Your data and your ads decide who the platform finds.
On Meta and TikTok, broad targeting usually wins, and we run it that way. That moves the work to two inputs the algorithm can’t supply for itself: what counts as a customer, and how many kinds of buyer your ads speak to. Most brands and agencies treat the first as a tracking setup and the second as a production schedule. We treat both as the targeting.
How we run each platform depends on how much it already knows about your buyer.
Meta and TikTok: go wide and teach the platform the customer.
Both have the signal to find your buyer without an interest stack. On the ads side, every distinct concept is a different buyer the platform can find, so we brief for range and refresh as creative fatigues. We keep audiences open and put our time into what they learn from: conversion signal, customer values, exclusions and product feeds built for catalog ads. Pinterest and Snap run the same way.
More on feeds
LinkedIn and Reddit: steer harder.
On LinkedIn, we target the whole buying committee, then split it by seat. The CTO gets the ad about security and integration, the CFO gets payback, and the product team gets the workflow it will live in. LinkedIn’s firmographic targeting makes that possible, and its algorithm can’t make that call on its own yet. We also send qualified-lead stages from your CRM back to LinkedIn, so it optimizes toward leads your sales team accepts. On Reddit, we show up in the communities where your buyers already compare options.
We define the customer.
Conversion events, values and exclusions that pay for new customers and qualified pipeline instead of easy clicks.
We widen who the ads speak to.
Founder-led spots, demos and creator ads, briefed for buyers beyond your core ICP.
We steer where targeting still works.
Buying-committee seats on LinkedIn and communities on Reddit, each with its own message.
We measure what platforms miss.
Lift tests across the funnel, with the learnings carried into streaming TV and audio.
Streaming TV & Audio
Broad targeting will find your next customer. Someone still has to tell it what one looks like.
With no audience settings left to tune, your ads decide who you reach. Personas used to be a targeting spec. On Meta and TikTok, they work as a creative brief. A portable power station sells to a camper, a homeowner bracing for an outage and a remote worker who wants out of the house, and each stops scrolling only for an ad about their own situation. Ten variants of one idea find one set of buyers. Ten ideas find ten. So our strategists map who else buys and why, check the account for ads that say the same thing, and brief what’s missing. See how we build and test it →
Brands can still aim broad targeting through customer definition and who your ads speak to.
What the algorithm learns from
We define what a customer is before we launch a campaign.
Left alone, a broad audience is very effective at finding whoever converts most cost-effectively. The challenge is that’s often repeat buyers, discount shoppers, and form fills that never close. We set the conversion event, what each one is worth and who gets excluded, so the platform hunts for the customer you’d pay to acquire.
