Real-time analytics in the manufacturing industry

NHL hockey is perhaps the fastest and most action-packed sport on the planet. With skating speeds up to 30mph and pucks regularly clocked over 100mph the tempo allows for virtually no time to stop and think and instead necessitates continuous split-second decisions. This year, in an attempt to have insight into the chaos, the NHL released an in-game app exclusively for coaching staff that collects 60 real-time statistics and analytics of individual player performance. The manufacturing industry—operating in a similar environment in which time-critical decisions present themselves at every point—has also been through this journey before with the recent introduction of real-time analytics on the shop floor. In both these environments, real-time access to data and analytics can often be the difference between success, shortcomings, and missed opportunities. As we continue through playoff season there are the obvious questions —will this new technology change the great game of hockey for players, coaches, and fans, especially in the fight for the Stanley Cup? What lessons can the league take from the introduction of real-time analytics into the manufacturing industry?

Lesson 1: New technology changes the decision-making process

The app will provide real-time analysis of 60 statistics, changing how coaches view their own players and the competition. Overall, coaches will have to decide which of the available statistics to focus on, how often they are reviewed, and how much influence the information will have on their decisions. In manufacturing, data can often be interpreted with relative ease because there is more control over the environment. In hockey, these data points become more complex and less controllable for desired outcomes. Players can’t always predict their opponents’ actions and require constant adaptation to the environment. However, with more access to real-time and historical data coaches can gauge the better probability of outcomes from their decisions.

Example 1: Pulling goalies from the game

Historically, when coaches were trailing in a close game they were reluctant to pull their goalie with less than one minute to go for fear of a blowout loss (e.g., 1-5 v. 1-3 final score). However, by taking the goalie out within the final minutes, they would actually increase their probability of winning. This decision typically reflects poorly on the team and coaching staff, even though statistically speaking all the historical data would increase win probability overall.

Example 2: Faceoffs

Historically, coaches selected which lines and players to put on the ice based solely on points (goals, assists) or perceived confidence (certain players just “seemed” better, or were having better games). However, with real-time data available, coaches are gauging how successful players are at winning these throughout the season or even within a given game. Now it’s not uncommon for coaches to put a certain player on the ice solely to take (and hopefully win) a faceoff and then change with a player that has a better matchup or skill needed for the remainder of the shift. A simple outcome, like which team wins a certain faceoff, can be the difference between a win and a loss. Similar lessons have come from the manufacturing industry. Better availability of data has changed how decisions are made in manufacturing as well. For example, by tracking real-time material efficiency manufactures can predictively reduce waste, scrap, warranty, and returns volumes. These can often make/break an efficient line run and change how manufacturers decide to evaluate changeovers and internal processes/capabilities, for example. Bottom line: Overall, by including data coaches can support the more rational outcomes that result in higher chances of success for their organizations if they’re willing to put pride aside and consider when to put pre-conceived opinions on reserve.

Lesson 2: New technology comes with a people impact

NHL coaches should be aware of the potential effects new technology like in-game tracking will have on their players and teams. In Industry 4.0, automation impacted the day-to-day jobs of many shop floor workers. When this new technology was introduced the workforce had mixed feelings—there was relief that mundane tasks were eliminated but lack of trust in the data and fear that jobs could be completely replaced by automation. It was important for manufacturing leaders to ensure these changes would not negatively impact company culture. While automation can’t replace the talent on the ice, there could be hesitancy among players to have every movement tracked, recorded, and scrutinized. While the metrics will give additional insights into how a player is performing, this intel could become overmanaged. Much like in manufacturing, it will be up to the coaches in the league to ensure the proper team culture and player attitudes are maintained.

Lesson 3: New technology requires higher investment in exchange for a greater return

The NHL built their app two years ago. While development dollars and costs to teams have not been released, there is no doubt that it was a significant investment. Beyond quantitative returns, there are qualitative returns that the league will benefit from in similar ways to the manufacturing industry introducing real-time shop floor analytics. Insight into time-on-ice and other real-time statistics will help prevent injuries: 

• Insight into on-ice time and other real-time statistics will help prevent injuries: Coaches can identify patterns and have more informed decision-making to refine strategies real-time based on how their team and competition is performing. If they notice a player has had an especially high time-on-ice during a game, they can take measures to reduce their shift times or pull them from the game early to reduce the chance of injury from overexertion. Likewise, manufacturing leaders can track asset utilization metrics to determine when machine repairs or maintenance is required to avoid longer or unpredicted machine downtime that would occur if the preventative service was not performed.

• Historical data will provide a base for key trends: Over time multiple seasons’ worth of data can help a team to define future strategies, player performance, and other key elements to a winning team. Similarly, manufacturers have been able to rely on KPIs to inform them of asset utilization, maintenance, etc. that impact game-winning strategies in the industry. The greater the volume of historical data to refer to and compare against the more confident coaches and manufacturing supervisors alike can be on their decisions.

• Team staff can focus on analysis: With the introduction of the in-game app the team staff can spend more time analyzing data vs. collecting and ensuring the quality of data. No longer does the team need statisticians to manually tally each shot on goal; instead they can simply glance at the shot count and use it as a decision criterion for whether to change goaltenders. Likewise, the production floor doesn’t need workers with stopwatches to track uptime, but can instead receive the input automatically and then interpret it from there.

As you are enjoying the remaining NHL playoff series and rooting for your hopeful Stanley Cup champion, remember that just like in manufacturing, each and every action is a data point that can be tracked, analyzed, and acted upon for better production and results the next time around. May the journey be filled with high-flying sauce, board-rattling hits, and bar-down snipes that lead to sick cellys. Just don’t forget all the data that went into – and will come out of – these glory moments!