Skip to content

Lookalike Audience Research & Build

YB
4.8(11 reviews)25 orders
3

About this service

A lookalike audience is only as good as its seed list. A 1,000-row seed of high-LTV purchasers will outperform a 100,000-row seed of "everyone who ever visited". This engagement exists for operators whose lookalikes have degraded since 2023 (most have) or who never built proper ones to begin with. You keep your ad account, your billing, and ownership of every audience. I research, score, and build the seed lists and the lookalike stack - one-time deliverable, no retainer.

Seed-quality scoring rubric

Every seed list is scored on five dimensions before it becomes a lookalike. The score drives the size recommendation (1% / 3% / 5% / 10%) and the audience-source priority.

DimensionScore 1 (weak)Score 5 (strong)
Recency of seed event> 180 days< 30 days
Event value densityPageView onlyPurchase / lead complete
Seed size< 1,000 rows1,000-50,000 rows
Seed deduplicationDuplicates, rawHashed, deduped
Source granularityOne mega-listSegmented by LTV / SKU / stage

Disclaimer: a perfect 5/5 across all dimensions is rare. Most accounts arrive with 2/5 average. The engagement lifts each dimension to 4/5 or higher where the data allows; dimensions limited by your account's actual history are documented, not over-promised.

What you get - deliverables

  • Seed-source audit: a written inventory of every potential seed source in your account (purchasers, leads, video viewers, engagement audiences, customer list), scored on the rubric above
  • Seed-list design: which sources to use, which to skip, and why - documented per source
  • Customer-list preparation: hashed (SHA-256), deduplicated, formatted to Meta's spec - you retain the raw list, I never store it
  • Segmentation strategy: seed lists split by LTV tier, SKU category, or funnel stage where the data supports it
  • Lookalike stack build: 1%, 3%, 5%, and 10% (or a custom percentage per source) per priority seed - the typical stack is 4-8 audiences
  • Audience naming convention: documented, inheritable, so your team can identify each lookalike by source and tier
  • Exclusion architecture: lookalikes layered with proper exclusion (e.g., 10% excludes the 5%, the 5% excludes the 3%) to prevent overlap
  • Source-priority recommendation: which lookalikes to test first, with the rationale per audience
  • A 30-day monitoring note: which lookalikes are converting, which are not, and the deprecation candidates
  • A one-page audience hygiene guide for your team

How it works - five steps

  1. Day 1: Access and audit. You grant partner-level access to your ad account, share your customer list (hashed if possible, raw if not - I hash it client-side and never store the raw), and confirm your funnel's primary conversion event.
  2. Days 1-2: Seed-source audit. Every potential seed source scored on the rubric. You receive a one-page written audit before any audience is built.
  3. Days 2-3: Seed-list design. Sources selected, segmented, and prepared (hashed, deduped, formatted). Customer list prepared client-side - the raw list never leaves your environment.
  4. Days 3-4: Lookalike stack build. 1% / 3% / 5% / 10% per priority seed, with documented naming conventions and exclusion architecture.
  5. Day 5 + 30-day monitoring. Audience stack delivered with a source-priority recommendation. A 30-day monitoring note tracks which lookalikes convert and which to deprecate.

Who this fits - and who it does not

A strong fit if:

  • You spend $1,500+/month on Meta Ads and have an existing purchaser or lead audience (≥ 1,000 rows)
  • Your lookalikes are 12+ months old and have not been refreshed
  • You have a customer list (CSV or from your CRM/Shopify export) that has never been used as a seed
  • You want a one-time deliverable, not an ongoing retainer

Not a fit if:

  • Your account has fewer than 1,000 qualifying seed rows - the lookalike quality ceiling is too low
  • You are looking for ongoing monthly management - that is a different engagement
  • You want me to fund spend on my accounts - different service model
  • Your seed is "all site visitors, all time" - I will tell you to segment it first

You bring / I run

You bring:

  • Active Meta ad account, with billing you control
  • Full ad spend budget, paid by you to Meta - I never bill media
  • Your customer list (CSV export from Shopify, WooCommerce, CRM, or your CMS), or the data sources already in your account (Pixel, custom audiences)
  • A 48-hour response habit on questions during the 5-day build

I run:

  • The seed-source audit, the seed-list design, the customer-list hashing (client-side, raw never stored), the lookalike stack build, the exclusion architecture, the naming conventions, and the 30-day monitoring note

Reporting and communication

A short written seed-source audit at day 2 (the scores per dimension, in writing). A daily progress note during the 5-day build. A final handover document on day 5 with the audience stack, naming conventions, exclusion architecture, and source-priority recommendation. Day 30 brings a stability note covering conversion performance per lookalike and deprecation candidates. Communication runs through the platform's message system; I respond within one business day.

Why this engagement

Three differentiators. First, the seed-quality rubric is documented before any audience is built - you see the score of every potential source, including the ones I will skip. Second, the customer list is hashed client-side and the raw list never leaves your environment - privacy by design, not by promise. Third, the engagement ends with a 30-day monitoring note - no monthly retainer required, no baked-in dependency. A clean lookalike stack should leave you free, not tied.

What is included

  • Base Facebook Ads campaign · Your ad account, your spend
  • Revisions · 2 rounds included
  • Delivery · 5 to 7 days

Frequently asked questions

Industries served

  • Consumer Electronics
YB
@ybrownUnited KingdomMember since Feb 2026
Online

Search and social, mostly Meta and Google

Completed orders
114
1 in the last 30 days
Rating
4.641 reviews
Active listings
6
Ad platforms
6

More from Yu Brown

View storefront