CTR in 2026: Click-Through Rate in the Andromeda Era

CTR — short for click-through rate — is the percent of people who click your ad after seeing it. The metric tells a different story in 2026 than it did three years ago. Meta's Andromeda model now decides who sees your creative before any human-facing number matters, which means CTR has shifted from a delivery lever to a diagnostic. The rate still pays the bills if you read it right. It misleads you if you read it wrong. This guide covers the formula, 2026 platform benchmarks, the decay curve that signals fatigue, and where the click-through rate fits when an AI grades your ads in real time.

TL;DR: CTR = (clicks ÷ impressions) × 100. Healthy 2026 ranges are roughly 1–2% on Meta feed, 0.5–1.2% on Reels, 3–6% on Google Search, 0.4–0.65% on LinkedIn, and 1–3% on TikTok. Andromeda has demoted click-through rate as a delivery signal, so use it as a diagnostic for hook quality, angle fit, and fatigue — not as a primary optimization target.

What CTR is and how the formula actually works

Click-through rate is one of the oldest performance signals in paid media and one of the most misunderstood. The formula has not changed: divide clicks by impressions and multiply by 100.

CTR = (Clicks ÷ Impressions) × 100

If your ad gets 350 link clicks on 25,000 impressions, the rate is 1.4%. Simple math. The interpretation is where it gets messy.

Three things complicate the number. First, an "impression" on Meta means the ad rendered, even if the user scrolled past in 0.4 seconds. Second, "click" can mean a link click, a profile click, a comment expansion, or a swipe — Meta exposes a dozen click flavors and most dashboards average them. Third, the denominator includes serves to cold, warm, and hot audiences mixed together, so a single rate across an ad set is the average of three different conversations.

That is why a 2.1% reading on one campaign can be excellent and the same 2.1% on another can be a disaster. Without segmenting by placement, audience temperature, and click type, the number is decoration. Pair it with CPM and ROAS for a complete read, and use the CTR calculator to model the dollar impact of moving the percentage one way or the other.

Outbound CTR vs all-CTR — the metric Meta hides by default

The single most expensive click-through-rate mistake in 2026 is reading the wrong column. Meta Ads Manager defaults to CTR (all), which counts every click on the ad creative — including reactions, comments, profile taps, and "see more" expansions. Outbound rate counts only clicks that send a user off-platform to your destination.

The gap between these two numbers is enormous. We see ads in adlibrary's enriched library where CTR (all) reads 3.8% but outbound clicks land at 0.7%. That is a creative people enjoy and never click. Meta's delivery system rewards engagement, so a high all-CTR can still keep an ad in rotation while it bleeds spend with no traffic.

Make outbound the default column in Ads Manager. Do this once, and a third of the "high engagement but no conversions" mysteries solve themselves. Reference Meta's own click metrics documentation and the Marketing API insights reference before you trust any dashboard's column label.

2026 click-through rate benchmarks by platform and placement

Benchmarks are noisy by definition. They drift by industry, season, and account history. The ranges below reflect what we observe across in-market ads on adlibrary in early 2026, cross-checked against WordStream's annual benchmark report, Databox's paid-social benchmarks, and Meta's advertiser performance disclosures.

A few rules pulled from creative testing work across the table. Cold audiences pull the rate down; warm retargeting pulls it up. Reels and Stories will always read lower than feed because the scroll velocity is higher. LinkedIn looks bad on paper but converts at 5–10× the price, so the math still works. Compare your numbers to the placement-and-temperature row, never the platform average.

Two more nuances buyers miss. Industry matters less than placement in 2026 — Reels variance between e-commerce and B2B is smaller than the variance between feed and Reels inside the same vertical. And iOS vs Android still splits Meta's outbound numbers by 10–20%, with iOS lower because of the post-iOS14 attribution gap.

If you want to model your own placement-by-placement plan before launch, the CTR calculator and the audience saturation estimator take impressions and reach assumptions and project where you should land. Pull the live competitor floor for your category from unified ad search before you commit to a number.

CTR decay as a fatigue signal — the 15% rule

The rate on a winning ad does not stay flat. It peaks in days 3–7 of in-market delivery, plateaus, then decays. The shape of that decay is the single best ad fatigue signal Meta gives you that is not behind a paywall.

