Paid Social

Paid social is the work of getting in front of a defined audience in their feed — interest-led, behaviour-led, lookalike-driven — at the moment they're scrolling. Different specialism to paid search (which is intent-led). We run paid social through the AOS across Meta, LinkedIn, TikTok, X and YouTube — same continuous-optimisation loop, calibrated to social-platform creative velocity and audience economics.

What we run

The standard programme covers:

• Meta (Facebook + Instagram) — feed, stories, reels, in-stream, lead-gen ads. The dominant channel for most B2C and a growing channel for B2B brand work.
• LinkedIn — sponsored content, message ads, lead-gen forms. The dominant channel for most B2B paid social.
• TikTok — in-feed and TopView formats. Increasingly significant for younger-skewing B2C and emerging B2B (the platform's professional ad inventory has grown materially).
• X — promoted posts and ads. Niche but useful for thought-leadership amplification and crisis-window targeting.
• YouTube paid — bumpers, in-stream, discovery. Often grouped with paid search (Google) but the creative and audience model is much closer to social.

How the work runs

Why creative velocity matters more than creative perfection

On social platforms, creative is the variable that moves performance most. Audiences saturate fast — 4-6 weeks per variant in B2C, 6-10 weeks in B2B. The agencies that win on social aren't the ones producing the single best creative; they're the ones producing the largest VOLUME of good-enough variants and letting the algorithm find the winners.

Most fixed-team operations cap out at 5-10 fresh variants per channel per month. The AOS handles 30-50 for B2C and 15-25 for B2B — variant production runs through the platform under brand and policy guardrails set by senior creative direction.

What stays human: campaign concept, brand voice, the big visual ideas, escalation when the platform surfaces something unusual. What runs through the platform: production-scale variant testing, headline/copy iteration, format adaptation across channels, performance-driven retirement of fading creative.

First-party audience strategy

Third-party cookies are degrading; iOS App Tracking Transparency has reduced the audience-signal Meta and other platforms can build from third-party data. The durable answer is first-party audiences:

• CRM-exported audiences (customer lists, lifecycle stage segments, lookalike seeds) uploaded server-side via hashed identifier matching
• Site-visitor audiences captured via server-side pixel + first-party cookies (resilient to ITP)
• Engagement audiences (video viewers, ad clickers, page engagers) built natively in each platform
• Email-engaged audiences synced from your email platform (engaged subscribers, recent purchasers)

These audiences perform materially better than third-party data in 2026 — they're durable, they're owned by you, and they survive platform privacy changes. We build the architecture to maintain them as part of the engagement.

How performance is measured

Same metric framework as paid search — channel-level performance for reallocation decisions, blended ROAS for board-level reporting, CRM-attributed revenue for the truth signal.

AI-powered vs traditional paid social delivery

What it costs

Same fee schedule as paid search — sliding-scale percentage of monthly media spend. Most programmes run paid search and paid social together; the combined media spend tier determines the fee percentage.

Where this service wins

• B2C brands with a deep creative library and the appetite for high variant velocity
• B2B programmes on LinkedIn with closed-loop CRM signal — the gap between platform CPL and CRM-qualified-CPL is usually 3-5x; closed-loop fixes it
• Brands hitting a creative-supply ceiling — fixed teams can't sustain 30-50 fresh variants/channel/month; the platform can
• Multi-channel social programmes (3+ channels) where the rebalancing capability across audiences and platforms compounds

Where it doesn't fit

• Single-channel programmes (e.g. only Meta) — the value comes from multi-channel reallocation; single-channel cuts off most of the upside
• Brands without a creative pipeline that can keep up — the platform produces variants under brand rules, but it doesn't invent the brand. Need a creative source upstream
• Heavily regulated sectors where every variant needs compliance sign-off — the velocity benefit is capped by approval throughput

Read deeper on this

• Paid Search & Display — the intent-capture half of the performance media stack.
• Conversion tracking foundations for AI-led marketing — the server-side tracking work paid social depends on (Meta CAPI, LinkedIn Conversions API, TikTok Events API).
• Why your CAC is climbing — and what to do about it — the privacy + signal degradation forces hitting paid social hardest.

Sources and further reading

• Meta — Conversions API documentation — Meta's official guide to server-side conversion tracking.
• Apple — App Tracking Transparency — Apple's developer documentation on the ATT framework.
• LinkedIn — Conversions API documentation — LinkedIn's official guide to server-side event tracking.