Meta Ads Playbook

Before you spend a single pound, your tracking needs to be airtight. This is the most overlooked part of running Meta Ads — and it's the reason most brands can't tell whether their campaigns are actually working.

The Meta Pixel records what happens after someone clicks your ad: pages they visit, products they view, items added to cart, purchases made. Without it properly configured, Meta is optimising blind.

"Most accounts we audit have the Pixel installed but the wrong events firing. Meta is optimising for the wrong thing and the brand has no idea."

The Essential Setup

Install the Pixel via Shopify Integration

Use the native Shopify integration in Meta Events Manager rather than pasting code manually. It enables automatic event matching, which improves signal quality considerably.

Enable the Conversions API (CAPI)

iOS 14 changes eliminated a significant share of browser-based tracking. CAPI sends event data server-side, directly from your store to Meta. Shopify has a native CAPI connector in the Meta app.

Why CAPI Matters

On a healthy account, CAPI recovers 15 to 40% of purchase events that browser tracking would miss. That's the difference between Meta thinking a campaign is breaking even versus profitable.

Configure All Five Standard Events

These must fire in the correct order. Missing or misfiring events mean broken optimisation signals.

Event	Where It Fires	Priority
PageView	Every page	Essential
ViewContent	Product detail pages	Essential
AddToCart	Cart add action	Essential
InitiateCheckout	Checkout page load	Essential
Purchase	Order confirmation	Essential

Each event must pass the correct value, currency, and content_id parameters. If Purchase isn't passing revenue, you can't calculate ROAS inside Meta.

Conversions API (CAPI) is now essential, not optional. iOS privacy changes have reduced pixel signal quality significantly. CAPI sends server-side events directly to Meta, bypassing browser restrictions. Brands running both Pixel and CAPI see 15–20% more purchase events reported. Check your Event Match Quality score in Events Manager — any score below 6.0 needs immediate attention.

Set Attribution to 7-Day Click / 1-Day View

This is the standard for DTC ecommerce. A 1-day click window underreports on higher-consideration products. A 28-day view window overreports massively. Pick one, stick to it.

UTM Parameters on Every Ad

Meta's in-platform reporting shows what Meta wants you to see. UTMs in GA4 show what actually happened. Use both.

utm_source=meta
utm_medium=paid_social
utm_campaign={{campaign.name}}
utm_content={{ad.name}}

Foundations Checklist

Pixel installed via Shopify Meta integration
Conversions API enabled and deduplication active
All 5 standard events firing with correct parameters
Purchase event passing value and currency
Attribution window set to 7-day click / 1-day view
UTM parameters added to every ad
All events verified green in Meta Events Manager

Most underperforming Meta accounts share the same problem: too many campaigns, too many ad sets, too many ads competing for the same audience. The result is overlap, auction competition, and fragmented signal data Meta can't learn from. Meta's Andromeda algorithm update in 2025 made this worse — the system now uses your creative to find audiences, meaning diverse creative variety matters more than audience segmentation. Brands still running 10+ ad sets with micro-audiences are fighting the algorithm instead of feeding it.

The Three-Campaign Model

Prospecting

Find New Customers

Broad targeting, interests, or lookalikes. Top of funnel. Bring in people who haven't heard of you.

Broad — let Meta optimise
Interest stacking as one ad set
LAL from purchasers

Retargeting

Re-engage Warm Audiences

Site visitors, product viewers, add-to-cart non-purchasers. Separate budget, separate creative, different message.

ATC non-purchasers — 30 days
Product page views — 14 days
Site visitors — 7 days

Retention

Win Back Past Buyers

Past customers who haven't purchased in 60–180 days. Often the most efficient spend in the whole account.

Past purchasers 60–180 days
New product launches
Replenishment offers

Budget Allocation by Stage

Stage	Prospecting	Retargeting	Retention
Early (0–12 months)	70%	20%	10%
Growth (12–36 months)	55%	25%	20%
Established (36+ months)	45%	25%	30%

CBO — Campaign Budget

Budget at Campaign Level

Meta shifts budget to the best-performing ad set in real time. Works well with 3+ ad sets and sufficient data.

Best for: accounts spending £100+/day per campaign.

ABO — Ad Set Budget

Budget at Ad Set Level

More control, more manual work. Use when testing a new audience and you want to control spend before Meta has data.

Best for: testing phases, budgets under £50/day.

The Consolidation Rule — Updated 2026

More than 3–4 ad sets per campaign is too many. Meta needs 50 optimisation events per ad set per week to exit the learning phase — this now applies at the weekly level, not daily. Fragmented structure makes that nearly impossible. The Andromeda algorithm rewards consolidated accounts: fewer campaigns with more data per campaign consistently outperform fragmented setups. Any edits during the first 7 days reset the learning phase entirely.

The conversation around Meta audiences changed after iOS 14. Broad targeting — giving Meta minimal restrictions and letting the algorithm find buyers — has become more effective for many accounts. That doesn't mean audiences don't matter. It means knowing when to use what.

