Tutorial

Audiences & Remarketing That Actually Move CPA

Layer in-market, remarketing, and custom audiences to improve efficiency — without choking volume or over-segmenting your campaigns.

Overview

Audiences turn decent keyword coverage into efficient acquisition. The goal isn’t to micro-target everything — it’s to bias delivery toward people most likely to convert while keeping enough scale to feed smart bidding.

Audience segments and dashboards

Audience types that work

  • In-market & custom intent — pre-built Microsoft segments and your own lists based on search behavior/URLs.
  • Remarketing (site audiences) — all visitors, product/category viewers, cart/checkout abandoners.
  • Customer lists — hashed emails for retention/upsell or exclusion of current customers.
  • Similar audiences — look-alike segments once seed lists are big enough.

Layering strategy: Observation vs. Targeting

Start with Observation (bid only): add audiences to campaigns/ad groups to collect data without restricting reach. Once you find clear winners, test Targeting in duplicate structures to isolate and scale (or use experiments).

Rule of thumb: Observation first for 2–4 weeks → add positive modifiers → then consider Targeting in a controlled test.

Remarketing lists that convert

Create a small portfolio of high-signal lists:

  • All visitors 30–90d — broad volume; conservative modifiers.
  • Product/category viewers 14–30d — mid-signal, moderate modifiers.
  • Cart abandoners 7–30d — high-signal; stronger modifiers or dedicated Target campaigns.
  • Past converters — exclude if new-customer CPA is the KPI; include for upsell/cross-sell structures.
Funnel stages and lists

Bid modifiers that make sense

Use ranges — then tune based on data:

  • All visitors: +5% to +15%
  • Product/category viewers: +10% to +25%
  • Cart/checkout abandoners: +20% to +40%
  • Customer list (new-customer focus): −100% (exclude) or small negative modifier

If running Target CPA/ROAS, keep modifiers modest. Overly aggressive modifiers can fight the bid strategy.

RSAs & assets by audience

  • All visitors: standard value props + social proof.
  • Product viewers: emphasize category fit, specs, and fast benefits.
  • Cart abandoners: address risk: returns, support, trials, financing.
  • Past converters: loyalty/upsell CTAs with tailored sitelinks.

Add audience-specific sitelinks/callouts and test pinning lines for “warm” segments.

Creative sequencing

Map ads to the customer journey. Example sequence for SaaS:

  1. Visit 1 (in-market): problem/solution headlines + pain-first copy.
  2. Visit 2 (site visitor): value prop depth + feature sitelinks.
  3. Visit 3 (product viewer): comparison/objection handling.
  4. Visit 4 (abandoner): risk reversal + urgency.

Exclusions & frequency

  • Exclude existing customers from acquisition campaigns to protect CAC.
  • Exclude irrelevant placements if using audience expansion/display.
  • Cap frequency where possible; rotate ad variants to prevent burnout.

Measurement & reporting

  • Segment by audience in reports; track CPA/ROAS deltas vs. non-audience traffic.
  • Build a simple dashboard: conversions, CPA, spend by list (All visitors, Product viewers, Cart, Customers).
  • Watch assist metrics — remarketing often boosts overall conversion rate beyond last-click.

Implementation checklist

  • ✅ Add in-market + remarketing lists in Observation to every search campaign.
  • ✅ Apply initial bid modifiers (see ranges) and review weekly.
  • ✅ Create cart/checkout abandoner lists with shorter membership windows.
  • ✅ Build audience-tailored RSAs and sitelinks.
  • ✅ Test Targeting in duplicates/experiments for high-signal lists.
  • ✅ Exclude customers from acquisition if CAC/LTV targets demand it.

Next steps

Pair your audience strategy with strong product intent via Shopping. Continue with Shopping Feeds: Titles, Images & Rules.