Skip to content

getaero-io/gtm-eng-skills / skills/deepline-ads-audiences

deepline-ads-audiences

Verified

Use this skill when building, enriching, auditing, or uploading B2B paid ads audiences to Google Customer Match, Meta/Facebook Custom Audiences, or LinkedIn Matched Audiences. Triggers on phrases like '/deepline-ads-audience', '/deepline-ads-audiences', 'upload this audience', 'create custom audiences', 'personal email hashes', 'increase Facebook match rate', 'Google ads audience', 'Meta audience', 'FB audience', or any workflow that turns CRM/customer/contact data into paid ads upload lists. Skip for outbound prospecting sequences, cold email, or pure campaign copywriting.

recipes/sample-abm-segment-example.md

View on GitHub
# Sample ABM Segment Ads Audience Example

Use this recipe for a concrete ABM audience pattern from a high-priority title or account segment. It is intentionally written as a reusable example, not as a customer-specific runbook.

## Scenario

The source audience is a high-priority B2B segment from CRM or Salesforce title segmentation:

- about 2,000 source people.
- most rows have LinkedIn URLs.
- about half of rows have uploadable work emails.
- US-only by default for personal identifier enrichment.
- Mobile phones skipped because they are too expensive for this audience test.

Observed provider planning numbers from the example run:

| Layer                               | Example result                          |
| ----------------------------------- | --------------------------------------- |
| LimaData hashed personal-email hits | low double-digit row lift               |
| Aviato hashed personal-email hits   | low-to-mid double-digit lift            |
| Raw personal-email fallback         | run only after explicit budget approval |

Do not treat these as Deepline benchmarks. They are example-run planning numbers that help agents preserve the intended shape of the workflow.

## Goal

Create separate enriched and unenriched hash-only audiences for Google Customer Match and Meta Custom Audiences so the user can compare actual platform match behavior.

The output should include:

- `unenriched_hash_only.csv`, built from first-party work emails and allowed source identifiers.
- `enriched_hash_only.csv`, built from source identifiers plus provider hashes and raw personal-email fallbacks hashed once.
- `provider_coverage.csv`, showing attempted rows, hits, unique hashes added, failures, and cost by provider layer.
- `no_double_hash_audit.md`, proving provider hashes were not hashed again.
- `upload_report.md`, listing platform account name and ID, audience name and ID, uploaded count, invalid count, request IDs, and readback status.

## Rights Gate

Before enrichment or upload, confirm:

- The CRM or Salesforce segment can be used for paid ads audience creation.
- Enriched personal identifiers can be used for paid ads matching.
- The audience is US-only, or the user has confirmed broader geography rights.
- The workflow is for ABM paid ads only, not outbound.
- Google and Meta account selection has been confirmed by account name and ID.

## Play Path

Create the working directory from the project root, then run the play commands from this skill's folder so `./plays/...` resolves. If `deepline run --help` reports an unknown command, ask the user to update the Deepline CLI (https://code.deepline.com/INSTALL.md) before continuing.

```bash
WORKDIR="$PWD/deepline/data/<segment>-ads-audience" && mkdir -p "$WORKDIR" && echo "$WORKDIR"
```

Build baseline and enriched objects. The build returns two datasets, `baseline_rows` and `enriched_rows`; write each to its own CSV in the same run:

```bash
deepline plays check ./plays/build-hash-only-audience.play.ts
deepline run ./plays/build-hash-only-audience.play.ts --input '{"file":"'"$WORKDIR"'/source.csv"}' --csv-out baseline_rows="$WORKDIR/unenriched_hash_only.csv" --csv-out enriched_rows="$WORKDIR/enriched_hash_only.csv"
```

Audit the enriched payload against the provider hash columns in the source CSV:

```bash
deepline plays check ./plays/audit-no-double-hash.play.ts
deepline run ./plays/audit-no-double-hash.play.ts --input '{"payloadFile":"'"$WORKDIR"'/enriched_hash_only.csv","providerHashFile":"'"$WORKDIR"'/source.csv","providerHashColumns":["aviato_hash","limadata_hash","email_sha256","hashed_personal_email_sha256"]}'
```

Upload enriched and unenriched as separate objects. Use the Google-only play for the Google comparison and the combined play when a Meta Custom Audience ID is available:

```bash
deepline plays check ./plays/upload-google-hash-only-audience.play.ts
deepline run ./plays/upload-google-hash-only-audience.play.ts --input '{"file":"'"$WORKDIR"'/unenriched_hash_only.csv","account_id":"<google_customer_id_without_dashes>","audience_name":"High-priority segment unenriched hash-only <date>"}'
deepline run ./plays/upload-google-hash-only-audience.play.ts --input '{"file":"'"$WORKDIR"'/enriched_hash_only.csv","account_id":"<google_customer_id_without_dashes>","audience_name":"High-priority segment enriched hash-only <date>"}'
```

## Acceptance Criteria

The run is complete only when:

- The play run IDs and output files are reported.
- The source count, LinkedIn URL count, baseline work-email count, and provider coverage counts are reported.
- Provider hashes are passed through as lowercase 64-character SHA-256 values.
- Raw personal emails, if any, are normalized and hashed exactly once.
- The no-double-hash audit passes before live upload.
- Enriched and unenriched objects are uploaded separately.
- Google and Meta account names and IDs are shown in the final report.
- Match-rate fields are reported only when the platforms return them.
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall