Amazon gave you the pipeline.
We keep it running for every advertiser.
Amazon's open-source reference application turns Stream deliveries into Iceberg tables. AdsLake deploys it in your AWS and handles what it leaves to you: finding every advertiser account, keeping subscriptions and tokens alive, watching delivery, and upgrading as Amazon changes.
Runs in your AWS · Your credentials stay with you · Open Iceberg tables
Stream fixes report polling. Running it for many advertisers is the hard part.
Marketing Stream pushes Amazon Ads data to you every hour, and Amazon's reference application gives you a solid way to merge it. It is a reference, not an operated product. Keeping it right across dozens of accounts is still your job.
Request, wait, retry, repeat
- Reports stuck in PENDING for minutes or hours, for every advertiser, every day.
- HTTP 429 throttling as soon as you add accounts or backfill.
- Limited lookback: miss a window and the data is gone.
- Numbers that shift later: attribution restates them, so yesterday's report is already stale.
Great foundation, ongoing upkeep
- Every advertiser subscribed by hand, dataset by dataset, with an access token that expires in an hour.
- New accounts don't subscribe themselves, and lost access goes unnoticed until a dashboard is empty.
- One delivery mixes advertisers, so sharing data with clients needs per-account scoping.
- Reference code may change without backwards compatibility. Someone has to own upgrades and new datasets.
From zero to every advertiser streaming
We deploy the pipeline in your AWS, then AdsLake keeps it current.
Deploy in your AWS
We set up the Stream destination and Amazon's merge pipeline in your account, sized for your volume and checked against numbers you already trust.
Connect Amazon Ads
Add your Amazon Ads API app. The refresh token is stored in your Secrets Manager and never sent to us. AdsLake finds every advertiser account it can reach.
Set a rule, not a to-do list
Choose all accounts, a filter or a hand-picked list. AdsLake creates the subscriptions, adds new accounts as they appear and alerts you when delivery stops.
Everything around the pipeline
The upkeep every team ends up building, done once and kept up to date.
Account discovery
Finds every advertiser account your API app can reach, including Sponsored Ads and DSP, and re-checks daily.
Self-healing subscriptions
Desired state in, actual state out. New accounts are added by rule, failed subscriptions are recreated, revoked access is flagged.
Token management
Access tokens are refreshed automatically from your stored refresh token. No more pasting one-hour tokens into a terminal.
Freshness per advertiser
See the latest data hour for each account, backfill progress and failed deliveries. Get alerted before dashboards go stale.
Upgrades & new datasets
When Amazon adds datasets, metrics or changes the reference code, we test the change and roll it out to your install.
Reconciliation soon
Compare Stream totals with the Reporting API per advertiser and day. Make the switch with proof that numbers match.
One lake for all your advertising data
Amazon Marketing Stream is where we start. The same approach (your cloud, open tables, upkeep handled) extends to the other sources media teams join it with.
Amazon Marketing Stream
Hourly Sponsored Ads and DSP performance and conversions, merged and kept current.
Amazon Ads reports & metadata
Reporting API for what Stream doesn't cover, plus campaigns, ad groups, targets and account details through the API.
Amazon Marketing Cloud
Scheduled AMC queries, including reach and frequency, landed as tables next to your Stream data.
Google Ads & Display & Video 360
Performance and entity data from Google's ad platforms, in the same tables and naming.
Google Ads Data Hub
Clean-room query results, reach and frequency for Google and YouTube campaigns.
Meta
Facebook and Instagram campaign performance, reach and frequency through the Marketing API.
Reach and frequency are reported per platform. Clean rooms don't share user-level data, so we don't claim de-duplicated reach across platforms. Tell us which source you need first.
Start with a deployment, keep it running with AdsLake
Pick what you need now. Everything runs in your AWS except Hosted.
Production setup
For teams moving off the Reporting API or Stream v1.
- Stream destination and merge pipeline in your AWS
- Subscriptions for your advertisers and datasets
- Numbers checked against your current reports
- Handover, runbook and Q&A
AdsLake in your AWS
For teams with their own Amazon Ads API app.
- Account discovery and subscription upkeep
- Token refresh, freshness and alerts
- Tested upgrades and new datasets
- Direct engineering support
AdsLake Hosted
For brands and agencies without their own API app or AWS team.
- Click “Connect Amazon Ads”, done
- We run the pipeline; you query the tables
- Optional export to your own bucket
- Monthly, self-serve
Your account. Your keys. Your data.
AdsLake is built to pass a security review on the first try.
- Runs entirely in your AWSPipeline, tables and credentials live in your account. Ad data never reaches our servers.
- Only counts leave your accountThe license check sends the number of active advertisers and subscriptions. No metrics, no campaign IDs, no report data.
- Encryption with your KMS keyDelivered data and tables are encrypted with a key you own and control.
- No hostage dataIf you stop using AdsLake, the pipeline and every table stay in your account and keep working. Only the upkeep stops.
Fixed-price setup, then per active advertiser
Setup is scoped and priced after a short call. AdsLake is billed per advertiser you keep streaming, not per row or gigabyte. Your AWS costs stay on your own bill.
We are working with a small group of design partners now, at early-access pricing in exchange for honest feedback.
- All Stream datasets per advertiser
- Hourly freshness and history backfill
- Upgrades as Amazon evolves
- Direct engineering support
Questions teams ask
How is this different from Amazon's reference application?
The reference application is the pipeline: it turns deliveries into correct tables, and we build on it. It leaves account discovery, subscriptions across many advertisers, token refresh, monitoring and upgrades to you. That is what AdsLake does.
Can we just deploy the reference application ourselves?
Yes, and many teams should. Talk to us if you want it in production quickly, if you manage many advertisers, or if you don't want to own the upkeep.
Do I need my own Amazon Ads API access?
For anything running in your AWS, yes. It uses your Amazon Ads API app, so the data relationship stays between you and Amazon. If you don't have API access, join the waitlist for AdsLake Hosted.
Is any of my advertising data sent to you?
No. Data flows from Amazon into your S3 bucket and is processed in your account. The only thing AdsLake reports to us is license and usage counts, such as the number of active advertisers.
We already use Stream v1. Can you help?
Yes. Stream v2 delivers data differently from v1, so existing pipelines need to be rebuilt. We can run v2 alongside your current setup while you compare numbers.
What happens if we stop using AdsLake?
The pipeline and your tables stay in your account and keep working. Only automatic subscription upkeep, monitoring and upgrades stop. There is no export step and no lock-in.
Get every advertiser streaming, and keep it that way
Tell us how you pull Amazon Ads data today and how many advertisers you manage. An engineer will reply, not a sales team.
Email hello@adslake.io