being piloted by select manufacturers, calculates the return on retail investment for each store and each in-store activity in real time. “RORI integrates retail coverage patterns and budget, store activity and labor data to provide logic-based assessments, enabling on-the-fly changes to the priorities, tasks and time spent in each store based on their economic impact — ensuring every minute is maximized,” Atlas CEO John Matise notes.

A Case in Point

Retail sales specialist Taylor begins each day by reviewing RORI’s initial plan on a tablet. Today, RORI has planned five store visits for Taylor and projects the return on investment (ROI) opportunity for the day to be $6,075. The system also provides the optimal routing sequence to minimize windshield time. However, Taylor knows RORI may update the plan based on variable inputs provided during each call or by external factors such as weather and traffic patterns. RORI’s plan has Taylor beginning with store No. 4108 and spending 45 minutes accomplishing 10 in-store activities with an ROI opportunity of $1,490. After Taylor arrives at the store and checks in, he completes a brief store assessment on his tablet that either validates the initial retail plan or causes RORI to re-prioritize his tasks. At this store, Taylor’s top priority is verifying a holiday display, part of a multi-million-dollar media spend, is on the sales floor. It’s not, so Taylor enters “No.” RORI immediately re-optimizes the in-store tasks to get the holiday display on the sales floor. Using conditional logic and proprietary algorithms, RORI reduces the number of tasks for store No. 4108 from 10 to seven, increases the projected in-store time from 45 minutes to one hour and changes the daily ROI for that store from $1,490 to $3,100. Once the display is in place, Taylor continues to address this store’s priorities. While checking backroom inventory, RORI alerts him that 55 of the 60 minutes allocated for this store have been used. RORI’s dynamic optimization engine has determined that proceeding to the next store offers a higher ROI than taking the extra time to complete the last priority, which is capturing competitors’ temporary price reductions. Before heading to the next store, Taylor sees that RORI has rerouted him based on inputs from the navigation app, Waze. By communicating with Waze in real time, RORI helps Taylor avoid the gridlock reported on two major thoroughfares and re-optimizes his remaining retail activities.

By applying new measurements such as return on retail investment, manufacturers will continue to improve performance and increase their ability to directly influence store-level outcomes. That influence will only grow as predictive retail decision-support systems like RORI, fueled by machine learning, become more intelligent over time.

Note: The RORI proof of concept is under way. Manufacturers interested in participating may contact Dan Lyons at