The operator who has had coordination delegated for 90 days faces a challenge nobody announces in advance: how to stay connected to the operation's reality without returning to direct presence. The initial instinct is to check the dashboard at random moments or send the coordinator a message asking how things are going. The problem is that neither practice produces a reliable reading: the dashboard at any given moment shows state, not trend; and the coordinator's summary — in good faith — naturally emphasizes what went well and minimizes what was resolved with difficulty. The operator two weeks into delegation may be looking at stable numbers while the operation is quietly accumulating small drifts that only become visible at the scale of a month.
The solution isn't more access to data — it's a structured weekly review with a fixed format, a maximum duration of thirty minutes, and a specific set of questions. A delegated operation's weekly review is worth more than any amount of ad hoc checks because it reads directional signals — is the operation improving, stable, or drifting? — rather than individual incidents, which at the scale of a week are noise. This article describes the practical format of that review: when to do it, which five indicators to read, what questions to ask about each, and how to distinguish results that require a decision from those that only require monitoring.
The silent drift that happens when the operator stops reviewing with structure
The real risk of delegation without weekly review isn't an operational catastrophe — it's the accumulation of small drifts that no coordinator will proactively report, because from the shift perspective each one looks like a manageable one-off situation. Acceptance rate falling one percentage point per week looks like daily stability and is monthly degradation. Average wait times in the northern zone climbing gradually during overnight hours look like normal variation until a customer complaint about excessive wait surfaces and the dashboard shows that indicator has been rising for three straight weeks. The coordinator didn't flag it because on each shift the variation was within the operational range — but the three-week trend never appeared in any shift close as a trend.
The weekly review isn't a control mechanism over the coordinator — it's the tool that gives the operator a temporal perspective the coordinator structurally cannot have. The coordinator operates at shift granularity: what happened in the last eight hours. The operator reviewing weekly operates at trend granularity: what changed in the last seven days compared to the seven before, and what that change means for next week's structure. These are complementary perspectives that produce different value — the coordinator manages incidents in real time, the operator reads the operation's direction and decides when to intervene before drift becomes a problem that can only be solved by direct presence.
The timing and duration that make the review useful
The weekly review doesn't work if it's done when time is available — because time is never available. It works if it has a fixed slot in the calendar: Tuesday or Wednesday morning, before the day's commitments, with the last seven days' shift closes already read in the five minutes before. Tuesday or Wednesday because it has the weekend's overnight period in the rearview mirror without losing the current week's reaction time. Morning because the quality of decisions a review produces is a direct function of available cognitive energy — and a review done between meetings produces shallow readings and deferred decisions. Friday evenings invariably produce reactive reviews of what was difficult that week rather than trend analysis.
Thirty minutes is the correct ceiling, not the target. If the review regularly requires more than thirty minutes, the problem isn't the duration — it's that the data isn't organized for quick reading, or the operator is doing coordination instead of strategic review. The first five minutes are for reading the coordinator's shift closes from the week. The next twenty are for reviewing the five indicators with each one's specific question. The last five are for recording conclusions in three categories: confirmed, being monitored, and decision this week. That record is the only output of the review — and without it, the review produces no asset that can be compared against the following week.
The five indicators that produce a reliable trend reading
Not every indicator on the dashboard is equally useful for a weekly trend review. Some capture operational noise: Monday's trip count varies between weeks without any structural variable having changed. The following five indicators measure variables that change slowly under normal conditions and sharply when something in the operation is drifting:
- **7-day average acceptance rate vs. the prior 7 days**: normal variation is ±1 percentage point. If the difference exceeds that range, there's a cause — and it's worth finding before another week of accumulation.
- **Average wait time in the two or three high-demand zones**: not the global average, but the zones that concentrate 50-60% of trips. If this indicator climbs consistently for two weeks, there's a coverage problem current incentives aren't resolving.
- **Coordinator-to-operator escalations per shift** (weekly average): the trend should be declining as the coordinator matures. If the number rises or stabilizes at a high level, the context file has specific gaps the review should identify.
- **Context file entries this week**: not how many entries were made, but whether any were new rather than threshold corrections. A week with no new entries means the operation learned nothing operationally in that period.
- **Active drivers vs. registered drivers with zero trips this week**: if this ratio worsens two consecutive weeks, there's an activation or retention problem that won't show up in acceptance rate until it's expensive to reverse.