Audience Types Ranked

Audience Type	Source	Quality	Best For
Customer List (purchasers)	CRM / Klaviyo export	Very High	LAL base, exclusions
Purchase Event LAL	Meta Pixel events	Very High	Top-of-funnel prospecting
Website Visitors LAL	Meta Pixel	High	Prospecting at scale
Engagement LAL	Instagram / Facebook	Medium	Scale when saturating
Interest Stacking	Meta's interest database	Medium	New accounts, small budgets
Broad (no targeting)	Algorithm driven	Variable	Large budgets, strong creative
Advantage+ Audience	Meta AI-driven	Very High	Default for ASC — automates all audience decisions

"Go broad before you go narrow. Meta's algorithm is better at finding your buyers than you are at defining them."

Retargeting by Intent Level

High Intent

Add to Cart / Initiate Checkout — 7 daysHighest priority

Mid Intent

Product Page Views — 14 daysCore retargeting

Low Intent

All Site Visitors — 30 days

Awareness

Social Engagers — 60 days

Creative is the single biggest variable in Meta performance. Two brands with identical structure, identical targeting, identical budgets can have completely different results based on what their ads look like. The goal is a repeatable system: a small number of formats, a clear testing process, and a way to identify winners fast and scale them before they burn out.

Creative Formats

Static Image

Product-Led Statics

Clean product photography or lifestyle imagery with a punchy overlay. Fast to produce, easy to test, still highly effective.

Best sizes: 1:1, 4:5

Video

Short-Form Video

Hook in the first 2 seconds. No longer than 30 seconds for prospecting. UGC-style typically outperforms polished production in cold audiences.

Best ratio: 9:16, 4:5

Carousel / DPA

Multi-Product Carousels

Multiple products or multiple angles. Dynamic Product Ads pull from your catalogue automatically. Best for retargeting where intent is established.

Best for: retargeting, catalogues

The Hook Framework

The hook is the first 2–3 seconds of video, or the first line of your headline for statics. It's the only thing that determines whether someone stops scrolling or keeps going.

Hook Type	Example	Works Best For
Problem-led	"Most brands don't realise they're doing X wrong"	Cold audiences
Outcome-led	"I trained 5 days a week for 3 months using this"	Aspirational lifestyle
Curiosity gap	"The scent people keep asking me about"	Beauty, fragrance, fashion
Direct offer	"Free shipping today. [Product] from £X."	Warm audiences, retargeting

Testing Process

Test One Variable at a Time

When an ad wins, you need to know why. Test hooks first — highest impact. Once you have winning hooks, test bodies. Then formats. One variable per round.

Launch 10–20 Variations and Keep Producing

Meta's algorithm needs 15–50+ active creatives to optimise properly in 2026 — this benchmark has doubled from two years ago. Brands producing fewer than 10 new creatives per month see a 35–45% increase in CPA when scaling spend by more than 30%. Load variations across at least three formats (static, video, UGC) and let the algorithm find winners. Give each creative a minimum 7 days and 1,000 impressions before calling it.

Define Kill Criteria Before Launch

Before the campaign goes live, decide what numbers mean a creative is dead.

2%+

CTR Min

Kill if below

<£1.50

Max CPC

Category dependent

Min ROAS

For prospecting

7 days

Min Run

Before decisions

Scale Winners, Kill Losers Fast

When a creative is above target ROAS for 7 days, scale by duplicating the ad set and increasing budget 20–30%. Don't increase the existing ad set's budget by more than 20% — it resets the learning phase.

Creative Fatigue

Every ad has a lifespan. CTR drops week on week, CPC climbs, ROAS falls. Don't wait for top performers to fully die before you have replacements ready. Always keep new creative in the pipeline.

How you launch has a direct impact on long-term performance. The learning phase is Meta's algorithm calibrating to your objective. Disrupting it through budget changes, creative swaps, or audience edits resets the clock and wastes your data.

Pre-Launch Checklist

Pixel and CAPI verified green in Events Manager
Purchase event passing correct value and currency
Custom audiences built and ready
Creative reviewed on mobile — 85% of Meta traffic is mobile
Copy reviewed — no grammar errors, no broken characters
Destination URLs tested, landing pages loading correctly
UTM parameters added to every ad
Campaign objective set to Conversions, not Traffic

Learning Phase Rules

What Resets Learning

Budget change over 20%
Editing audience targeting
Adding or removing ads
Changing bid strategy
Changing optimisation event
Pausing for more than 7 days

How to Exit Faster

Launch with enough budget (50 events in 7 days)
Fewer, larger ad sets
Use Advantage+ placements
Don't touch it in the first 7 days
Optimise for ATC if purchase volume is low

Budget Calculation

To exit learning in 7 days, Meta needs 50 purchase events. At a £30 CPA, that's a minimum of £1,500 over 7 days (£215/day). If budget is lower, optimise for Add to Cart until you build enough purchase data.

