1. The "Formula - Faction Madman" in the NBA Data Circle: Huolingge's PER and BAD Values
In the NBA data analysis circle, Brother Huoling is well - known for "trusting formulas rather than words." He never uses vague evaluations like "a certain player is very hard - working" or "a certain player plays selfishly." Instead, he breaks down each of a player's actions into numbers. In the early years, he became well - known through the PER (Player Efficiency Rating). This index integrates "positive contributions" such as points, rebounds, and assists with "negative consumption" like turnovers and fouls through a weighted calculation, and then divides the result by the playing time to directly calculate the "real value of a player per minute." For example, a guard who averages 20 points per game but has 5 turnovers will have a lower PER than a guard who averages 18 points per game but has only 1 turnover because the PER doesn't consider "padding statistics" but the "efficiency of each point contributed."
Recently, he has come up with an even more extreme idea: the BAD value (Team Disruption Index), which is specifically designed to expose the disguise of the "ball - hog". This indicator calculates the difference between the "ball - possession rate" (such as a usage rate per possession exceeding 30%) and the "true shooting percentage" (efficiency considering shooting difficulty), and then adds the "opponents' scoring rate off turnovers". It transforms the "toxic behavior" of "holding the ball and playing recklessly" and "satisfying oneself while causing the team to lose" into a quantifiable number. For example, a certain star averages 25 points per game, but has a 40% ball - possession rate and only a 42% true shooting percentage. The BAD value can directly tell you that "his scoring comes at the cost of the team's loss" - this is more heart - wrenching than the criticism of "he is selfish".
2. The premise for data not to deceive: Cover every "action - result" link
Zhang Gongzi, a basketball commentator, has penetrated the essence of data: "Data can only avoid being deceived by the impression-based view if it can restore 'the output efficiency of each input'." Traditional data only focuses on "results" (such as points and rebounds), but ignores the "process". For example, is a player's score from an open shot created by a teammate's pick-and-roll or from a foul caused by a strong drive? Is a rebound an offensive rebound (which allows for an immediate second chance at offense) or a defensive rebound (which only protects the ball possession)? These "process data" are the key to distinguishing between "truly excellent" and "falsely impressive".
Brother Huoling's logic is very simple: Each touch of the ball by a player is an "input" (ball possession, opportunity), and the corresponding result is an "output" (points scored, assists, turnovers). Only by quantifying the entire "input - processing - output" link can we break free from the subjective judgment of "so - and - so plays beautifully" and restore the true value of the player. For example, a player's "screen quality" can be calculated by "the increase in the shooting percentage of teammates after his screen"; a player's "defensive value" can be measured by "the decrease in the shooting percentage of opponents under his defense". These "detailed data" are the core of the fact that data doesn't lie.
3. The "pain points in identifying people" in the production workshop: From experience-based judgment to data-driven decision-making
If we apply this logic to the production workshop, the problem becomes even more urgent. Evaluating outstanding employees based on "the word - of - mouth of senior employees" or "judging by how hard - working they seem to the naked eye" often overlooks "hidden costs". For example, a senior "skilled worker" may have a high output but a defect rate three times that of new employees. An operator who "works overtime every day" may have 20% more equipment downtime than others (because they don't know how to quickly handle minor malfunctions). Only data can expose these "blind spots of experience".
The production workshop needs to "identify the right people" and break down "excellence" into quantifiable action indicators. For example, for the operators, there are production volume (the number of pieces completed per hour), defective product rate (the number of unqualified pieces per 100 pieces), abnormal handling time (the repair speed when the equipment breaks down), and cooperation degree (the waiting time for connection during team assembly). For the supervisors, there are the team's compliance rate, the response speed to abnormal events, and the skill improvement rate of subordinates. These numbers will not show favoritism to "senior employees" nor miss out on "silent masters." For instance, a new employee who just keeps his head down and works with a 0.5% defective product rate and a production volume 15% higher than the average, the data will directly place him in the position of an "outstanding employee."
4. Three practical values of the data system: competition, quick wins, and system upgrade
When these indicators are converted into comparable scores, the management of the production workshop will become "dynamic".
I. Replace the closed - door selection with transparent competition: Convert production volume, defective product rate, and response speed into monthly efficiency scores and post them on the workshop bulletin board. Master Zhang got 92 points because he shortened the time of a certain process by 15%. Master Li got 68 points because his defective product rate was 8% higher than the average. Without the need for leaders to scold, employees will find their own gaps: It turns out that I'm slow because I didn't check the part tolerances and It turns out that he's fast because he optimized the tool placement.
II. Rapidly replicate "implementable small wins" (QuickWin): For example, Aunt Wang on a certain production line has a rework rate of only 0.5% (while others' is 3%). By analyzing her operations, it's found that she checks the part tolerances before assembly, while others assemble directly. After promoting this habit to the entire production line, the rework rate dropped by 2% that month. This is the "low - cost and quick win" brought by data, which can improve efficiency without major overhauls.
III. Solve systemic problems with projects: For example, if the overall good product rate in the workshop is only 85%, it's not the fault of any individual, but a flaw in the process design (such as the parts storage area being too far from the production line, which makes it easy for parts to get bumped during assembly). The data will tell you the problem lies in the process rather than the employees are incompetent. Then, carry out a layout optimization project. Move the storage area next to the production line, and the good product rate will directly increase to 90%.
5. Landing logic of the SC workshop: Anchor the core with data and respond to changes with fluidity
This week, we are going to launch this "data-driven employee evaluation system" in the SC workshop. There are essentially two core goals: retain those who can take on responsibilities and maintain the flexibility to adapt to changes.
Target core employees: Use data to identify the "irreplaceable people" - for example, the operators of key equipment (imported CNC machine tools), whose abnormal situation handling time is 40% faster than others and equipment downtime is 30% less; or the senior employees who can quickly perform changeovers. When dealing with order changes, they can shorten the changeover time from 2 hours to 40 minutes. These people are the "stabilizers" in the workshop. We will use data to provide them with "special incentives" (such as additional bonuses and priority for promotion) to ensure they won't leave.
Maintain a reasonable turnover rate: Instead of "randomly laying off employees", use data to determine "which employees need optimization" — for example, employees whose defective product rate has been 15% higher than the average for three consecutive months and have shown no improvement after training, or those who are slow to respond and hold the team back. Maintaining a turnover rate of 10%-15% can not only bring in young people who can operate new equipment but also cope with order fluctuations (hiring more people during peak seasons and making adjustments during off - seasons).
The "knob" of flexible production: When orders change from "mass production" to "small-batch customization", our indicators can quickly switch - from "prioritizing output" to "prioritizing changeover speed" and "prioritizing the passing rate of customized parts". The data system is like an "adjustment switch", enabling the workshop to adapt to orders instead of being driven by them.
The essence of this system is not to "assess people with data", but to use data to restore "who is making real contributions" and "who is holding back" - to turn the vague "excellent" into clear "numbers", so that employees know "how to work to get better" and managers know "how to retain talent to stabilize the overall situation". After all, what the workshop needs is not "people who seem excellent", but "people who can really help the workshop make money".