Most marketers waste their first £500 on Meta Ads making the same three mistakes, wrong objective, bloated audience, and no creative strategy. This workshop gives you the exact build process that Byter uses with real clients, so you skip the expensive lessons and go straight to campaigns that convert.
Workshop: Build a Meta Ads Campaign from Scratch
This is a build session, not a theory lecture. By the end of this 60-minute workshop, you will have a structurally sound, strategically considered Meta Ads campaign live in Ads Manager. Not clicked around the interface hoping for the best. Not watched someone else do it on a screen recording. Built it yourself, correctly, from the ground up. We are going to move through every layer: objective selection, audience architecture, placement strategy, creative briefing, and budget logic. The sequence matters. Do not skip ahead.
Pull up your Meta Ads Manager. We are building as we go.
Why Meta Ads Still Matter in 2025
Before we touch the interface, let us ground this in reality.
PA406-01: Workshop: Build a Meta Ads Campaign, Key Concepts
Meta's Q4 2024 earnings report confirmed the platform now reaches over 3.27 billion daily active people across its family of apps. For UK advertisers specifically, the picture is even sharper: Meta's average revenue per user in Europe grew by 21% year-on-year, a direct signal that advertiser competition is intensifying and that precision is now the differentiator. According to Ofcom's 2024 Online Nation report, 71% of UK adults use Facebook or Instagram at least weekly, with Instagram reach among 18 to 34-year-olds exceeding every other paid social channel. That is the audience. The question is how to reach them without haemorrhaging budget.
According to HubSpot (2025), 67% of marketers cite Meta as their primary paid social channel, yet fewer than a third feel confident they are using campaign objectives correctly. That gap is exactly where money gets wasted.
Meta's advertising ecosystem has also changed substantially since the iOS 14.5 privacy changes of 2021. Modelled conversions, aggregated event measurement, and the shift toward server-side tracking via the Conversions API have all altered how campaigns are structured and measured. If you learned Meta Ads more than two years ago and have not revisited your fundamentals, significant parts of your knowledge base are likely out of date. This workshop is built around how the platform actually behaves today, not how it behaved in 2020.
Phase 1, Campaign Level: Choosing Your Objective
The first decision in Ads Manager is also the most consequential. Meta's Advantage+ Campaign Budget system will optimise your entire spend around whatever objective you declare. Choose incorrectly and you will get exactly what you asked for, just not what you wanted.
Meta currently organises objectives into six categories under the ODAX framework (Outcomes-Driven Ad Experiences): Awareness, Traffic, Engagement, Leads, App Promotion, and Sales.
Here is how to choose correctly:
Awareness: use this only when brand recall is the genuine goal. Do not use it expecting clicks or conversions. It is appropriate for top-of-funnel brand building when reach and impressions are the success metrics.
Traffic: optimises for link clicks or landing page views. Useful for content-led funnels, but it attracts a broader audience than pure converters. A good stepping stone when your pixel is new.
Engagement: useful for growing social proof on a post before boosting spend, or for warming cold audiences with low-cost interactions before following up with a Sales campaign.
Leads: activates Meta's native Lead Forms or connects to your CRM. Strong for service businesses, financial advisers, estate agents, and anyone whose sales cycle involves a discovery call or consultation.
Sales: the workhorse objective for e-commerce. Requires your Meta Pixel or Conversions API to be firing correctly before you launch. Meta will optimise specifically for purchase events if that is what you designate as your conversion event.
A practical rule of thumb: choose the objective that corresponds to the action you are willing to pay for. If you are paying for purchases, use Sales. If you genuinely want leads into a CRM, use Leads. Never use Traffic and hope people will buy. Meta will send you people who click, not people who convert.
Warning
If your Pixel has fewer than 50 conversion events in the last 7 days, the Sales objective will struggle to exit the learning phase. Consider starting with Traffic and warming the pixel before switching objectives. Alternatively, set your conversion event to something higher in the funnel, such as Add to Cart or Initiate Checkout, until purchase volume is sufficient.
