Skip to content

Data checklist before the 11.11 and 12.12 sales

20 data tasks in the 4 weeks before 11.11 and 12.12 to measure true profit, catch errors early and avoid arguing about numbers after the sale.

By The DataxLand team· · 6 min read

Preparing for 11.11 and 12.12 — the Nov 11 and Dec 12 "double-day" sales that are the biggest shopping events of the year on Shopee, TikTok Shop and Lazada across South-East Asia — usually revolves around stock, vouchers, livestreams and ad budgets. The data side tends to wait until later — and then in mid-December, when the whole team asks "did 11.11 make or lose money?", nobody can answer because cost of goods was not updated, ads were not assigned to channels, and returns are still coming in.

This checklist collects the data tasks to do before the year-end sales, split across four weeks. The goal is simple: when the sale ends, you have one true profit figure that the owner, marketing and accounting all trust, instead of three figures and an argument.

Why prepare the numbers before, not after

Three things can only be done before the sale:

  1. A baseline. To know how much the sale added, you need a clean "normal" period to compare against. Picking the baseline after the sale makes it easy to pick the period that favours the conclusion you want.
  2. Naming conventions. Ad campaigns, voucher codes and bundles named by convention from the start can be filtered later; ad hoc names mean checking every row by hand.
  3. Cost of goods with effective dates. Stock bought specifically for the sale often has a different cost. Without an effective date, the new cost overwrites old orders or vice versa.

Week −4: clean the foundations

  • Update the cost-of-goods table for every SKU you will sell during the sale, with the effective date of each cost.
  • Check cost coverage of revenue. SKUs without a cost are counted as if cost were zero, which inflates profit. In the DataxLand sample data (a simulated scented-candle shop, 1 Aug – 29 Oct 2026), this ratio is 92.35% — nearly 8% of revenue has no cost of goods. Bring it close to 100% before the sale.
  • Declare bundles and gifts: components or a separate cost for each bundle code.
  • Reconcile SKUs across marketplaces and your inventory software; a row with an unknown SKU is a row with no cost of goods.
  • Check the order status map: have new statuses (for example a new cancellation type from a marketplace) been mapped to cancelled, returned or delivered?

Week −3: lock the baseline and conventions

  • Choose the baseline period: for example the 4 weeks before the sale, excluding days with big promotions. Write it down so everyone uses the same baseline.
  • Name ad campaigns by convention with a sale prefix (for example 1111-, 1212-) and the destination sales channel.
  • Name voucher codes by convention for each programme so the sale's voucher cost can be separated later.
  • Record the fee schedule in force for each marketplace and the fee-bearing programmes you sign up for (vouchers, free shipping). Marketplace fees change several times a year; double-check in your shop's Seller Centre (see Marketplace fees in 2026).
  • Compute break-even ROAS by channel from current contribution margin, to set an ad-spend ceiling for the sale (see What ROAS is profitable?).

Week −2: set up where you read the numbers

  • One sale tracking page with 6–8 tiles: orders, net revenue, contribution margin, ad spend, profit after fees, cancellation rate, return rate, provisional fee rate. All by day and by channel.
  • A daily report sent to the team chat early each morning, for the "Yesterday" period.
  • Alerts for sale days: cancellation rate above a threshold, daily ad spend above the ceiling, an unusual drop in orders on one channel (often a sign of a sync error or a hidden listing). For setting thresholds that are not noisy, see Anomaly alerts: when to act and when it is noise.
  • Test-run the reports and alerts with current numbers to make sure the content is right and reaches the right group.

Week −1: dry run and roles

  • Refresh data on a tighter schedule during the sale days, so the morning report shows yesterday's numbers.
  • Run the data checks once more: duplicate rows, mismatched SKUs, unmapped statuses. The five most common errors are listed in 5 sales data errors to check.
  • Assign who reads what: who watches ad alerts, who watches cancellations and returns, who may switch a campaign off.
  • Agree on one headline number for the wrap-up: profit after fees and after ads, based on net revenue, not GMV.

After the sale: do not close the books too early

Sale-day profit is not true profit for at least a few weeks, for these reasons:

FactorWhy the number still movesWhat to do
Provisional feesMarketplace fees are only final once the order completes and is settledWatch the provisional fee rate fall; do not close while it is high
ReturnsReturns come back over several weeks after deliveryClose profit once most orders are past the return window
Pulled-forward revenueBuyers wait for the sale, so the week before sells lessCompare the whole before, during and after period with the baseline
Front-loaded adsAd spend ramps up sharply before sale dayAssign spend by sale campaign, not only by date

For the full method of reading the results — baseline, pulled-forward revenue, profit after returns — see After 11.11 and 12.12: measuring the true profit of a sale.

Five wrap-up questions you should be able to answer

  1. What was profit after fees and ads for the sale period (including the week before and after) compared with the baseline?
  2. Which channels made a profit, and which only made revenue?
  3. Which best-selling SKUs lost money after vouchers and programme fees?
  4. How many percentage points higher were cancellation and return rates during the sale than on normal days?
  5. Which ad campaigns had ROAS below break-even?

How DataxLand does this

DataxLand keeps cost of goods by effective date, calculates profit after fees with the same definition for every channel, and shows two health metrics on the dashboard: cost coverage of revenue and the provisional fee rate (see Metrics and the semantic layer). The checks after every build catch duplicate rows and mismatched SKUs (Verified numbers).

A sale tracking page is built by copying the overview page and adding tiles (Workspace pages). Daily reports and alerts on thresholds, period-over-period changes or anomalies are set up in Reports and alerts and sent to Telegram, Slack or Microsoft Teams. After the sale, ask Si "profit after fees 1–20 Nov versus the previous 4 weeks, by channel" for a wrap-up figure with sources. See also the multi-channel profit and loss solution page.

Frequently asked questions

How long before the sale should I start preparing the numbers?

About four weeks is enough for small and mid-sized shops: the first week cleans cost of goods and SKUs, the second locks the baseline and conventions, the third builds the page and alerts, and the last is a dry run.

Can I use GMV for the sale wrap-up?

GMV tells you scale, not profit. Sales come with discounts, vouchers and higher return rates than normal days, so the gap between GMV and net revenue is bigger than usual. Wrap up with net revenue and profit after fees.

How long after the sale can profit be closed?

When most orders are past the return window and the provisional fee rate is low. Until then, report the figure with a note that it "may still change".

Do I need a separate dashboard page for the sale season?

Yes, it helps. The everyday page has many long-term metrics; a sale page only needs a few metrics by day and by channel so you can read it quickly on the busiest days.

Read next: P&L dashboard: 8 metrics to track · Automated reports on Telegram · Retail and consumer goods

Related posts

Verified dashboards from your own data, this afternoon

Free plan with no time limit · No card required · Sample workspace ready in minutes