For Amazon Agencies & Brand Managers

Your AMC Data.
Automated. Analysed. Actioned.

Most agencies log into AMC manually, run a query, download a CSV, and repeat. I built a pipeline that does all of it automatically — and adds AI-generated insights on top.

By Shuqing Ke — UK-based independent AMC analytics consultant

✓ No S3 bucket needed ✓ Works with any AMC instance ✓ EU + NA supported ✓ Claude AI insights included

The Problem with Manual AMC

⏱️

Hours wasted every week

Logging in, writing queries, downloading CSVs, formatting reports — all done manually, all repeatable.

📊

Data sits unused

AMC has LTV, NTB, lapsed buyers, time-to-convert — most agencies only scratch the surface.

🔁

No systematic retargeting insight

Without automation, you can't consistently identify at-risk buyers before they churn.

🌐

Multi-account chaos

Managing 5+ client AMC instances manually? Each one needs separate logins, separate queries.

How the Pipeline Works

One script. Runs automatically. Pulls live data from every connected AMC instance.

Step 1
Authenticate
Refresh token → access token (auto-renewed)
Step 2
Submit SQL
POST query to AMC API with dynamic date range
Step 3
Poll Status
Every 15s until SUCCEEDED
Step 4
Download
Pre-signed S3 URL → CSV (no bucket needed)
Step 5
AI Insights
Claude analyses data, generates narrative

What Happens Under the Hood

POST
api.amazon.com/auth/o2/token
grant_type    = refresh_token
refresh_token = {your_refresh_token}
client_id     = {your_lwa_client_id}
client_secret = {your_lwa_secret}
↳ Returns access_token — valid 60 mins, auto-refreshed
POST
advertising-api-[region].amazon.com /amc/reporting/{instanceId}/workflowExecutions
{
  "workflow": {
    "sqlQuery": "SELECT COUNT(DISTINCT user_id)
                 AS buyers, ..."
  },
  "timeWindowStart": "2025-03-23T00:00:00",
  "timeWindowEnd":   "2026-03-23T23:59:59",
  "timeWindowType":  "EXPLICIT"
}
↳ Dates always dynamic — no hardcoded ranges
GET
/workflowExecutions/{executionId}
Poll every 15 seconds

PENDING → RUNNING → SUCCEEDED
                  → REJECTED  (SQL error — check statusReason)
                  → FAILED
↳ Typical query time: 2–4 minutes
GET
/workflowExecutions/{executionId}/downloadUrls
{
  "downloadUrls": [
    "https://s3.amazonaws.com/...presigned..."
  ]
}

→ Download CSV directly
↳ Direct download — no S3 bucket configuration required

Queries We Run Automatically

Query 1

LTV Dashboard

Total buyers, revenue, lifetime value per buyer, and repeat purchase rate — from Amazon's purchase attribution table.

↳ e.g. 420,000 buyers · £7.60 LTV · 7% RPR
Query 2

Lapsed Buyer Segments

Buckets all buyers into Active / At Risk / Lapsed / Churned based on days since last purchase — as of today.

↳ e.g. 96k At Risk + Lapsed buyers identified for retargeting
Query 3

Time to Conversion

Days between first ad exposure and purchase — from Amazon's traffic and conversion timing table.

↳ e.g. 90.2% convert same day → 7-day retargeting window is optimal
Query 4

NTB by ASIN

New-to-brand purchases, NTB%, attributed revenue and ad spend per ASIN per month.

↳ Identify gateway ASINs that drive the most new customers
Query 5

Ad Spend + ROAS

SP and DSP spend by month from Amazon's ad spend tables, with total ROAS across both channels.

↳ Full-funnel spend visibility in one number
Query 6 — DSP Only DSP Required

Viewed But Didn't Convert

DSP is the only Amazon ad type that tracks impressions at user level — meaning AMC can tell you exactly who saw your ads. We join DSP exposure data against conversion events to isolate high-intent non-buyers.

↳ e.g. "68,000 users saw your DSP ads but didn't buy — push them directly into a retargeting segment now"

Example Output — What the Pipeline Produces

Trailing 12 months · Illustrative figures based on typical Amazon brand accounts · Data structure is exactly what you receive for your clients

LTV Summary
MetricBrand ABrand B
Total Buyers420,00018,000
Total Sales£3,200,000$210,000
LTV / Buyer£7.60$11.70
Repeat Buyers29,400360
Repeat Purchase Rate7.0%2.0%
Same-day converters90%90%
Lapsed Buyer Segments — Brand A
SegmentBuyersRevenue
Active 0-30d28,000£240,000
At Risk 31-60d52,000£400,000
Lapsed 61-90d44,000£340,000
Churned 90d+296,000£2,220,000

What the Data Tells You

~70% of buyers have churned

Roughly 296,000 buyers haven't purchased in 90+ days. That's over £2.2M in revenue from people who already trust the brand — the most efficient retargeting audience you have.

96k buyers about to churn

The At Risk + Lapsed segments represent around 96,000 buyers who purchased in the last 30–90 days. A well-timed DSP retargeting push now costs a fraction of acquiring new customers.

90%+ convert same day

Both brands show ~90% same-day conversion from ad exposure. This means 30-day retargeting windows are wasteful — 7 days is optimal, saving budget on non-converting impressions.

Brand B: 2% RPR — biggest opportunity

Nearly all buyers are one-time purchasers. A structured retargeting programme could realistically move this to 5–7%, directly impacting LTV without increasing ad spend on new customer acquisition.

Running DSP? We can find your non-converters

SP impressions don't carry user IDs — only DSP does. If you're running DSP, AMC can identify every user who saw your display ads and didn't purchase. That audience is already warm. Retargeting them costs a fraction of prospecting cold traffic.

Want This for Your Agency?

I set up the full pipeline for your AMC instances — automated queries, lapsed buyer segments, AI-generated insights, and retargeting audience builds. Ready in days, not months.

Book a Free Discovery Call →

Or email shuqingke1404@gmail.com

About

I'm Shuqing Ke, a UK-based independent data consultant specialising in Amazon Marketing Cloud (AMC) analytics. I build and operate automated reporting pipelines for marketing agencies and DTC brands selling on Amazon — turning raw AMC data into weekly decision-ready dashboards with AI-grounded insights.

Current clients include marketing agencies serving DTC brands across Japan, the UK, and EU. Engagements range from one-off template builds (LTV, NTB, retention, churn analysis) to ongoing monthly retainers covering pipeline maintenance, new-seller onboarding, and feature additions.

Get in touch: Book a 15-min call · shuqingke1404@gmail.com · LinkedIn

Built and operated by Shuqing Ke · UK-based independent AMC analytics consultant · shuqingke1404@gmail.com
Data shown is from live client accounts, anonymised. Results will vary by brand, category, and ad investment. · Built with Amazon Ads API + Claude AI