A real example: A Byter client selling artisan coffee subscriptions came to us having run a Traffic objective for three months, puzzled as to why their cost per purchase was £48 when their product sold for £18. The campaign was doing exactly what it was told, driving link clicks from curious browsers, not buyers. Within two weeks of switching to Sales with an optimised Conversions API setup, cost per purchase dropped to £14.20. Same creative. Same budget. Wrong objective fixed.
Workshop Action, Phase 1:
Open Ads Manager, click "Create", and select your campaign objective. Name your campaign using a clear convention: [Client/Brand] | [Objective] | [Date] | [Test Label]. For example: ByterDemo | Sales | Jan2025 | ColdAudienceTest. Good naming hygiene will save you hours when you are managing multiple campaigns across multiple clients. Agree on a naming convention before you build anything, and document it in a shared Google Doc or Notion page.
Phase 2, Ad Set Level: Audience Architecture
This is where most beginners over-engineer. They stack interest upon interest, layer demographic restrictions, and end up with an audience of 40,000 people that Meta cannot optimise effectively. The algorithm needs room to breathe. When you shrink your audience too aggressively, you force Meta to show your ads to the same small group of people repeatedly, driving up frequency and burning out your budget without meaningful results.
The modern Meta Ads approach is built on what practitioners call the Broad-Core-Retarget model:
Broad: minimal targeting, age range and location only, letting Meta's algorithm find converters through its own signals. This is increasingly effective and often outperforms interest targeting, especially for accounts with healthy pixel data.
Core: interest and behaviour stacks for specific personas. Aim for audiences above 1 million in the UK for most campaigns. Stack no more than five interests per ad set. More than that and you lose the ability to diagnose what is actually working.
Retarget: custom audiences built from pixel data, video viewers, Instagram engagers, or customer lists. This tier should always carry your highest-value messaging since these people already know who you are.
This maps directly to the Byter 3-Layer Targeting Model: Cold audiences (Broad and Core interest targeting for awareness), Warm audiences (video viewers, page engagers, site visitors), and Hot audiences (cart abandoners, enquiry form submitters, lapsed customers). Every pound you allocate should have a clear layer designation before you touch Ads Manager. If you cannot answer which layer an ad set belongs to, you are not ready to build it.
For a new campaign with no pixel history, start with one Broad ad set and one Core ad set. Do not build a retargeting ad set until you have at least 1,000 pixel events to work with. Otherwise the audience is too small for meaningful optimisation and you will waste budget cycling through the same tiny pool of people.
Audience sizing guidance for UK campaigns:
Cold audiences: 1M–5M is a healthy range
Lookalike audiences: typically 500K–2M depending on the percentage used (1% lookalikes in the UK are roughly 600K–700K)
Retargeting audiences: 5,000–200,000
Lookalike audiences are worth a specific mention. A 1% lookalike of your customer list, the 1% of the UK population most statistically similar to your existing buyers, is one of the most powerful tools in Meta's arsenal. You need a minimum of 100 matched customers to generate a lookalike, but ideally 500 or more for reliable results. Upload your customer list as a custom audience first, then build the lookalike from it.
Budget split suggestion for beginners: 60% Broad, 40% Core. Review after 7 days. Do not adjust allocations mid-learning phase.
Byter Tip
Byter Insider: We ran a cold audience build for a fitness apparel brand based in Shoreditch. The client had been running five tightly interest-stacked ad sets for four months and was convinced they knew their audience. We stripped it back to two ad sets: one Broad (UK women 22 to 40, no interests) and one Core (fitness, athleisure, and sportswear interests, audience size 1.4M). Within the first 14 days under CBO, the Broad ad set was taking 68% of budget and delivering a cost per purchase of £11.20 versus the Core set's £18.40. By month two, we had cut the Core set entirely. The client's blended ROAS moved from 1.8x to 3.4x on the same monthly spend.
