Data that tells you what happened
Twelve weeks of a Swedish brand's China launch, exactly as a +8h client sees it: daily numbers, the events behind them, and the actions they triggered. The tour plays itself, or you can step through the scenes.
The real Signal cockpit behind this demo. Password: Demo123456
Demo Brand AB is a fictional demo brand; every number is synthetic.
Loading the tour…
What happened, and what we did about it
A dashboard is only useful if it changes decisions. Seven moments from the launch, each with the reading, the goal check, and the action it triggered.
Soft launch on Tmall Global
Demo Brand AB goes live with 48 SKUs. Traffic is thin and CAC is high, expected while the store builds review volume and search relevance.
On plan. Launch quarter is about baseline and data, not efficiency; blended ROAS below 2 was priced in.
Hold spend steady, collect first reviews, prepare the KOL seeding wave.
Xiaohongshu KOL seeding wave
Seeding drives a 4x traffic surge. Conversion dips to 0.8%: discovery traffic saving and following, not buying yet. Favorites run 3x baseline.
On plan: this wave carries an awareness KPI, not a sales KPI. Watch the delayed conversion, not the day-of ROAS.
Retarget the saved/favorited audience over the following two weeks; measure the delayed conversion.
First Douyin livestream test
A single evening livestream lifts the day to 2.4% conversion at a 3.2 day ROAS, the best acquisition efficiency of any channel so far.
Beats target. Livestream commerce earns a permanent slot in the channel mix.
Book a second test before 618, then a weekly cadence through H2.
618 mega festival peak
The pre-sale adds convert: ¥900k GMV in the 72-hour window, a ¥458k peak day at 2.9% conversion and ¥536 AOV. Peak-day media ROAS holds at 3.3.
June closes at 113% of the ¥1.45M plan.
Post-festival: watch refunds and logistics SLA through the delivery wave.
Hero SKU sells out
The festival oversells the hero line: 7 SKUs go offline, conversion halves to ~1.0% through Jun 28, and refunds spike as delivery promises slip.
Risk to the July plan. Every festival stresses the replen cycle. This one found the weak link.
+8h action, Jun 22: cut media spend 55% mid-flight, expedite the replen PO, renegotiate the safety-stock rule with the trade partner (TP).
Replenishment lands + 2 new SKUs
Stock restored in 9 days (vs the 15-day standard). Spend resumes; conversion recovers to 1.6–1.8% and July tracks ahead of the raised goal.
July goal raised from ¥900k to ¥1.05M on the recovered run-rate, and tracking ahead of it.
Lock the new safety-stock rule; pre-build festival replen POs for 8.8 and Double 11.
First member day (CRM)
A members-only drop to the ~26k-follower base lifts conversion to 2.1% with only a modest spend bump. The store’s own audience is starting to compound.
Ahead: organic + CRM share of sales is closing on half, pulling blended ROAS past 4.5.
Monthly member-day cadence; grow the follower base as a stated KPI alongside GMV.
Ask your data. Get grounded answers.
Every +8h reporting setup ships with an MCP connector for Claude. Ask questions in plain language. The answers are read live from your governed metric catalog, the same numbers as the dashboard.
- Answers cite the exact metrics and periods they were read from, so they stay anchored to your data rather than guessed.
- One governed catalog: blended ROAS means the same thing in the dashboard, the weekly review, and every Claude answer.
- Works wherever Claude does: desktop, mobile, or inside your own team's workflows.
claude → mcp: query_metric(shopTotalSales, 2026-06)
claude → mcp: list_goals(client: Demo Brand AB)
claude: "June closed at ¥1.64M, 113% of the ¥1.45M goal. The 72-hour festival window contributed ¥900k…"