The trigger most experienced buyers use: a 15% drop in 7-day rolling outbound click-through rate versus the prior 7 days, on the same ad and same audience, is the action threshold. Below 15%, treat it as noise. At 15% or beyond, the creative has saturated its addressable pocket inside that audience and frequency is doing more harm than good.

The mistake is reacting to one bad day. The metric is volatile by hour and by day-of-week. A 7-day rolling window strips out that noise. The frequency cap calculator helps you set the upper bound on impressions-per-user before that decay even starts, which is cheaper than reacting after the fact.

Pattern we see often: a Reels ad will hold 0.9% outbound for two weeks, then drop to 0.6% in three days. That is not a bad week. That is the audience telling you the hook has been seen. Cross-check with the ad timeline analysis view to confirm whether the same audience pocket is being hit by two of your ads simultaneously, which compresses the decay curve.

Frequency above 3.5 in 7 days will accelerate the decay regardless of creative quality. Hit that ceiling and even the best creative angle saturates inside a week.

CTR vs hook rate vs ThruPlay — which to actually optimize for

Click-through rate is one of three creative-quality signals on Meta. The other two — hook rate and ThruPlay — measure earlier and later moments in the same view, and confusing them is one of the costliest analytics mistakes in 2026.

• Hook rate: percent of impressions that watch past 3 seconds. This is the gate. If hook rate is below 25% on Reels, nothing else matters.
• Outbound rate: percent of impressions that click through. The mid-funnel signal. Tells you the ad earned the click after the hook held.
• ThruPlay: percent that watched 15s or to completion. The signal Meta's delivery system actually rewards, because watch-time is what keeps users in-app.

Optimize hook rate first, ThruPlay second, the click-through metric third — in that order. A 2% rate with 18% hook rate will lose to a 1.3% rate with 38% hook rate every time, because Andromeda will give the second ad more impressions at lower CPM. The click-through metric alone, optimized in isolation, leads to clickbait that Meta then suppresses.

Looking at thumbnail patterns across in-market video ads on adlibrary, the first frame predicts hook rate, which predicts the rate, which predicts conversion rate. The chain breaks at the weakest link, and 80% of the time that link is frame one. The creative testing framework treats hook rate as the gating test before any other variant gets budget.

The hierarchy also explains a counterintuitive 2026 outcome: a lower-clicking ad sometimes scales better than a higher-clicking one. If watch-time and post-click dwell are both stronger, Meta's auction model — see Meta's ranking documentation — will pay you for the better signal.

The Andromeda era — why the click-through rate is now diagnostic

Meta launched Andromeda — its next-generation ranking and retrieval model — across the auction stack in late 2024 and expanded its remit through 2025. The short version: Andromeda evaluates roughly 10,000× more creative-and-audience pairings per second than the prior system. It uses dwell, watch-time, comment sentiment, and post-click signals as primary delivery inputs. The click-through metric is read, but it sits well below those signals in the model's ranking.

That has two practical consequences for 2026 buyers.

First, you can no longer "win" delivery on a high click-through number alone. An ad with strong outbound clicks but weak ThruPlay and short dwell will be throttled. Andromeda treats the click without the watch as a low-quality signal — possibly clickbait, possibly bot behavior. We have seen ads with 2.4% outbound get less impression share than ads with 1.1% outbound and 35-second average dwell on the same audience.

Second, the metric's diagnostic value is higher than ever. Because the model is making decisions you can't fully see, segment-level reads are one of the few transparent windows into how each creative is landing. A sudden drop on Reels but not feed tells you the vertical hook broke, not the angle. A flat reading with falling conversion rate tells you the angle promises something the landing page doesn't deliver. A drop on iOS but not Android usually means a tracking break, not a creative break.

Read the click-through rate like a doctor reads a thermometer. It does not tell you what is wrong. It tells you something is. See Meta's Andromeda announcement for the official framing and the Advantage+ shopping documentation for how the auction now consumes those signals.

Step 0 — angle research moves the rate more than headline tweaks

Before any of this matters, the angle has to land. The single biggest mover of the click-through metric is not button color, copy length, or thumbnail crop. It is whether the ad makes a promise the in-market audience already wants to believe.

Step 0: Find the angle on adlibrary first, then run the test.

Pull the top-performing ads in your category from the last 90 days. Cluster them by promise — speed, status, savings, social proof, fear, identity. The winners are usually 2–3 angles repeated by 12–15 advertisers. That clustering is the directional bet. Headline tweaks inside the wrong angle move the rate by 5–10%. Switching to the right angle moves it by 80–200%.

