Skip to main content

My Ad Is Above or Below Benchmarks. What Can I Do?

Kyrios evaluates your ad performance against industry benchmarks after launch; actions depend on whether your ad is above or below benchmarks and whether it is still in the learning phase.

Updated over 3 months ago

Who This Is For / When to Use

  • Users running ads through Kyrios AI Ads Manager

  • Advertisers reviewing benchmark performance scores

  • Anyone unsure whether to edit, pause, or relaunch an ad

Use this article when you see a message indicating your ad is above or below benchmarks.

How Benchmark Scoring Works

Kyrios compares your ad’s performance to:

  • Industry benchmarks for your platform and objective

  • Kyrios internal performance data

Benchmark scores are shown:

  • On or before Day 7 after launch

  • Again on or before Day 14

Benchmark scores do not retroactively change.

Important: The Ad Learning Phase

All major ad platforms (Facebook, Google, TikTok, and others) use a learning phase.

What Happens During the Learning Phase

  • Platforms collect data from ~50 optimization events

  • The system learns who to show your ad to and when

  • Performance may fluctuate

What You Should NOT Do

  • Do not edit ads during the first 7 days

  • Do not frequently change targeting or creative

  • Do not judge results too early

Edits during this phase can reset learning and delay optimization.

If Your Ad Is Above Benchmarks

If your ad is marked Above Benchmarks:

  • No action is required

  • Let the algorithms continue optimizing

  • Avoid unnecessary edits

Kyrios will continue automated testing and optimization in the background.

If Your Ad Is Below Benchmarks

If your ad is marked Below Benchmarks, Kyrios is already:

  • A/B testing creative

  • Optimizing targeting

  • Adjusting delivery automatically

If performance remains below benchmarks, you can take the following actions.

Option 1: Leave the Ad Running

This is often the best first step.

Why:

  • The ad may still need more optimization events

  • Your business may be seeing value from metrics Kyrios cannot track (calls, walk-ins, offline sales)

  • Benchmarks are based on averages, not your full business context

Recommendation:
Wait until at least 14 days before making major decisions.

Option 2: Edit Targeting

Use this option if the audience is not engaging.

You can:

  • Narrow targeting

    • Add negative keywords

    • Target specific cities or regions

    • Switch to retargeting

  • Broaden targeting

    • Add more keywords or interests

    • Expand audience size

  • Change audiences entirely

    • Test a different demographic or intent group

Make targeting changes only after the learning phase.

Option 3: Edit Creative

Use this option if the message or visuals are not resonating.

You can:

  • Switch from images to stock videos

  • Add a clear offer (e.g., “20% Off”)

  • Replace or refresh ad copy

  • Remove low-performing headlines, images, or videos

  • Add new creative variations

Use Asset Reporting to identify low-performing elements.

Option 4: Start Fresh

Use this option when learnings are clear.

You can:

  • Create a new ad on the same platform

  • Launch on a different ad platform

  • Apply insights from what did not work

Starting fresh avoids learning-phase penalties from repeated edits.

How Benchmark Timing Works

  • Week 1 score reflects performance from Days 1–7

  • Week 2 score reflects performance from Days 8–14

  • Editing an ad does not change past scores

Example:

  • If you edit an ad on Day 8:

    • Week 1 score stays the same

    • Updated performance appears in Week 2

Common Issues and Fixes

I Don’t See a Benchmark Score Yet

  • Your ad is still within the first 7 days

  • No action is required

My Score Didn’t Improve After Edits

  • Edits reset optimization

  • Wait for the next benchmark window (up to 7 days)

My Ad Is Below Benchmarks but Getting Leads

  • Benchmarks do not capture all business outcomes

  • If ROI is positive, changes may not be necessary

Did this answer your question?