AMC Workflow Playbook · 04

The 4-Month AMC Window

Amazon just made its premium AMC datasets free through 12/31/2026. Was $1-20K/mo per account. Now $0. The 3 datasets that matter, the use cases they unlock, and the order to subscribe.

$1-20K/mo → $0

Per advertiser account

Under 4 months

Through 12/31/2026, then back to paid

3 datasets

FSI, ARP, Brand Store Insights

Why this matters now

Amazon turned its premium measurement layer free, and the window is closing

3 datasets that previously ran $1-20K/mo per advertiser account just went $0 through 12/31/2026. This isn't a discount -- it's a strategic window. Data backfill starts the day you subscribe, doesn't extend, and reverts to paid pricing on 1/1/2027. The agencies that subscribe + build measurement queries before it closes own a CFO-level data layer their competitors will need to catch up to.

The asymmetric opportunity. Most agencies will subscribe their clients, sit on the data, and not know what to do with it -- because the analytical use cases require knowing what to ask of the new tables. Knowing which 3 questions to point at each dataset (and in what order) is what turns a free subscription into a renewal moat.

The honest urgency. Backfill doesn't retroactively appear. Accounts that subscribed in June already have three months of signal you cannot get back. Every week you wait is another week of history you will never be able to query. Speed-to-subscribe compounds.

The datasets

The 3 that matter (and which to skip)

6 first-party datasets went free. Only 3 deliver immediate leverage for SaaS agencies serving brand clients. The other 3 (Prime Video Insights, Prime Video Channel Insights, Amazon Your Garage) are niche to specific eligibility -- mention them only if the client qualifies.

Dataset 1 · The killer one

Amazon Insights (FSI)Was $1-20K/mo · Now $0

Combined ad-exposed AND un-exposed conversion events for tracked ASINs. Until now, AMC only showed ad-attributed conversions. FSI exposes the missing baseline -- the buyers who converted without ever seeing an ad. That's the data that finally enables true incremental ad lift measurement WITHOUT running a holdout test.

exposure_type field on every conversion: ad-exposed (saw an ad in the 28d before buying) or non-ad-exposed (bought without any ad). The difference between the two cohorts = true ad lift.

Organic vs ad-driven NTB customer sizing

Compare new-to-brand buyers who were ad-exposed vs those who weren't. Reveals your organic NTB acquisition strength per ASIN -- the new customers you'd get without spending. Critical input to budget reallocation: which ASINs have strong organic pull vs which depend on ad spend for new-customer acquisition.

Total customer LTV (not just ad-attributed)

For brands wanting "total customer" or "new customer" numbers that reconcile to Seller Central. Standard AMC undercounts because it only sees ad-touched buyers. FSI gives the full picture -- enables a "true customer" LTV vs the "ad-attributed customer" LTV.

Audience segment over/under-indexing

Combined with Audience Segment Membership (also part of Amazon Insights), shows which Amazon-defined behavioral segments your buyers fall into vs the baseline market. Informs media planning + audience refinement -- where you're winning, where you're under-indexed.

Dataset 2 · The long-LTV one

Amazon Retail Purchases (ARP)Was $1-20K/mo · Now $0

5 YEARS (60 months) of ASIN-level retail purchase data across Amazon Stores -- regardless of advertising. ARP closely matches Vendor Central's "Retail Analytics > Distributor View" or Seller Central's "Business Reports > By ASIN Detail" -- the dataset that reconciles AMC to the reports brands already trust.

ARP covers all Amazon Store purchase activity, not just ad-attributed. Includes signals standard AMC doesn't expose: is_business_flag, purchase_program_name (Subscribe & Save), purchase_order_method (Cart / Buy Now / 1-Click).

5-year customer LTV trajectory

Standard AMC is limited to ~13 months of ad-attributed data. ARP gives you 60 months of total purchase data per customer. Build LTV cohorts that span 3-5 years -- critical for long-LTV categories: supplements, beauty, CPG, pet, baby.

