Facebook advertising remains one of the most effective paid marketing channels available. Despite competition from TikTok and changing privacy regulations, Meta's platforms (Facebook and Instagram combined) reach over three billion monthly active users. The targeting capabilities are unmatched, the cost is accessible to businesses of any size, and the platform has more performance data than any other advertising network except Google.

The problem is not that Facebook advertising does not work. The problem is that most people set it up badly — wrong objective, wrong audience, wrong creative, wrong budget allocation — and then conclude that the platform is broken.

This guide covers how to set up, run, and optimise Facebook campaigns that actually generate results. Whether you are selling products, generating leads, or driving website traffic, the principles are the same.

How Facebook Advertising Works

Facebook advertising operates on an auction system. When you create a campaign, you are competing against other advertisers who want to reach the same audience. Facebook decides which ads to show based on three factors.

Bid. How much you are willing to pay per result. This can be automatic (Facebook decides) or manual (you set a cap).

Estimated action rate. Facebook's prediction of how likely the person is to take your desired action — click, purchase, sign up. This is based on the person's past behaviour and the ad's historical performance.

Ad quality. How relevant and engaging your ad is, measured by feedback signals. Positive engagement (clicks, comments, shares) improves quality. Negative feedback (hiding the ad, reporting it) lowers quality.

The ad with the highest combined score wins the auction. This means a lower bid with excellent creative can beat a higher bid with poor creative. Quality matters as much as budget.

Campaign Structure: The Three Levels

Every Facebook campaign has three levels. Understanding this hierarchy is essential for organising your advertising effectively.

Campaign level

This is where you choose your objective — what you want the campaign to achieve. Facebook's current objectives are awareness (reach as many people as possible), traffic (send people to your website), engagement (get likes, comments, shares), leads (collect contact information), app promotion (drive app installs), and sales (drive purchases on your website).

Choose the objective that matches your actual goal. If you want website visitors, choose traffic. If you want purchases, choose sales. If you want email signups, choose leads. Facebook's algorithm optimises toward whatever objective you select. Choosing the wrong one means Facebook optimises for the wrong outcome.

The most common mistake beginners make is choosing engagement (likes and comments) when they actually want sales or leads. Likes do not pay bills. Optimise for the action that generates revenue.

Ad set level

This is where you define your audience, budget, schedule, and placement. You can have multiple ad sets within one campaign, each targeting a different audience or using a different budget.

Ad level

This is the actual creative — the image or video, headline, primary text, description, and call-to-action button. You can have multiple ads within each ad set, which is how you run A/B tests.

Audience Targeting

Facebook's targeting is its superpower. You can reach people based on demographics, interests, behaviours, and connections — or you can let Facebook find the right audience for you.

Core audiences (interest-based)

You define the audience manually using criteria like location (country, city, or radius around a point), age range, gender, language, interests (hobbies, brands, topics they engage with), and behaviours (purchase history, device usage, travel patterns).

Start broad, then narrow. A common mistake is creating an audience that is too small by stacking too many interest filters. Facebook's algorithm needs a large enough pool to find the best people within it. For most campaigns, an audience of 500,000 to 5 million is a good starting point.

Layer interests logically. If you sell running shoes, target people interested in running AND fitness — not running OR cooking OR travel. Narrowing ensures relevance.

Custom audiences (your data)

Custom audiences are built from people who have already interacted with your business. You can create custom audiences from website visitors (requires the Meta Pixel installed on your site), email lists (upload your customer or subscriber list), app users, people who engaged with your Facebook or Instagram content, and video viewers.

Custom audiences are typically your highest-converting audiences because these people already know your brand. Retargeting website visitors who did not convert is one of the most profitable advertising strategies available.

Lookalike audiences (Facebook finds similar people)

A lookalike audience tells Facebook to find new people who are similar to an existing audience you provide. Upload your customer list or use your website visitors as the source, and Facebook finds people with similar characteristics, interests, and behaviours.

Lookalike audiences are the best prospecting tool on Facebook. A 1% lookalike (the top 1% most similar to your source) is the most precise. A 5% or 10% lookalike is broader but still targeted. Start with 1% and expand if results are strong.

Advantage+ audience (AI-driven)

Facebook's newest targeting option lets the algorithm find the best audience with minimal manual input. You provide broad targeting suggestions, and Facebook's AI uses its data to find the people most likely to convert.

This works surprisingly well for accounts with existing conversion data. If your pixel has tracked hundreds of purchases or leads, the algorithm has enough data to find similar people effectively. For new accounts with no data, manual targeting is usually more reliable.

Ad Creative: What Makes People Click

Your ad creative is the single biggest factor in campaign performance. The same audience with the same budget will produce dramatically different results depending on whether the creative is compelling or forgettable.