The right questions to read each indicator
The most common error in the weekly review is treating indicators like traffic lights: green is fine, red is bad, yellow bears watching. That framework produces inaction when the number is green and disproportionate reaction when it's red — without the operator understanding why it changed in either case. The right question for each indicator isn't 'is this good or bad?' but 'what changed this week compared to last, and do I have a hypothesis for why?' A bad indicator with a clear hypothesis is more useful than a good indicator with no understanding of what's sustaining it. Acceptance rate that dropped two points because Wednesday brought heavy rain and six drivers cancelled their shifts is completely different from the same drop with no identifiable cause in the coordinator's shift closes.
The operator who reviews with the 'why did it change' question and finds no answer in the coordinator's shift closes has exactly the input they need for the only intervention of that week: a specific message to the coordinator asking for the cause of that concrete variation, not a general analysis of the week. 'Wait time in the airport zone climbed from 6 to 9 minutes between Thursday and Sunday — what happened with coverage in that zone during those shifts?' is a question the coordinator can answer using panel data and shift closes. 'How was the week?' is a question that produces a curated summary that omits exactly what the operator needs to understand in order to decide with judgment.
When the weekly review escalates to an intervention
The weekly review should close with decisions, not worries. The difference between those two outcomes lies in having clear thresholds that convert a concerning indicator into a concrete action. The three thresholds that justify direct operator intervention — not just monitoring until next week — are: acceptance rate falling more than 3 percentage points week over week without an identified cause in the coordinator's shift closes; wait time in high-demand zones exceeding the context file threshold for more than four consecutive shifts; or escalation count that fails to decline for three straight weeks with no extraordinary events to explain it. Any of those three conditions doesn't warrant waiting until next Tuesday — it requires a working session with the coordinator that same week to identify the root cause.
The operator's intervention in a delegated operation isn't returning to the shift — it's asking a different question than the one the coordinator is answering with panel data. 'Why does the southern zone have twice the cancellations it had three weeks ago?' is a question the coordinator can answer using that zone's data but has no incentive to ask spontaneously if the global indicator was within range that week. The operator who intervenes early doesn't interrupt the coordinator's autonomy — they protect it, by identifying a problem before it reaches the level where the only available solution is the owner's direct presence on shift. That's the value difference between reviewing with structure and reviewing out of anxiety.
What the operator records after each review
The weekly review produces value only if it closes with a written record. The operator who reviews mentally, draws conclusions, and moves on builds no asset: they have no history of what they were monitoring last week, can't compare their readings week over week, and have no record of decisions made and their outcomes. The minimum record format has three sections: confirmed — indicators that read within range with an understood cause, where the right action is none; being monitored — indicators that changed with a hypothesis and a named person responsible for reporting the data next week; and decision this week — a maximum of three concrete actions with the responsible person's name and expected result date. Those three sections fit in ten lines of text.
That record isn't bureaucratic — it's the difference between an operator who manages the business from working memory and one who manages it from a system. After three months of documented weekly reviews, the operator has a history that allows them to answer questions that without the record are impossible to answer precisely: in which week did acceptance rate start falling the last time zone expansion caused coverage problems? What incentive decision did we make when wait time exceeded 11 minutes in December, and how long did it take to return to normal threshold? Those answers matter when the same situation recurs — and in a regional ride-hailing operation, situations recur with enough regularity that the history is always relevant.
The first month after delegating I coordinated by instinct — I checked the dashboard three times a day and asked the coordinator how everything was going. It was micromanagement dressed up as supervision. When I started doing the Tuesday morning weekly review with specific indicators and stopped checking the dashboard the rest of the week, the first month was uncomfortable. By the third month I stopped feeling like I was losing control. Now I open the dashboard on Tuesday at nine and by Wednesday I have the two decisions I need to make that week. The rest of the time I can work on what actually moves the business.
The operator who reviews weekly with a structured format isn't doing more work than one who reviewed ad hoc — they're redistributing the same management energy in a way that produces useful signals instead of noise. The difference between checking the dashboard out of anxious habit when the thought 'how is everything going?' surfaces, and reviewing with a five-indicator agenda and thirty allocated minutes, is the difference between confirming existing worries and generating new information. The first consumes energy without producing decisions. The second produces decisions even when the result is 'everything is within range and this week I can focus on commission restructuring without interrupting the coordinator.'
The practical test of a weekly review that's working is what appears in the decisions section. An operator who closes the review with 'everything in order, nothing to decide' for four consecutive weeks should check whether their intervention thresholds are sensitive enough — or whether they're avoiding the signals that require work. An operator who closes with two or three concrete decisions most weeks has evidence that the operation is communicating at the right resolution. That's what well-instrumented delegation produces: not silence, but the right signal at the moment when it's still possible to act at low cost and without needing to reclaim the direct presence the system spent months making unnecessary.
