The Weekly Ecommerce Ops Dashboard For India Teams

Most weekly operations reviews in Indian ecommerce businesses run on forty numbers and produce no decisions. The fix is not a better dashboard tool. It is fewer numbers, named owners, and a meeting that discusses only exceptions.

The numbers a weekly review should run on

Twelve to fourteen metrics, grouped into five blocks. If a number does not sit in one of these blocks, it belongs in a monthly pack.

Orders:

• Orders received and net order value, split by channel.
• Pre-dispatch cancellation rate.

Fulfilment:

• Same-day dispatch percentage against your stated cut-off.
• Pick and pack accuracy. Median operations sit at 98 to 99 percent, a competent one at 99.5, top quartile at 99.9.
• Inventory accuracy from cycle counts.
• Ageing of unshipped orders, measured in hours past promise.

Delivery:

• On-time delivery against the promise shown to the customer. 95 percent is healthy, leaders run near 98.
• First-attempt delivery rate. Above 80 percent is upper tier, below 75 percent signals address quality or last-mile execution problems in specific pin codes.
• NDR resolution within 24 hours.

Returns:

• RTO rate, split by payment mode. India’s D2C average sits around 20 to 30 percent, roughly 23 percent across large mixed samples, and COD-heavy fashion can touch 40 percent.
• Customer return rate, which runs 15 to 35 percent by category and 25 to 35 percent in fashion and ethnic wear.
• Return-to-resale cycle time. Best in class is 48 to 72 hours.

Inventory:

• Days of cover and stockout rate on A-class SKUs.
• Fill rate or OTIF to marketplace and quick commerce partners.

Cadence and who owns each number

Three rhythms, not one.

• Daily, fifteen minutes, standing: five numbers only. Yesterday’s orders, same-day dispatch, unshipped ageing, NDR queue, A-class stockouts. This meeting exists to catch a fire, not to analyse one.
• Weekly, forty-five minutes, seated: the full set, exceptions only.
• Monthly: trends, money and structural decisions. Cost per order, contribution after logistics, courier mix, SKU-level profitability.

Every metric gets exactly one named owner. Not a team and not a function. Write the name next to the number on the dashboard. If two people own a number, nobody does. If the owner cannot influence the number, you have assigned it to the wrong person and you will receive explanations instead of actions.

Ownership also decides where a number is measured. If same-day dispatch comes from a warehouse spreadsheet and on-time delivery comes from the courier panel, the two will disagree and the review will spend twenty minutes on reconciliation instead of decisions. Pick one system of record per block: the order management system for orders and fulfilment, the courier or aggregator panel for delivery, the WMS for inventory. Where a number has to come from a partner, agree the extraction method in writing at onboarding rather than arguing about it in month four. The owner of a metric owns its source, its definition and its refresh time, not only the commentary attached to it.

Leading and lagging indicators

Most operations dashboards are built entirely from lagging indicators, which is why the reviews feel like post-mortems. Pair each lagging number with the leading one that moves it.

• RTO rate is lagging. Address quality score, prepaid share and pre-dispatch confirmation contact rate are leading.
• On-time delivery is lagging. Cut-off compliance, courier pickup time and manifest handover time are leading.
• Pick accuracy is lagging. Share of certified staff on shift and audit sample coverage are leading.
• Stockouts are lagging. Forecast error and purchase order acknowledgement time are leading.

Set a control band for each number rather than a single target. Green, amber, red, with the thresholds written down. A metric sitting inside its band gets no discussion time at all, however interesting the movement looks. This is the discipline that keeps a forty-five minute review inside forty-five minutes.

Keeping the review from becoming a status meeting

The failure mode is predictable. Somebody reads numbers aloud, everyone nods, no action is recorded, and the same slide reappears the following week. Five rules prevent it.

• Circulate the pre-read twelve hours before. Nobody presents numbers in the room. If you have not read it, you do not attend.
• Open with last week’s actions. Every one gets closed, extended with a stated reason, or killed. Do this before any new number is discussed.
• Discuss only what is outside its control band. A green metric gets zero minutes, even if someone wants to celebrate it.
• Every exception leaves the room with one owner, one action and one date. If you cannot name all three, the item is not ready and moves to a separate working session.
• Hard stop at forty-five minutes. A review that runs long is carrying analysis that should have happened offline.

Two habits keep the dashboard honest over time. First, retire any metric that has not generated an action in eight weeks. It is reporting rather than management, and it can move to the monthly pack without loss. Second, write down the definition of every number, including which timestamp starts the clock and which orders are excluded. Half the arguments in Indian ops reviews are definitional rather than factual, and they disappear the day the definitions sit on the same page as the numbers.

A dashboard is not a report card. It is the shortest list of things that, if they move, the business moves. Twelve numbers, twelve owners, one meeting a week that ends in dated commitments.