A sale day war room for Diwali, built before the traffic arrives

There are four lanes. Workflows bring data in. A dashboard shows it. Monitors decide when someone should look. A Relay team does the first ten minutes of investigation and proposes an action, and a shift lead decides.

Lane 1: Data in, as it happens

Everything the war room shows lives in Tables, written by workflows:

• Orders. A Shopify orders webhook points at a workflow’s Webhook trigger. Each order lands in the orders table with value, payment method, pincode, SKU lines and UTM source.
• Payments. Your payment gateway’s webhooks for failed and captured payments hit a second workflow. Gateways such as Razorpay are not native integrations; the webhook and the API block do the job, with the key stored as a workspace secret.
• Stock and ads. A scheduled workflow runs every 15 minutes. The Shopify block lists inventory levels for the sale SKUs, the Google Ads block pulls campaign performance, and the API block reads Meta spend from the Marketing API.
• RTO risk. As each COD order lands, a rule flags it as risky when its pincode has a high return history, the value is above your limit and it is the buyer’s first order. Those orders go to your COD confirmation flow.

Lane 2: One screen

The war room is a dashboard on a TV and on every laptop, with row-level access so an agency partner sees only ad panels. Five panels are enough:

1. Orders per 15 minutes against the same window last year.
2. Payment success rate by method.
3. Spend against revenue by channel, with delivered ROAS from metrics defined once.
4. Stock cover in hours for the sale SKUs.
5. Share of COD orders flagged risky, by region.

Panels that read the webhook tables update as rows land. Stock and ad panels move every 15 minutes. That is near-real-time, and on sale day it is enough: the decisions take minutes, not seconds.

Lane 3: Five monitors, every 5 minutes

Monitors re-check every 5 minutes and alert in-app, by email and in a Slack channel called #war-room:

• Order drop. Orders per 15 minutes fall below half of the expected curve. Often a checkout problem, not demand.
• Payment failures. Failure rate above twice the normal rate for two checks in a row.
• Stock cover. A sale SKU under 3 hours of cover.
• ROAS drop. Delivered ROAS on a campaign falls below your floor after enough spend.
• Risky COD share. Flagged COD orders above your limit in one region.

Lane 4: A Relay team on watch

Start from the Ecommerce ops team template and keep three Relays in Assist mode. Each one starts when its monitors alert, writes findings to the team’s shared notes board and hands off by rule: if the stock watcher finds a SKU about to sell out, it hands the ad question to the ads watcher.

• Ads watcher. Reads Google Ads reports and, through Meta’s ads MCP server connected under Tool connections, campaign insights. A budget or status change is a spend call, so it waits for the shift lead.
• Stock watcher. Reads stock cover and orders, finds which campaigns still send traffic to a SKU that will sell out within hours.
• Payments watcher. Breaks failures down by method and bank, checks whether anything changed, and writes the incident note. It contacts no customers.

By default a Relay handles 25 tasks a day and 2 at once, so plan which alerts each one owns on sale day. The team kill switch pauses all three at once if anything looks wrong.

The rehearsal

Run a full rehearsal a week before. Replay last year’s order file into the orders table through a workflow, let the monitors fire, and put the Relays in paper mode, which simulates spend calls instead of making them. Check that the right person gets each alert and that every proposal makes sense.

On the day, the approval queue is the shift lead’s inbox. Approvals cover the exact change shown, run once and are written to the audit log, so the post-sale review starts from a record of what was proposed, approved and declined, not from memory.

After the sale, keep the dashboard and turn the monitor thresholds back to normal. For the everyday version, see the live store dashboard and ad budget pacing.