Subscribe & Save subscriber dynamics

ARP's purchase_program_name field separates Subscribe & Save subscribers from one-time buyers across 5 years. Surfaces S&S retention curves per ASIN, churn timing, and which products convert one-time buyers into subscribers. For high-repeat categories (supplements, pet, baby, beauty consumables), this is the data that drives subscription growth strategy.

Purchase method intent signals

ARP's purchase_order_method field shows whether each purchase was Cart, Buy Now, or 1-Click. Different intent levels per method: 1-Click = high-loyalty repeat buyer, Cart = considered purchase (often multi-item baskets), Buy Now = mid-intent impulse. Combined with ASIN-level data, surfaces which products are loyalty-buyer concentrated vs impulse-buyer dependent.

Dataset 3 · The brand store one

Brand Store Insights (BSI)Was $1-3K/mo · Now $0

Page-level and engagement-level events from Amazon Brand Stores. Tracks visits, dwell time, widget interactions, and -- critically -- the traffic source for each visit. Subscribe only if the client has an active Brand Store; skip otherwise.

ingress_type field shows how each visitor reached the store: 1 = search · 2 = detail page byline · 4 = ads · 6 = store recommendations · 7-11 = experimentation. Different intent per source.

Brand store funnel attribution

Joined with conversion events, shows whether store visits actually drive purchases -- broken down by traffic source. Answers the long-standing brand store ROI question: "is our brand store a real conversion asset, or just a vanity page?"

Engagement quality scoring

Dwell time + widget interactions per page = quality of visit. Surfaces which brand store sections (hero, A+ modules, video, comparison) drive engaged behavior vs bounces. Direct input to A/B testing the store's structure.

Ad-driven vs organic brand store traffic

Isolate ad-driven visits (ingress_type = 4) from organic. Compare conversion rates. Answers: "do shoppers who arrive at our brand store via ads convert at the same rate as those who arrive via search?" -- informs whether brand store optimization should be uniform or segmented by traffic source.

The order

Subscribe in this sequence, not all at once

Subscribing is one click each, but the analytical work that follows is finite. Sequence the value: highest-leverage first, narrowest scope last.

Subscribe to FSI first (every eligible client)

This unlocks the CFO-level conversation. True incrementality is the answer to the "is this ad spend working?" question every agency faces at renewal time. Subscribe today, run the first incrementality split query within a week, present findings to the client team within 30 days.

Subscribe to ARP second (every eligible client)

The 5-year purchase data is irreplaceable for long-LTV categories. Eligibility requires Vendor Central or Seller Central to be associated with the AMC instance -- verify before subscribing. Start with simple per-ASIN LTV trajectories before attempting complex joins (mind the time-window mismatch with other AMC tables).

Subscribe to BSI third (only if eligible)

Only if your client has an active Brand Store. Useful but narrower scope than FSI/ARP. Skip otherwise. For brands with stores, the funnel attribution use case usually pays for the subscription work alone.

The bigger context

Why this matters MORE in the Alexa era

Amazon's AI shopping layer (Alexa, built on 2 years of Rufus learnings) is compressing the buyer journey -- discovery → click → buy → reorder happens in days, not weeks. In that world, attributed ROAS becomes more inflated, because more sales get credited to ads that would've converted organically through the AI-driven funnel.

The conventional reading. Alexa changes how shoppers discover. Brands need to restructure their listings, optimize for AI summaries, and front-load structured attributes. All true at the discovery layer.

The reading nobody's connecting. Alexa accelerates the funnel -- so attribution overcounting compounds faster. The brands that finally prove TRUE incremental ad lift (vs what Alexa would've converted organically) win the CFO budget conversation in 2026 and 2027. The agencies that wire that proof layer before the free window closes own the renewal cycle that follows.