Workshop Action, Phase 2:
Create two ad sets. Ad Set 1: Broad (UK, ages 25–54, no interests). Ad Set 2: Core (add 3–5 relevant interests, keep the audience above 1M). Name them clearly: AdSet_Broad_UK and AdSet_Core_UK. Set identical budgets at this stage. We will let Meta's CBO logic operate. Resist the urge to add exclusions beyond your existing customer list. Exclusions reduce reach and can actually hurt learning phase performance.
Phase 3, Ad Level: Creative Strategy
Creative is now the primary targeting mechanism on Meta. According to Hootsuite's Social Media Trends Report (2025), ads with native-feeling creative (shot on mobile, minimal branding, text-on-screen) achieve up to 32% lower CPMs than polished studio content in cold audience testing.
This does not mean low quality. It means contextually appropriate quality. A beautifully produced brand film belongs on YouTube pre-roll, not in an Instagram Stories placement where a lo-fi, subtitled talking-head video will outperform it at a fraction of the production cost.
For each ad set, you should test a minimum of three creative variants across two formats: static image and short-form video (under 30 seconds). This is your 3x2 creative matrix: three messaging angles, two formats.
Your three messaging angles should map to different psychological triggers:
Problem-aware: speak directly to a pain point your audience already recognises. "Still spending 3 hours a week on your accounts?"
Solution-focused: lead with the outcome or transformation. "What 200 business owners did to halve their admin time."
Social proof: testimonials, reviews, user-generated content. Screenshots of genuine five-star reviews perform particularly well as static ads because they feel earned rather than manufactured.
For each creative, your copy structure should follow the Hook-Bridge-CTA model:
Hook (first line of primary text and first 3 seconds of video): stop the scroll. This must be specific and immediate. Vague hooks kill performance.
Bridge (body): connect the hook to the offer with one or two sentences of substance. Do not over-explain. You are creating curiosity, not delivering a brochure.
CTA (final line + button): tell the user exactly what to do next. "Shop now", "Get your free quote", "Download the guide". Specificity outperforms generic calls to action every time.
This hook-first approach is the foundation of Byter's Hook-Hold-Convert Method. Hook in the first 3 seconds, hold attention for at least 15 seconds with substance, then convert with a single unambiguous CTA. Every ad you build in this workshop should be evaluated against those three beats before it goes live. If you cannot identify a clear hold mechanism in your creative, the ad is not ready.
Format-specific guidance:
Static images should be 1080x1080px (square) for feed placements and 1080x1920px (vertical) for Stories and Reels. Minimal text overlay. Bold contrast. Clear focal point.
Video should front-load the hook in the first two seconds. Add captions. 85% of Meta video is watched without sound. Keep total duration under 30 seconds for cold audiences. 15 seconds is a strong target.
Carousel ads work well for product catalogues or multi-step storytelling. Each card should be able to stand alone. Do not rely on users swiping through all cards.
Tool recommendations:
Canva Pro: rapid static ad production with brand kit consistency. Use the batch resize feature to produce all format variants from a single design in minutes.
CapCut: mobile-first video editing that produces native-feeling content. The auto-caption feature is excellent and saves significant post-production time.
AdCreative.ai: useful for generating initial static concepts at scale, though always refine with human judgement. The AI tends to produce visually safe, generic executions without a strong creative brief to direct it.
Foreplay.co: an ad inspiration library for researching competitor creative before you brief your own. Build a swipe file of ads in your vertical before beginning any creative production.
Motion: for tracking creative performance metrics across your ad account over time, helping you identify which angles and formats are fatiguing and what to test next.
Workshop Action, Phase 3:
Write the copy for one static ad using the Hook-Bridge-CTA model. Keep your primary text under 125 characters for the hook line. Select or design one image. Upload to your ad set and set the destination URL with UTM parameters appended. Your UTM parameters should at minimum capture utm_source=facebook, utm_medium=paid-social, and utm_campaign=[your campaign name]. This data flows into Google Analytics 4 and becomes invaluable when you are reporting to a client or stakeholder.