First 30 Days

Period	Focus	Action
Days 1–7	Learning phase	Do not touch the campaign.
Days 8–14	First review	Pause clear underperformers. Keep top creatives.
Days 15–21	Optimise	Consolidate learnings, refresh creative if CTR dropping.
Days 22–30	Scale decision	If CPA and ROAS on target, begin cautious budget increases.

Scaling is not just increasing budget. It's maintaining or improving efficiency as you grow. Most accounts hit a wall because they push spend up faster than their creative and audience strategy can support.

Vertical vs Horizontal

Vertical Scaling

Increase Budget on Winners

Raise budget 15–20% every 3 to 4 days. More than 20% risks resetting learning.

Use when: clear winner, want to push harder.

Horizontal Scaling

Duplicate Into New Audiences

Duplicate the winning ad set into a new audience. Expand LAL percentages, test new interests. Keep the same creative.

Use when: frequency above 3, audience saturating.

Scaling Signals

Signal	What It Means	Action
ROAS above target 7+ days	Efficient and stable	Increase budget 20%
Frequency above 3.5	Audience overexposed	Refresh creative or expand audience
CTR dropping week on week	Creative fatigue	Test new hooks, pause losers
CPA rising, volume holding	Audience saturation	Horizontal scale to new audiences
Learning phase recurring	Too many edits	Consolidate, stop touching, let it run

"The brands that scale Meta Ads sustainably treat it like a system. Test, learn, iterate. The ones that burn out treat it like a tap — just turn it up."

Advantage+ Shopping Campaigns (ASC) — 2026 Standard

ASC is now the default campaign structure for ecommerce on Meta. Accounts with ASC as their primary campaign see 15–25% lower CPA than manually targeted campaigns at spend levels above £1,000/day. The recommended 2026 structure: one ASC as your scaling engine (60–70% of budget), one broad prospecting campaign (20–30%), and a retargeting campaign capped at 5% of budget. Set the existing customer budget cap to 20–30% to prevent over-indexing on warm audiences. Minimum £100/day to give the algorithm enough data to exit the learning phase.

Meta Ads Manager surfaces dozens of metrics. Most are vanity numbers or misleading without context. Knowing which numbers tell you whether campaigns are actually working is what separates good decisions from gut-feel reactions.

Metrics That Matter

Metric	What It Measures	Good Benchmark	Priority
ROAS	Revenue per £1 of spend	ASC average 4.52x (2026); 3x+ prospecting minimum	Primary
CPA	Cost per purchase	<35% of AOV	Primary
CTR (Link)	Creative performance	0.9–2% average; 3%+ strong (2026)	Secondary
CPC (Link)	Cost per click	£0.40–£1.20 UK typical	Secondary
CPM	Cost per 1k impressions	£8–£20 UK Q1–Q3	Context
Frequency	Avg times ad seen	<3 for prospecting	Diagnostic
Add to Cart Rate	Landing page strength	3–8% of clicks	Diagnostic

Ignore: Reach & Impressions

These tell you how many people saw your ad — not whether it drove revenue. 2 million impressions with no purchases is a failure.

Ignore: Post Engagement

Likes, comments, and shares have almost no correlation with purchase behaviour in DTC. Don't optimise for them.

Reporting Rhythm

Daily

5-Minute Check

Confirm campaigns are active and spend is tracking to plan. Flag CPA shifts over 40% above target. Do not make optimisation decisions daily.

Weekly

30-Minute Review

Review last 7 days. Identify top and bottom performers. Pause losers, scale winners. Check frequency on all ad sets.

Monthly

2-Hour Analysis

Review against targets. Is ROAS improving or declining? Make structural decisions: budget reallocation, audience expansion, next month's creative strategy.

The Blended ROAS Problem

Blended ROAS — total revenue divided by total spend — can be misleading. A high blended ROAS might be driven entirely by your retention campaign reaching people who were going to buy anyway, masking weak prospecting performance. Always look at ROAS at campaign level.

The Meta Ads

Install the Pixel via Shopify Integration

Enable the Conversions API (CAPI)

Configure All Five Standard Events

Set Attribution to 7-Day Click / 1-Day View

UTM Parameters on Every Ad

Prospecting

Find New Customers

Retargeting

Re-engage Warm Audiences

Retention

Win Back Past Buyers

CBO — Campaign Budget

Budget at Campaign Level

ABO — Ad Set Budget

Budget at Ad Set Level

Static Image

Product-Led Statics

Video

Short-Form Video

Carousel / DPA

Multi-Product Carousels

Test One Variable at a Time

Launch 10–20 Variations and Keep Producing

Define Kill Criteria Before Launch

Scale Winners, Kill Losers Fast

What Resets Learning

How to Exit Faster

Vertical Scaling

Increase Budget on Winners

Horizontal Scaling

Duplicate Into New Audiences

5-Minute Check

30-Minute Review

2-Hour Analysis

The Blended ROAS Problem

Paid Media Management

Creative Strategy

Full-Funnel Growth