The honest caveat. Alexa internal data is NOT in any of these free datasets. Amazon won't expose Rufus/Alexa query data -- it's their moat. What FSI gives you is the next-best thing: the un-exposed conversion baseline, which makes incrementality measurable regardless of which discovery surface (search, Alexa, brand store) the shopper used.

Operational

5 things to verify before you click subscribe

Subscribing is one click, but the prerequisites matter. Get any of these wrong and the data won't flow.

Check	Required for	Why
AMC instance access granted	All 3	You can't subscribe on a client's behalf without instance-level access. Confirm before promising a timeline.
Vendor / Seller Central associated	ARP	Without the selling-account association, ARP returns no data. Verify in: Ads Console > Administration > Account access & settings > Selling account.
Active Brand Store exists	BSI	If no Brand Store, skip BSI -- the tables will be empty.
"Free trial complete" status	FSI	Accounts that ran a prior FSI trial may have different re-subscription eligibility under the new free pricing. Check before assuming.
Data backfill window	All 3	Typically 24-72h after subscribing before data is queryable. Set client expectations accordingly.

3rd-party signals still cost. Experian Vehicle Insights and other 3P paid features are NOT included in the free 1P tier. Don't promise them as part of the free window. NCS CPG Insights Stream has been discontinued by Amazon altogether and is no longer available at any price.

Watch out

5 mistakes that waste the window

The free subscription is easy. Using it well is the actual work. Most agencies will trip on one of these.

1. Subscribe and forget to query. The most common failure. the window passes, data accumulates, you build nothing, and on 12/31 you have a free trial expiring with no measurement output to show. Set a calendar reminder for "first query against new dataset" 1 week after every subscription.

2. Joining ARP with other AMC tables without time-window filters. ARP has a 60-month window; standard AMC tables have ~13 months. Joining them without explicit date filters produces analytical errors. Amazon explicitly warns about this. Always hardcode date filters when joining ARP.

3. Running production queries before sandbox testing. The new tables have new schemas, new constraints, new aggregation thresholds. Test every query pattern in sandbox before promoting to production -- otherwise you'll burn query credits on failed runs.

4. Ignoring Audience Segment query constraints. The audience_segments_* tables time out on extended date ranges and can't be queried across multiple regions in one SQL statement. Filter aggressively by segment_marketplace_id before joining.

5. Not telling your clients. If your client finds out about the free window from their next agency RFP instead of from you, you've quietly leaked credibility. Day-of-subscription Slack/email: "we just enabled these datasets for you -- here's what we'll be measuring."

The plan

What to actually do in the time left

A working sequence -- not theoretical. Each phase is one calendar block. By the time the window closes on 12/31/2026, you have a measurement infrastructure your competitors are starting from zero on.

Phase 1

Week 1 · This week

Subscribe all eligible client accounts to FSI + ARP + BSI (where applicable). Verify eligibility checks. Confirm backfill timeline with each client.

Phase 2

Weeks 2-3 · Rest of September

Build sandbox queries for the top use case per dataset per client. Validate against expected ranges. Document any schema gotchas per client account.

Phase 3

October 2026

Run incrementality + LTV analyses on each client's highest-spend brand. Present findings to client teams. Identify the 1-2 use cases per client worth productizing.

Phase 4

November-December 2026

Build production query pipelines. Accumulate historical data. Integrate findings into existing reporting cadence. Validate against Seller Central / Vendor Central for credibility.

Phase 5 · After the window

January 2027 onwards

You have months of backfill + working pipelines + presented findings + earned client trust. Competitors who didn't subscribe are starting from zero -- with the data layer now back to $1-20K/mo per account. This is the renewal moat.

Want it built for your accounts?

I wire the FSI + ARP measurement layer for Amazon agencies and SaaS

If you'd rather skip the build and have the FSI + ARP measurement layer deployed against your AMC instances directly -- queries written, methodology designed, and findings packaged into client-ready reports -- I do this as a fixed-fee engagement. 15-min discovery call to scope which datasets to start with for your portfolio.

Book a 15-min call →