Exercise
Complete the Hook-Bridge-CTA structure for a fitness app targeting busy professionals: Hook: 'No time to train? ___' Bridge: 'ByterFit's 15-minute workouts are built for people who ___ but still want ___.' CTA: '___'
Understanding the Learning Phase
The learning phase is one of the most misunderstood concepts in Meta Ads, and failing to respect it is responsible for a disproportionate number of campaign failures. When you first launch a campaign, or make a significant edit to an existing one, Meta's delivery system enters a period of exploration. It is testing different users, times, placements, and creative combinations to understand how best to deliver your ads.
During this period, performance is intentionally unstable. You will likely see higher CPMs, lower conversion rates, and erratic daily results. This is normal and expected. The mistake is to react to this instability by making changes: adjusting budgets by more than 20%, swapping creatives, editing audiences. Each of those actions resets the learning phase and traps the campaign in a perpetual state of exploration.
The learning phase officially exits when your ad set accrues 50 optimisation events within a 7-day period. For a purchase-optimised campaign with a £20 CPA target, this means you need to be spending enough daily budget to generate roughly 7 purchases per day. If your budget is too low to hit 50 events in a week, consider switching to a higher-funnel conversion event temporarily (Add to Cart, View Content) to accumulate enough signal for Meta to work with.
Signs your campaign is stuck in learning:
"Learning" badge remains on the ad set after 14+ days
Highly erratic day-to-day cost per result
Delivery warnings flagging "Learning Limited" status
If you see "Learning Limited", the most common causes are: audience too small, budget too low relative to target CPA, or too many ad sets competing for the same audience.
The Meta Learning Phase explained: what happens, what resets it, and how to exit it faster
Phase 4, Budget Logic and Campaign Structure
Budget decisions on Meta operate at two levels: the campaign level (CBO, Campaign Budget Optimisation, now branded as Advantage+ Campaign Budget) and the ad set level (ABO, Ad Set Budget Optimisation). Understanding the difference is critical.
CBO (Campaign Budget Optimisation): You set one daily or lifetime budget at the campaign level and Meta distributes it dynamically across ad sets based on real-time performance signals. This is the recommended approach for most campaigns and is what we use at Byter for the majority of client accounts.
ABO (Ad Set Budget Optimisation): You set individual budgets for each ad set. This gives you more manual control but requires more active management. It is useful when you have a specific retargeting ad set that needs a guaranteed minimum spend, or when you are deliberately isolating variables in a structured test.
Budget sizing guidelines for UK market campaigns:
Proof of concept testing (new account, new product): £20–£40 per day minimum across the campaign. Below this, Meta lacks sufficient daily signal to make meaningful optimisation decisions.
Scaling phase (proven offer, exited learning): Increase budgets by no more than 20% every 3–4 days to avoid re-entering the learning phase. Avoid doubling budgets overnight.
Mature campaigns: Consider using Advantage+ Shopping Campaigns for e-commerce, which automate much of the audience and placement logic and have shown strong ROAS improvements for accounts with rich pixel histories.
Daily versus lifetime budgets:
Daily budgets give you predictable daily spending but do not allow Meta to front-load spend on high-performing days.
Lifetime budgets with a defined end date allow Meta to redistribute spend across the schedule based on when it expects the best results. Useful for time-limited promotions and event-driven campaigns.
This budget sequencing also connects to the Byter 72-Hour Rule: the first 72 hours of any campaign determine its trajectory. Front-load your attention here. Check Events Manager on day one to confirm pixel events are firing. Check delivery on day two to confirm the campaign has exited any initial review period. Check early CPM and frequency on day three to identify any immediate structural problems before they compound. What you do in those first three days sets the ceiling for the campaign.
Byter Tip
For evergreen campaigns, those with no defined end date, always use daily budgets. Lifetime budgets on campaigns with no end date can result in erratic spend patterns as Meta lacks the scheduling context it needs to distribute budget sensibly.
