How to replace a dying ad using competitor research (under 20 minutes)

Your CTR just dropped below 2%, frequency is climbing, and ROAS has been sliding for three days straight. Now what?

Most media buyers have a detection checklist. Almost none have a replacement pipeline.

This article is the replacement pipeline. A 20-minute workflow that takes you from "this ad is dying" to "I have three new concepts worth testing," without guessing and without briefing a creative team off thin air.

Your fatigue dashboard is working, your replacement process isn't

Every media buyer reading this already knows the fatigue signals: CTR drops under 2%, frequency hits 3-4, ROAS declines 3-5 days in a row, cost per result creeps up. You catch it and pause the ad set.

Then you sit there.

"The ad is dying" isn't the hard part. The hard part is what you put in its place, and most people default to two options that both lose money.

Option 1: rerun the old playbook. You tweak the hook, swap the opening clip, and relaunch, but it buys you maybe a week before the same fatigue comes back faster.

Option 2: greenfield brainstorm. You write a brief from scratch, send it to the editor, wait four days, and spend $800 testing, only to find the new creative doesn't clear the old one's floor half the time.

Both options are expensive because both treat replacement as a creative problem. It's not. It's an intelligence problem.

Detection tells you what died, not what's alive

Here's the gap in every fatigue article published in 2026: they all teach you how to spot a dying ad. None tell you where the replacement comes from.

You don't need to invent a new winning creative. Someone in your niche already has one running right now.

Your competitors launched new ads this week, and some have been running for 25+ days with heavy spend, which is the only signal that means something in the ad library. If an ad is still live at 30 days, it's making money because nobody keeps losing ads up at that spend level.

That's your replacement pool. Not your imagination.

The workflow below uses three things: your own fatigue dashboard as the trigger, Brandsearch Discovery as the inspiration layer, and Brandsearch Spectre as the competitor-specific intel feed. Twenty minutes end to end.

The 20-minute replacement workflow

Here's the whole thing, broken into four phases.

Phase 1: confirm fatigue in your own numbers (2 min)

Before you go anywhere else, confirm the ad is actually dying, not having a bad day. Pull the last 7 days of CTR, CPC, frequency, and ROAS and compare to the trailing 30.

If CTR has dropped below 2% for three consecutive days and frequency is above 3, it's fatigued. If it's one bad day, wait.

This is the only phase that happens in your own dashboard. The other three are competitor-driven.

Phase 2: pull new winners in your niche (5 min)

Open Discovery. Filter to your niche, set format to video, running days to 25+, and phase to winning. Sort by total reach.

You're looking at every video ad in your category that's been live long enough to prove it makes money, not new tests but actual winners that survived past the breakeven window.

I typically land on 12 to 20 ads in any busy niche. Scroll it once and save the three or four that open with a completely different angle than your dying ad: different pain, different promise, different visual setup.

The goal is pattern recognition, not plagiarism. You want to see what angle is pulling right now so you don't brief a new creative on a theme that already fatigued across the whole niche three weeks ago.

Phase 3: check what your tracked competitors just launched (8 min)

Now switch to Spectre.

If you're tracking the five or ten competitors that matter in your niche, Spectre shows you every new ad they pushed in the last week. Not the archive but the deltas: new launches, new landing pages, new hooks.

Open each tracked competitor and look at two things. What did they launch in the last 7 days, and which of those new ads are already showing heavy reach.

The ads still live at 7 days with climbing reach are the early winners. Discovery shows you what's already won, and Spectre shows you what's about to win.

If a direct competitor just launched three new video ads in the last four days and one is already pulling serious reach, that's not a coincidence. Someone in your exact market already validated a concept, on your exact buyer, last week.

Phase 4: mine hooks from Brand Analysis (5 min)

Pick your closest tracked competitor and open their Brand Analysis page. Traffic trends, ad scaling, creative tests, and landing page history are all in one screen.

Open the Hooks tab. It lists the opening lines of every video ad the competitor is running, transcribed, searchable, grouped by performance.

Hooks fatigue faster than anything else in a video ad, because the first three seconds are the first thing viewers get tired of and the first thing the algorithm starts underpricing. If you can replace a tired hook with a proven one, you buy another 2-3 weeks of life out of the same body and CTA.

Pull the top 10 hooks from the Hooks tab and look at which ones are running on the ads with the longest run times. Those are the hooks that kept working past the normal fatigue curve.

Write down the pattern, not the exact words but the structure. Pain-first? Question? Direct offer? Demo? Social proof?

That structure becomes your brief.

What a 20-minute brief actually looks like

After running this workflow on a live campaign last quarter, here's what the output looked like for a supplement brand whose flagship video ad had been running 42 days and finally tipped into fatigue.

Fatigue signal: CTR 1.6% (down from 3.1%), frequency 3.8, ROAS 1.9 (down from 3.4) across three days.

Discovery pull: 14 winning video ads in the supplement niche running 25+ days. Of those, six opened with a pain-point hook, four with a doctor-authority hook, two with a before/after.

Spectre pull: Two direct competitors launched new hooks in the last 7 days. One was a 3-second "You're taking the wrong dose" opener that was already at 600K reach in four days.

Hooks tab: The three top-performing long-runners in the niche all started with a contrarian claim in the first 2 seconds. None led with product.

Brief: "Contrarian pain hook + doctor cutaway + existing testimonial body + existing offer CTA." Edit time: 90 minutes from a single existing asset. Test launch the next morning.

New creative cleared the old ad's floor on day two and was the brand's top spender within a week.

The brief wasn't guessed. It was assembled from four separate competitive signals, each pointing at the same angle.

The replacement pipeline is a system, not a sprint

If you're in a fatigue cycle right now, the instinct is to panic and greenfield a new creative. Resist it.

A dying ad is a scheduled event: you know it's coming 3-5 days before the full ROAS collapse. That gives you enough time to run the competitor loop calmly.

The whole point of using competitor research as the replacement pipeline is that you're never starting from zero. The market is already telling you which angles work, and your job is to listen faster than your competitors do.

The 20-minute replacement loop

When your dashboard flags fatigue, don't brainstorm. Run the loop.

1. Confirm fatigue in your own dashboard: CTR below 2%, frequency above 3, ROAS declining 3+ days.
2. Pull new winners from Brandsearch Discovery: filter your niche to video ads, running 25+ days, winning phase. That's your shortlist of ads that survived the breakeven window.
3. Check tracked competitor launches in Brandsearch Spectre: every new ad pushed in the last 7 days, filtered to ones with climbing reach.
4. Mine opening hooks from the Brandsearch Brand Analysis Hooks tab: pull the top 10 hooks on the longest-running creatives. Copy the structure, not the words.
5. Brief the new creative from the combined pattern: not from your imagination. Launch within 24 hours.

Detection without replacement is an incomplete system. Competitor research is the replacement layer: twenty minutes, four phases, and a brief that was pre-validated by the market before you ever wrote it.