The thumb-stop test

People scroll their Facebook and Instagram feeds at speed. Your ad has roughly one second to stop the scroll. If it does not pass the thumb-stop test — making someone pause mid-scroll — nothing else matters because they will never read your headline or click your link.

What stops thumbs: bold colours and contrast, human faces showing emotion, movement in the first frame of video, text overlays with a surprising or provocative statement, before-and-after visuals, and anything that disrupts the visual pattern of the surrounding feed.

Image ads

A single static image with your message. Best for simple, clear offers. Use a high-quality image that tells the story at a glance. Minimal text on the image — Facebook historically penalised ads with too much text, and while the strict 20% rule is gone, clean images still outperform cluttered ones.

Video ads

Video consistently outperforms images in engagement and conversion rates. Keep videos under 30 seconds for feed placements. Hook the viewer in the first 3 seconds. Add captions because 85% of Facebook video is watched without sound. Demonstrate your product or service in action rather than just talking about it.

Carousel ads

Multiple images or videos in a swipeable format. Each card can have its own headline, description, and link. Carousels are excellent for showcasing multiple products, telling a sequential story, or presenting multiple benefits of a single offer. They tend to have higher click-through rates than single images because the swipe mechanism creates engagement.

The ad copy formula

Your primary text (the text above the image or video) should follow a proven structure.

Hook. The first line must grab attention. Ask a question, state a bold claim, or identify a pain point the reader recognises. "Struggling to get your website on page 1 of Google?" is stronger than "We offer SEO services."

Body. Explain what you are offering and why it matters to the reader. Focus on benefits, not features. "Get more customers from Google without spending a fortune on ads" is a benefit. "We provide comprehensive SEO services" is a feature.

Social proof. Include a result, testimonial, or statistic that builds credibility. "We helped 50+ UK businesses double their organic traffic" is more convincing than "We are a great marketing agency."

Call to action. Tell them exactly what to do next. "Click below to get your free SEO audit" is clear and specific. "Learn more" is weak but acceptable.

Keep the primary text to 3 to 5 short lines for most placements. Longer copy can work for cold audiences who need more convincing, but test both.

Budget and Bidding

How much to spend

There is no universal minimum budget, but practical reality sets a floor. With less than £5 per day, Facebook does not have enough budget to test effectively. The algorithm needs data to optimise, and tiny budgets produce too few events for the learning phase to complete.

Recommended starting budget: £10 to £20 per day per ad set. This gives the algorithm enough data to optimise within 3 to 7 days.

Testing budget: When launching a new campaign, allocate enough budget to generate at least 50 conversion events (clicks, leads, or purchases depending on your objective) per ad set per week. This is the threshold Facebook needs to exit the learning phase and optimise effectively.

Campaign budget optimisation vs ad set budgets

Campaign Budget Optimisation (CBO) lets Facebook distribute your total campaign budget across ad sets automatically, giving more budget to ad sets that perform better.

Ad set budgets let you control exactly how much each ad set spends.

For beginners, ad set budgets give you more control and clearer data. Once you have experience and multiple performing ad sets, CBO can improve efficiency by automatically reallocating budget.

Bidding strategies

Lowest cost (default). Facebook gets you as many results as possible within your budget. Best for most advertisers.

Cost cap. You set a maximum cost per result. Facebook tries to stay at or below this amount. Useful when you know your target cost per lead or purchase.

Bid cap. You set a maximum bid for each auction. This gives the most control but can limit delivery if your cap is too low.

Start with lowest cost. Move to cost cap once you know what a profitable cost per result looks like for your business.

The Meta Pixel: Essential Tracking

The Meta Pixel is a piece of code installed on your website that tracks what visitors do after clicking your ad. Without it, Facebook has no way to measure conversions and no data to optimise your campaigns.

Install the pixel before running any campaign. Go to Events Manager in your Meta Business Suite, create a pixel, and add the code to your website header (same process as Google Analytics — paste in Ghost's Code Injection or your platform's equivalent).

The pixel tracks standard events like page views, add to cart, initiate checkout, purchase, lead, and complete registration. You can also create custom events for specific actions.

Once the pixel is collecting data, Facebook uses it to find more people likely to convert, retarget visitors who did not convert, measure your return on ad spend accurately, and build lookalike audiences based on actual converters.

Testing: How to Find What Works

No one — not the best marketer in the world — knows which ad will perform best before testing. Testing is not optional. It is the core skill of successful Facebook advertising.

What to test

Creative first. The ad image or video has the biggest impact on performance. Test 3 to 5 different creative concepts per ad set. Different images, different video styles, different angles.

Headlines second. Once you find creative that works, test different headlines against it. A benefit-driven headline versus a question versus a social-proof headline.

Audiences third. Test different audience definitions — interest targeting versus lookalikes versus broad targeting. Keep the creative constant so you isolate the audience variable.

One variable at a time. If you change the image, headline, and audience simultaneously, you have no idea which change caused the result. Change one thing, measure the impact, then change the next.

