Driver Performance

Driver Performance is where you answer "who's carrying the fleet and who needs a conversation." Open it from Performance → Drivers in the left sidebar. The page is titled "Driver Performance" and subtitled "Identify your highest and lowest performing drivers," and everything on it is driven by the week range in the top right.

That range is two dropdowns, Start Week and End Week, and it always snaps to whole weeks — you never pick a mid-week date. It opens on the last five weeks (four weeks back through the week you're in), which is usually enough history to tell a bad week from a bad habit. Change it and both tabs on the page redraw against the new range.

The first tab, Team Performance, is four side-by-side leaderboards for the selected weeks.

Overall Ranking is the headline. Each driver gets a single Overall Score, and the table lists them Rank, Driver, Overall Score, best first. Next to it, Agent Call Compliance ranks drivers by the percentage of Viki's calls they responded to, Netradyne Hits ranks them by number of infractions, and HR Incidents ranks them by incident count. Each of those three shows Rank, Driver, and the value.

The Overall Score is worth understanding before you act on it. It's a relative score, not an absolute grade: for each metric you've chosen, the app compares every included driver against the best and worst performer in the range, flips the direction for metrics where lower is better (Netradyne hits, DPMO), then blends them using the weights you've set. A score of 100 means "best in your fleet on these metrics over these weeks," not "perfect." If a driver is missing data for one metric, that metric is simply left out of their score and the remaining weights are rebalanced, so a driver with a gap isn't unfairly punished.

Under the Overall Ranking table there's a Configure button. It opens "Configure driver ranking," where you pick the metrics that make up the score and give each one a percentage weight. The weights must add up to 100% — there's a "Distribute weights equally" shortcut if you'd rather not do the arithmetic — and each metric can only be used once. Out of the box the score is split evenly across overall tier, CDF DPMO, packages delivered, and Netradyne infractions, but the picker offers the full catalogue of per-driver metrics, grouped by how they're measured, so you can weight the score toward whatever your station is being graded on this quarter.

The same dialog controls who is ranked. "Include all drivers in the ranking" is on by default and covers every active driver. Turn it off and you pick specific drivers from a list — useful when you want new hires or a particular route group scored separately. Save, and you'll see a confirmation that the ranking was updated; the tables recalculate with the new weights.

The second tab switches from leaderboards to one driver at a time. Pick a driver from the Driver dropdown and you get four stacked tables, each laid out with metrics down the side and one column per week in your range, so trends read left to right. A cell showing "..." means there's no data for that driver that week.

Scorecard Performance covers overall tier (shown as the usual Fantastic/Great/Fair/Poor rating), total Netradyne hits, CDF DPMO, DCR, and DSB DNR. Non-scorecard Performance covers shift confirmation rate, days not confirmed, AI agent call compliance, and HR incidents — the operational behavior Amazon's scorecard never sees. Work Contribution shows packages delivered, days rostered, and hours worked, which is the context that stops you from comparing a five-day driver to a two-day driver. Incentive Performance lists the incentives that driver was part of during the range, with each one's current status.

Once a driver is selected, a second dropdown appears so you can put another driver's four tables side by side — the fastest way to show someone that their week really was an outlier. Hide Comparison clears it again.

Team Performance has a Download CSV button above the tables that exports all four leaderboards as one combined file, including any custom metrics you weighted into the score. Every individual table on the page also has its own download icon in its header, which exports just that table for the selected weeks. Files are named for the table and the week range, and on the Individual tab, for the driver as well.

This page doesn't replace the Scorecards page — it reframes it. The scorecard figures here are the same Amazon weekly numbers, cut per driver and lined up week over week instead of week by week for the whole station. Netradyne counts are the same camera alerts you work through on the Safety page, totalled per driver so you can see who is generating them repeatedly rather than handling each alert in isolation. Nothing you do here changes an alert, a scorecard, or an incident; the page is for spotting the pattern, and you go to those pages to act on it.