How to run
Meta ads.

Nobody opens Instagram to buy your product. Your ad has to earn attention first, which is why creative matters far more here than in search.

It also means demand generation. Meta works well for products people did not know existed, and less well for urgent, specific needs — someone with a burst pipe is searching Google, not scrolling.

The Pixel tracks what happens after the click. Without it, Meta cannot optimise toward buyers and you cannot tell what worked.

Browser tracking now misses a significant share of events because of privacy controls and blockers. The Conversions API sends events server-side and recovers much of that gap. Running both, properly deduplicated, is the current standard.

Verify your domain and configure your priority events, or your tracking will be limited regardless of setup.

Meta optimises hard toward the objective you choose. Choose traffic and it finds people who click but do not buy — genuinely, that is what you asked for.

For leads or sales, choose the conversion objective and let it optimise toward the event you care about, even though the reported cost looks higher. The cheaper number on a traffic campaign is usually cheaper worthless clicks.

Detailed interest targeting matters much less than it used to. Meta's delivery system finds buyers from creative signals and conversion data, often better than manual targeting does.

Broad audiences with strong creative now routinely beat narrowly stacked interests. Over-narrowing raises costs and starves the system of the data it needs to learn.

Where targeting still earns its place: custom audiences from your customer list or site visitors, and lookalikes built from your best customers rather than all of them.

Test genuinely different concepts, not colour variations. Different hooks, different formats, different angles.

The first two seconds decide everything on video. Lead with the problem or the result, never with a logo animation.

Design for sound-off with captions, and shoot vertical for feeds and stories.

Creative fatigues. When frequency climbs and click-through falls, the audience has seen it enough — that is a creative problem, not a bidding problem.

Change one thing at a time. Testing new creative, a new audience and a new landing page simultaneously tells you nothing about which one mattered.

Give the system time to exit the learning phase before judging. Editing campaigns constantly resets learning and guarantees poor performance.

Do not judge on a handful of conversions. Small numbers swing wildly and look like signal when they are noise.