Then build the creative around that angle and test variants of execution — not variants of the angle itself. A practitioner workflow:

1. Open unified ad search, filter to your category and last 90 days
2. Sort by ad timeline duration — longer-running ads have proven angles
3. Pull 8–12 winners, write down the promise in one sentence each
4. Cluster the sentences. The 2–3 dominant clusters are the angles to test
5. Build 4–6 creatives per angle. Hold the angle constant, vary the hook
6. Launch into a clean ad set, segment by placement, watch the outbound metric by placement, and track decay by audience

The media buyer workflow walks the same loop end to end. The creative testing post covers test design once you have the angles. The creative angle post covers the cluster step in depth, and the creative iteration loop shows how to keep the angle fresh once it ships.

A pattern worth naming: when a category has 12+ advertisers running the same angle for 60+ days, that angle is the floor — not a winning bet. Differentiation lives in the second-most-common angle, which usually clears 1.5× the click-through baseline because it is still distinctive in the feed.

Common click-through rate mistakes in 2026

Five mistakes show up over and over in account audits.

1. Chasing the rate at the expense of conversion rate. A 3% click-through with 0.8% conversion loses to a 1.5% click-through with 3.5% conversion on a CPA basis every time. The first metric is mid-funnel. Conversion rate is bottom-funnel. Optimize for the dollar, not the click.

2. Comparing across placements without segmenting. A 1.0% feed reading is healthy. A 1.0% Reels reading is excellent. A 1.0% Audience Network reading is suspicious. Same number, three different verdicts. The media buyer workflow covers placement segmentation in depth.

3. Ignoring audience temperature. Cold prospecting at 1.4% is a winner. Retargeting at 1.4% is a problem — that audience already knows you and should click harder. See retargeting segmentation for temperature breakdowns.

4. Reading CTR (all) instead of outbound clicks. Already covered. Half of "great rate but no traffic" mysteries are this.

5. Reacting to daily noise. Day-of-week swings hit 30% on most accounts. Use 7-day rolling windows, not yesterday's number. Pair with CPM and ROAS trends — if CPM rose and the rate fell on the same day, the auction got harder, not your creative worse.

The deeper pattern: this metric is relative. It only means something compared to your own baseline, the placement benchmark, and the angle's expected ceiling. Stripped of that context, the number is a vanity stat dressed up as a KPI. Buyers who treat it as one signal in a chain — alongside hook rate, ThruPlay, frequency, and dwell — outperform buyers who treat it as the score.

Frequently asked questions

What is a good CTR for Facebook ads in 2026?

A good Facebook feed CTR is 1.0–2.0% outbound, with 1.2% as a healthy cold-prospecting target. Reels run lower at 0.5–1.2% and Stories at 0.4–1.0%. Anything above 2.5% on cold feed traffic is unusually strong and worth investigating for click-bait risk under Andromeda's quality scoring.

How is CTR calculated?

CTR is calculated as (clicks ÷ impressions) × 100, expressed as a percent. On Meta, always use outbound CTR (link clicks that leave the platform) rather than CTR (all), which inflates the number with reactions, profile taps, and comment expansions.

Does CTR still matter with Meta's Andromeda model?

Yes, but as a diagnostic rather than a delivery lever. Andromeda weights dwell, watch-time, and post-click quality above CTR when ranking ads. Use CTR to spot fatigue, hook problems, and placement mismatches — not as the primary optimization target. ThruPlay and hook rate matter more for delivery in 2026.

When should I refresh creative based on CTR decay?

The standard trigger is a 15% drop in 7-day rolling outbound CTR versus the prior 7-day window on the same ad and audience. Below 15% is normal lifecycle noise. At 15%+ the audience is saturating and fresh angles or hooks should rotate in within 3–5 days.

What's the difference between CTR and hook rate?

Hook rate measures the percent of impressions that watch past 3 seconds — a video-only metric. CTR measures the percent of impressions that click. Hook rate sits earlier in the funnel and is the gate; CTR sits after, measuring whether the ad earned the click after the hook held attention.

Bottom line

CTR in 2026 is a diagnostic, not a destination. Read it segmented by placement, by audience temperature, and against a 7-day rolling window — and pair it with hook rate and ThruPlay before you draw any conclusion. The buyers winning under Andromeda are the ones who treat CTR as one signal in a chain, not the chain itself.