Attribuly Blog
A
Alex Li
28 min read

Stage-Based Frequency Caps for Ecommerce Retargeting

Channel-specific daily/weekly caps, recency tiers, fatigue signals, and alert rules for ecommerce retargeting—practical benchmarks for DTC paid-media and growth teams.

Stage-Based Frequency Caps for Ecommerce Retargeting

If you’re retargeting without stage-based caps, you’re either burning creative faster than you can produce it or leaving profit on the table. The smartest teams anchor caps and fatigue alerts to a business outcome, then adjust by recency tier and channel realities. Below is a practical, research-backed playbook with conservative-to-balanced defaults you can deploy today.

Key takeaways

  • Use blended MER and ROAS as the governing signal; treat channel metrics as diagnostics.

  • Cap by recency tiers to protect margins and creative longevity while preserving incremental reach.

  • Start with balanced weekly caps and rotate creative when CTR drops more than 20 to 30 percent or CPA rises more than 20 percent week over week.

  • Rely on platform guidance where it exists and label practitioner defaults where the public documentation is thin.

  • Automate alerts that combine MER drops with channel fatigue signals and escalate actions in 72 hour windows.

The anchor KPI that governs caps

Blended MER or brand-level ROAS should anchor your frequency policies. It keeps decisions tied to profit instead of silo metrics. Think of channel metrics like CTR, CPM, CPA, and in-platform ROAS as your instrument panel for diagnosing why MER moved, not as the pilot of your caps. When the governing KPI dips materially, your alert logic should check stage-level exposure and per-channel fatigue signals before pulling back budget or tightening caps.

Why not lead with channel metrics alone? Because retargeting often looks efficient inside a walled garden while cannibalizing other channels or organic demand. Anchoring to blended MER and ROAS ensures caps and rotations are aligned to business outcomes across the mix.

Recency tiers and stage mapping

Define clear tiers and keep your exclusions tight so you’re not paying to reach recent converters.

  • Hot 0–3 days: carts and deep product viewers

  • Warm 4–7 days: product viewers and active site browsers

  • Tepid 8–14 days: soft site engagers

  • Cold 15–30 days: broad site engagers and list match

  • Dormant 31–90 days: low intent or lapsed users

Hygiene basics: exclude purchasers for 7 to 30 days depending on replenishment cycles, exclude high LTV subscribers from aggressive caps, and mirror lookback windows to your typical decision cycle. Keep audiences deduped across channels where possible to avoid double counting frequency.

Channel benchmarks for frequency capping ecommerce retargeting

These benchmarks balance control and scale. Where platform guidance exists, we cite it. Where it doesn’t, we label as practitioner defaults and recommend validating with holdouts.

Channel or Inventory

0–3d Hot

4–7d Warm

8–14d Tepid

15–30d Cold

31–90d Dormant

Notes

Meta FB and IG

1 per day or 3–5 per week

1 per day or 2–4 per week

1 every 2 days or 2–3 per week

1–2 per week

1 per week

Practitioner defaults pending newer public Meta docs

TikTok

2–3 per week

2 per week

1–2 per week

1–2 per week

1 per week

Weekly creative refresh and modest weekly exposure per TikTok guidance in the Creative Best Practices library: see the official guidance in the article titled Creative best practices at TikTok Ads Help at https://ads.tiktok.com/help/article/creative-best-practices and in seasonal playbooks

YouTube or Demand Gen

Prefer campaign level caps sparingly at 3–5 per week for warm tiers

Same

Same

Same

Same

Google notes performance optimized Demand Gen already accounts for exposure; campaign caps apply across surfaces. See Google Ads and DV360 support articles for Demand Gen and frequency measurement

Display via DV360

1–2 per day or 5–10 per week

1–2 per day or 5–10 per week

1 per day or 3–7 per week

1–5 per week

1–4 per week

Viewable impressions count toward line item caps per DV360 documentation

CTV household level

2–3 per day

2–3 per day

1–2 per day

1 per day or 3–7 per week

1–3 per week

Use identity to cap at household and individual levels; follow IAB ACIF and identity best practices

Email Klaviyo

Not applicable

Not applicable

About 1 per week

1 per week

1 per week

Use Smart Sending and escalate during promotions with engaged cohorts

SMS Klaviyo

Not applicable

Not applicable

1–2 per week

1–2 per week

1 per week

Certain flows limited to one SMS per recipient and respect quiet hours

Evidence and references you can verify now:

Fatigue signals and creative rotation cadence

How do you know caps are too high for a tier? Watch for fatigue patterns that tend to precede MER decay.

