Meta's ad platform has changed significantly in the last two years. Most of the Meta advertising advice you'll find online is written for a platform that no longer exists. Here's what's actually working now.
← Back to InsightsMeta has progressively removed manual targeting options and pushed advertisers toward Advantage+ campaigns. This isn't a trend. It's the direction the platform has committed to. Broad audiences, AI-optimised delivery, algorithm-controlled spend allocation. Fighting it is counterproductive. The better move is to use Advantage+ intelligently: give the algorithm the inputs it needs to optimise correctly.
That means conversion events properly configured. Not "link clicks" or "landing page views," but actual business outcomes: form submissions, purchase events, qualified leads. It means high-quality creative in multiple formats, so the algorithm has something to test. And it means a clear conversion objective that reflects what you actually want, not what's easiest to set up.
The founders who succeed with Meta now are those who understand they're directing the algorithm, not replacing it. Your job is to give it accurate signals, strong creative and a realistic conversion window. Its job is to find the people most likely to act. That division of labour, when set up correctly, is genuinely effective. When set up lazily, it burns budget on entirely the wrong audience.
As audience targeting has become less precise (a combination of Apple's ATT changes in 2021, Meta progressively removing detailed targeting options, and Advantage+ broadening reach by default) the creative itself has become the primary targeting mechanism. The right creative self-selects its audience. Someone who isn't a potential customer scrolls past. Someone who is stops.
This is a significant shift from three or four years ago, when you could build a tightly defined audience of 300,000 people and be reasonably confident your ads were reaching exactly that profile. That precision is largely gone. What replaced it is creative quality and creative variety. A single static image running for six months no longer works. It fatigues fast, and a fatigued ad in a broad Advantage+ campaign is a budget drain. Video, UGC-style content and carousel formats that tell a story outperform static in most verticals, and consistently so.
The practical implication: treat your creative brief with as much rigour as your targeting strategy. The hook in the first two seconds of a video is more important than most targeting decisions you'll make. If the creative doesn't stop the scroll, nothing else in the campaign matters.
A three-layer funnel remains the most reliable structure for Meta: awareness at the top (broad, Advantage+ or interest-based), consideration in the middle (video view retargeting, engagement retargeting), conversion at the bottom (website retargeting, lead form campaigns). The percentages shift by sector, but the architecture holds.
The mistake most small businesses make is running conversion campaigns only, targeting cold audiences and asking for the enquiry immediately. For low-cost, low-consideration purchases, this can work. For B2B services and considered-purchase B2C products, it almost never does. Someone who has never heard of your business needs multiple touchpoints before they enquire. Running a conversion campaign cold is asking for the sale before the introduction. The funnel exists to build those touchpoints in a measurable, controllable way.
If your retargeting pool is too small (under a few thousand people) you won't have the data volumes to run effective retargeting. In that case, consolidate budget into the awareness layer and let it build before moving down-funnel. Patience here isn't indifference to results; it's the structurally correct approach.
Meta's native Lead Gen forms, which collect information inside the platform without sending the visitor to your website, produce higher lead volumes at lower cost per lead. They also produce lower-quality leads. People who haven't visited your website, don't have context on your offer and who sometimes forget they submitted the form by the time you follow up. The data is pre-filled from their Facebook profile, which means almost no friction and almost no intent signal beyond a tap.
Landing pages produce fewer leads at higher cost but higher intent. Someone who clicked through to your website, read your offer and completed a form is demonstrably more interested than someone who tapped a pre-filled native form in their feed.
The right answer depends entirely on your follow-up speed and your sales process. If you can call within five minutes of a Lead Gen form submission, they work well. The enquiry is warm enough to convert. If your follow-up is 24 hours or more, send them to a landing page first. The intent created by the visit will make that delayed call land significantly better.
Boosting posts. Running the same creative for more than three to four weeks without refresh. Targeting audiences under 50,000 people: the algorithm doesn't have sufficient data to optimise, and you'll spend the entire campaign in or near the learning phase. Daily budgets under £10 for the same reason: not enough spend for the algorithm to exit learning and start performing.
Optimising for link clicks instead of conversions. This one persists because conversion events can feel complex to set up. But optimising for clicks tells the algorithm to find people who click things, not people who buy things. Those audiences overlap less than you'd expect. Set up your Pixel events, configure your conversion objective properly and let the algorithm optimise for the outcome that actually matters to the business.
All of these are common. All of them consistently underperform. If your current Meta Ads setup involves any of the above, fixing them is the most valuable thing you can do before you increase spend.
Meta's attribution is optimistic. The default attribution window includes view-through conversions, counting a conversion if someone saw your ad (not clicked it) and then converted within a window that can extend to seven days. On a broad Advantage+ campaign reaching tens of thousands of people, this significantly overstates Meta's contribution. Someone who saw your ad in their feed three days before converting via a Google search gets attributed to Meta.
The honest way to measure Meta Ads performance is with a blended approach. Use platform metrics for directional signals: CPM, CTR and cost per result tell you whether creative is working and whether the algorithm is finding the right people. Use UTM parameters tracked in GA4 for last-click attribution to understand actual traffic quality and on-site behaviour. And use a question on your enquiry form ("how did you hear about us?") to capture assisted attribution that neither platform will give you.
No single attribution model is correct. A combination of three (platform-reported, GA4 last-click and self-reported) tells you more than any one alone. If all three point in the same direction, you have a meaningful signal. If platform numbers are strong but GA4 shows nothing, you have an attribution inflation problem worth investigating before you scale spend.
Mode runs Meta Ads as part of the Scale Sprint, alongside Google, LinkedIn and organic channels, with shared audience data and coordinated creative.
TALK ABOUT META ADS