How to evaluate tests

Give each test enough budget and time to produce statistically meaningful data. A minimum of 1,000 impressions and 20 to 50 results (clicks or conversions) per variant before drawing conclusions.

Compare variants on cost per result, not vanity metrics. An ad with a higher click-through rate but higher cost per conversion is worse than an ad with lower CTR but lower cost per conversion.

Kill underperformers quickly. If an ad is clearly losing after sufficient data, stop it and redirect the budget to the winner. Do not let emotional attachment to a creative you liked keep a bad ad running.

Campaign Optimisation

Once your campaign is live and generating data, optimisation is the process of improving results over time.

Monitor the learning phase

New campaigns enter a learning phase while Facebook's algorithm figures out the best audience and placement. During this phase, performance is unstable and costs may be higher. The learning phase ends when an ad set reaches approximately 50 optimisation events per week. Do not make significant changes during the learning phase — it resets the process.

Scale what works

When you find an ad set producing results at an acceptable cost, scale it gradually. Increase the budget by 20 to 30% every 3 to 5 days. Large sudden budget increases can reset the learning phase and destabilise performance.

Refresh creative regularly

Even the best ad eventually fatigues. The same audience seeing the same creative repeatedly leads to declining performance — lower click-through rates, higher costs. Monitor frequency (average number of times each person has seen your ad). When frequency exceeds 3 to 4, it is time for new creative.

Retarget warm audiences

Set up a retargeting campaign for website visitors who did not convert. These people have already shown interest — they clicked your ad or visited your site. A retargeting ad with a different angle, additional social proof, or a special offer often converts them on the second or third exposure.

Retargeting typically produces the lowest cost per conversion of any campaign type. Even a small daily budget (£5 to £10) on retargeting can produce outsized results.

Common Facebook Advertising Mistakes

Choosing the wrong objective. Optimising for engagement when you want sales means Facebook shows your ad to people who like posts, not people who buy things.

Too small an audience. Audiences under 100,000 give the algorithm too little room to optimise. Go broader and trust the algorithm to find the right people within the larger pool.

Only testing one ad. Running a single ad and judging the platform based on its performance is like eating at one restaurant and judging an entire city's food scene. Test multiple creatives.

Changing things too quickly. Making changes every day prevents the algorithm from learning. Let campaigns run for at least 5 to 7 days before making optimisation decisions.

No pixel installed. Without the pixel, you are flying blind. You cannot track conversions, build retargeting audiences, or create lookalikes.

Ignoring mobile. Over 90% of Facebook traffic is mobile. If your ad creative or landing page does not work perfectly on a phone, you are losing the vast majority of your potential customers.

Sending traffic to your homepage. Always send ad traffic to a specific landing page relevant to the ad. Someone who clicked an ad about SEO services should land on your SEO services page, not your homepage where they have to find it themselves.

Measuring Success

The metrics that matter depend on your objective.

For traffic campaigns: Cost per click (CPC), click-through rate (CTR), and landing page views (not just link clicks — landing page views confirm the page actually loaded).

For lead generation: Cost per lead, lead quality (do they convert into customers?), and lead volume.

For sales: Cost per acquisition (CPA), return on ad spend (ROAS), and average order value.

For all campaigns: Frequency (how often the same person sees your ad), relevance score (Facebook's assessment of ad quality), and cost trends over time.

Track these in Meta Ads Manager. For a complete picture, combine with Google Analytics 4 data to see what people do after they arrive on your site.

Facebook Advertising for Small Budgets

You do not need thousands of pounds to advertise on Facebook. A £10 per day budget (£300 per month) is enough to test, learn, and generate results for a small business.

Start with one campaign, one ad set, and three to five ad variants. Run for two weeks. Analyse the results. Kill underperformers, scale what works, and launch a new test. This iterative process produces steady improvement regardless of budget size.

The businesses that succeed with small budgets are the ones that test relentlessly and make data-driven decisions. The businesses that fail are the ones that create one ad, run it for a month without checking, and wonder why it did not work.

Need Help with Facebook Campaigns?

Running profitable Facebook campaigns requires ongoing testing, creative production, and optimisation. If you would rather focus on running your business while experts handle your paid media, Kael Tripton offers performance marketing services for businesses of all sizes. From campaign setup to creative strategy to monthly optimisation, we help businesses generate leads and sales from paid social media.

Final Thoughts

Facebook advertising is not magic. It is a system — set up correctly, fed with good creative, measured accurately, and optimised consistently. The businesses that treat it as a test-and-learn process generate real results. The businesses that set up one campaign and hope for the best waste their money.

Start small. Test everything. Scale what works. And never stop testing — because the creative that works today will stop working eventually, and the next winner is always one test away.

Last updated: March 2026. Meta's advertising platform and features evolve regularly. The principles in this guide reflect current best practice.