Usage rate is a workload stat, not a skill stat
Usage rate gets cited as if it were a measurement of star caliber. A player with a 32% usage rate sounds important. A player with a 19% usage rate sounds like a role player. The broadcast convention treats the number as a proxy for how much of the offense is being run through someone, and from there as a proxy for how good they are. That second step is the part that breaks. Usage rate measures the share of possessions a player consumes when they are on the floor. It does not measure whether the team was better for it. The single highest usage rate in the league last season belonged to a player whose team finished 24-58. That is not a coincidence. It is a structural feature of how the stat works.
What usage rate actually counts
Usage rate is defined as the percentage of team possessions a player ends while on the floor. A possession ends with a field goal attempt, a trip to the line, or a turnover. If you took the shot, drew the foul, or coughed it up, the possession is on your ledger. If your teammate did any of those things, it is on theirs. The stat is normalized so that a five-man lineup always adds to 100%, which means high usage on one player guarantees suppressed usage on the other four.
Nothing in that definition tracks whether the play worked. A contested 28-foot pull-up with three seconds on the shot clock counts the same as a wide-open layup off a kick-out. A possession that ends in a turnover counts the same as one that ends in a made three. Usage rate is volume of finishing actions, weighted by nothing. The stat is doing exactly what it was designed to do. The problem is that the broadcast framing has wandered well past what the stat measures.
The bad-team inflator
The fastest route to a high usage rate is to be the only viable shot creator on a team without other options. When Detroit ran their offense through Cade Cunningham last season, his usage rate climbed into the top five in the league. The reason was partly that he is a real lead guard with star tools. It was also that the alternative on most possessions was a rookie wing forcing a contested mid-range or a center who could not create his own shot. The defense knew where the ball was going. He knew where the ball was going. He still had to take it.
A high usage rate on a roster like that is a description of the team's offensive poverty as much as a description of the ball-handler's gravity. Move the same player to a roster with two other plus shot creators and the usage rate falls into the 24-26 range, because there are other places the ball can legitimately go. The player did not get worse. The supporting cast got better. That is the part of the equation that usage rate cannot see.
What gets buried by the totals
The number that does the actual work in evaluating offensive impact is true shooting percentage at a given usage rate. High volume at low efficiency does not produce wins. The league has a long history of guards who posted 28% usage rates on 52% true shooting and were genuinely valuable, alongside guards who posted 30% usage rates on 50% true shooting and were actively dragging their offenses down. The two profiles look almost identical in a usage-only table. They look very different in any model that prices possessions.
Russell Westbrook's late-career years are the canonical example, but the same shape recurs every season at the second-banana level. There is always a wing on a non-contender averaging 22 points on 28% usage and 53% true shooting whose agent will cite the points and the usage in a max negotiation. The team paying the bill cares about the true shooting line. The team that already paid usually does not get a second chance to revise.
Why role-player usage gets undercounted the same way
The mirror image of the high-usage critique is the low-usage one. A 3-and-D wing whose usage rate sits at 14% is not necessarily a bad offensive player. They might be a 45% three-point shooter generating gravity on the weak side and finishing the few possessions that come their way at an elite clip. Their usage is low because the system does not ask them to create. Their offensive value is high because the system gets a lot from the actions they do take.
Usage rate cannot distinguish that player from a 14% usage wing who simply cannot generate offense. Both look like role players in a usage table. Both are role players. One is a meaningful positive contributor and the other is a drag. The only way to tell them apart is to layer efficiency, positional value, and lineup data on top of the usage figure, which is what front offices do and what broadcast graphics usually do not.
What usage is genuinely useful for
The stat does carry real information, just not the information it gets used to deliver. Usage rate is useful as a description of role and as a fairness adjustment when comparing efficiency stats across players. A 30% usage scorer on 58% true shooting is doing a harder job than a 18% usage scorer on 60% true shooting, because the harder shots come with the volume. Usage-adjusted efficiency models like the ones that drive most modern player ratings explicitly weight true shooting against usage to capture that trade-off.
Usage rate is also a useful indicator of fit and stylistic compatibility. A roster with three players over 28% usage is going to fight over possessions. A roster with no one over 22% usage is probably missing a closer. Front offices use the distribution of usage across a lineup as a quick diagnostic for whether the offense will have a coherent late-game identity. That is a legitimate use of the number. Treating the highest single value on the team as evidence of star quality is not.
The honest read
The right way to read a usage rate is as a description of role weighted by supporting cast. A 28% usage rate on a contender means the offense is built around that player and the alternatives are good enough that the choice was real. The same 28% on a lottery team means the alternatives were worse and the player took what was left. The skill content of the rate is not knowable from the rate alone. It only becomes knowable when you pair it with efficiency, with on-off splits, and with whatever the rest of the roster could plausibly have done with the same possessions.
The stat that ate the offense is not automatically the stat that drove the offense. Sometimes it is. Often it is not. Usage rate cannot tell you which case you are looking at, and the broadcast convention of citing the rate as evidence of impact keeps mistaking workload for skill. The players doing the most are not always the players doing the most good. The league's salary cap sheets are full of the difference.