  • CTR down more than 20 to 30 percent versus the prior seven day baseline

  • CPA up more than 20 percent week over week on stable spend

  • ROAS down more than 20 percent on a 14 day basis adjusted for promo noise

  • Frequency hitting upper bounds without incremental reach growth

Rotation triggers: refresh creative when any of the above trip, or when average frequency over the last 14 days exceeds these soft ceilings: Meta 5 to 7, TikTok 3 to 4, YouTube or Display 5 to 7, CTV households 7 to 10. Keep at least three to five unique creative variants live in warm tiers to slow fatigue, and use value led messaging for dormant tiers to avoid unsubscribes and ad avoidance.

Alert rules and automation playbook

Your alerts should tie a business outcome drop to specific exposure signals by tier and channel. Here’s a simple and extensible pattern.

  • Global governing trigger: blended MER over the last 14 days is down more than 20 percent versus the prior 14 days

  • Channel confirmation windows: use seven day frequency and CTR or CPA for near real time detection

  • Stage guardrails: only trigger actions when the affected tier shows elevated frequency or decaying response

A compact rule you can adapt:

{
    "scope": "brand",
    "governing_kpi": {"metric": "blended_mer", "window": "14d", "delta": "-20%"},
    "confirmations": [
      {"channel": "meta", "frequency_7d_gt": 4, "ctr_drop_7d_gt": "20%"},
      {"channel": "tiktok", "frequency_7d_gt": 3, "cpa_increase_7d_gt": "20%"},
      {"channel": "dv360_display", "viewable_freq_7d_gt": 7, "roas_14d_drop_gt": "15%"},
      {"channel": "ctv", "hh_freq_7d_gt": 10, "completion_drop_7d_gt": "10%"}
    ],
    "stage_filters": ["0-3d", "4-7d", "8-14d"],
    "actions": [
      {"type": "notify", "targets": ["slack", "email"]},
      {"type": "tighten_caps", "adjustment": "-25%"},
      {"type": "rotate_creatives", "min_new_variants": 3}
    ]
  }

Practical workflow with Attribuly as a neutral example

  • Use blended MER and ROAS available in Attribuly reporting to watch the 14 day trend at the brand scope. See the page Attribuly Metrics for how ROAS variants are defined at https://support.attribuly.com/en/articles/9004248-metrics

  • Configure an alert that checks when blended MER falls more than 20 percent and the warm tier shows frequency above target in the last seven days for Meta or TikTok. Then route a Slack notification to the channel owner with a short action list tightening caps by 25 percent and rotating in three new creatives

  • If the alert persists for 72 hours, run a stage constrained budget reduction to hot tiers and schedule a creative audit. If it persists for seven days, schedule a geo holdout test to validate incrementality

For teams that prefer product documentation when setting up alerting workflows, review the Attribuly Product page for an overview of capabilities at https://www.attribuly.com/product and the Alerts feature page at https://www.attribuly.com/product/alerts

Implementation notes for measurement and privacy

Cross channel deduplication and clean measurement make these caps meaningful.

  • Server side tagging: use sGTM or CAPI to improve event stability under privacy constraints and to align impression and conversion timestamps across platforms

  • Viewable versus served counting: DV360 counts viewable impressions toward certain caps and reports reach and frequency accordingly, which you can confirm in Measuring reach and frequency at the DV360 Help Center at https://support.google.com/displayvideo/answer/15272050?hl=en

  • Creative identity: adopt the IAB Ad Creative ID Framework to reconcile creative performance and manage rotations consistently across CTV, display, and social. The IAB announcement Improve Connected TV Advertising Experience outlines how this helps cross platform control at https://iabtechlab.com/iab-tech-lab-announces-ad-creative-id-framework-to-improve-connected-tv-advertising-experience-for-consumers-and-brands/

  • Lookbacks and exclusions: align conversion lookbacks with typical decision cycles by product category. Keep conversion exclusions strict so hot tiers aren’t inflating frequency on recent buyers

  • Holdouts and PSA tests: when fatigue persists despite rotations and cap changes, use geo holdouts or PSA suppression to validate incrementality before large budget shifts

Next steps

Start with the balanced defaults above and wire the alert JSON to your analytics stack. If you need a simple way to monitor blended MER and route fatigue alerts across platforms, explore Attribuly Alerts at https://www.attribuly.com/product/alerts