Workshop Action, Phase 4:
Set your campaign budget using CBO at £30 per day. Confirm your billing threshold in Ads Manager settings so you know when Meta will charge your payment method. Set an account spending limit as a hard ceiling. This is a safety net that prevents unexpected overspend if you forget to pause a campaign. Navigate to Billing > Account Spending Limit in your account settings and set it now before you launch anything.
Common Mistakes Practitioners Make
Understanding where campaigns break down is as valuable as knowing how to build them correctly. Here are the five most damaging errors we see, drawn directly from client audits and account rescues over several years of agency work:
Launching without a working Pixel or Conversions API. If Meta cannot see your conversion events, it cannot optimise for them. Verify your Pixel is firing in Events Manager before spending a single pound. Use the Meta Pixel Helper Chrome extension to confirm events are firing correctly on every key page: landing page, add to cart, checkout, and purchase confirmation.
Editing campaigns during the learning phase. Every significant edit resets the learning phase. Set your campaign, then resist the urge to tweak for at least 7 days or 50 optimisation events, whichever comes first. Print this rule out and pin it above your monitor if necessary.
Running too many ad sets on a small budget. A £30/day budget split across five ad sets gives each one £6, not enough data to learn. Consolidate. The modern best practice is fewer, larger ad sets rather than a fragmented structure with many small audiences.
Ignoring creative fatigue. Meta's own research suggests ad frequency above 3.5 within a 7-day window begins to hurt performance. Monitor frequency in your Ads Manager columns and refresh creative proactively, not reactively. Build a content calendar for creative rotation before you launch, not after performance drops.
Measuring the wrong metric. Click-through rate is vanity on Meta. Your north-star metric should be Cost Per Result (CPR) aligned to your objective: cost per purchase, cost per lead, or cost per landing page view. A 0.5% CTR with a £9 CPA is infinitely better than a 3% CTR with a £40 CPA.
A sixth mistake worth highlighting: not excluding existing customers from cold audience campaigns. Upload your customer list as a custom audience and exclude it from cold ad sets. Otherwise you are paying to advertise to people who have already bought from you, inflating your apparent conversion rate with warm buyers and skewing your data.
The 3×2 Creative Matrix: how to structure your ad creative testing across three angles and two formats
Phase 5, Review and Launch Checklist
Before you publish, run through every layer of the campaign with fresh eyes. Many costly errors are caught at this stage: a mismatched pixel event, a destination URL that 404s, or a creative uploaded in the wrong aspect ratio. These are entirely preventable problems that disappear with a disciplined pre-launch review.
Action Steps
0/17 completed
Key Takeaways
The campaign objective is the single most important decision in Ads Manager. Choose based on where your pixel data is, not just where you want to be.
The Broad-Core-Retarget audience model is a proven structure for scaling Meta campaigns methodically. Broad often wins. Do not dismiss it in favour of over-engineered interest stacks.
Creative is now audience targeting. Invest time in your 3x2 creative matrix before launch. Three messaging angles across two formats gives you six testable ads and a structured path to identifying your winning angle.
Respect the learning phase: 50 optimisation events or 7 days minimum before drawing conclusions or making changes. Every edit resets the clock.
The minimum daily budget formula, (Target CPA × 50) ÷ 7, tells you if your budget is sufficient to exit learning within a week. If the maths do not work, adjust your conversion event or increase your budget before launch.
Cost Per Result aligned to your objective is your true north-star metric. Not CTR, not reach, not impressions.
Naming conventions and UTM tagging are not optional extras. They are what make campaigns manageable at scale and reportable to clients and stakeholders with confidence.
Action Step
Use the checklist above to audit either a live campaign you are currently running or the campaign you have just built in this workshop. For every item you cannot tick, note the specific fix required and a deadline to implement it.
If you are building fresh, your immediate next step is to verify your Pixel is firing correctly in Meta's Events Manager before setting your campaign live. No Pixel, no meaningful optimisation. Open Events Manager now, navigate to the Test Events tab, and trigger each key event on your website manually to confirm they are being received. Only once every event shows a green confirmation should you proceed to launch.