Home/Playbook/Paid social advertising for pizza locations

Channels

Paid social: geo-fenced, appetite-led, and easy to overspend

Running Meta and TikTok for a delivery business: polygon targeting, creative that actually stops the scroll, and the three campaigns worth running.

Channels4 sections3 questions answered
Channels illustration

Paid social is the most accessible paid channel for a pizza location and the one where money disappears most quietly. The platform will happily spend a budget across an entire metropolitan area on a creative that nobody watched, and report it as reach. Getting value out of it comes down to three things: the geography, the creative, and running only the campaigns that have a job.

Target the polygon, not the radius

Every social platform offers radius targeting and it is wrong for the same reason a mailing circle is wrong. Upload your actual delivery-zone polygon, or approximate it with postcodes, so you stop paying to reach households you cannot serve.

For multi-unit operators, run one ad set per store with its own polygon and its own store-specific creative and ordering link. A single campaign covering all stores produces a lower cost per thousand and a worse cost per order, because the message stops being local.

The three campaigns worth running

Almost everything else is an experiment. These three earn their place.

  1. Acquisition to first order. Cold, polygon-targeted, appetite-led creative, with a fenced first-order offer and a direct ordering link. Judge on cost per first order, never on cost per click.
  2. Retargeting people who abandoned or browsed. Small audience, short window, 24 to 72 hours, and it is the highest-return social spend most shops have.
  3. Customer-list lookalikes for expansion. Upload high-value customers, build a lookalike, constrain it to the polygon. Works when the file is large enough to be meaningful, which usually means several thousand customers.

Creative rules for food

Food advertising on social has a well-established grammar and most local operators ignore it.

  • The product in the first second. Not your logo, not a title card. Cheese pull, oven, the box opening.
  • Vertical video. The feed is a phone. Square is a compromise; horizontal is a waste.
  • Sound-off first. Assume no audio. Text overlays carry the message; sound is a bonus.
  • Real and local beats polished. Footage from your own kitchen consistently outperforms stock food photography in local campaigns, because it is recognisably you.
  • Refresh every three to four weeks. Small local audiences see the same creative repeatedly, and fatigue arrives far faster than it does for national brands.

A small trade area means a small audience, which means creative fatigue in weeks rather than months. Budget for new footage, not just for media.

Judge it on orders, not on engagement

Platform reporting is optimistic by design and last-click by default, and it will confidently claim credit for orders from customers who order every week regardless. Pass conversions back to the platform for optimisation, but judge the channel on incremental order lift measured against a geographic or audience holdout.

Cost per first order is the number that matters for acquisition. Cost per click and cost per thousand impressions are inputs, not results.

Questions

How much should a single location spend on paid social?

Enough to generate a readable result, which in most trade areas means a few hundred dollars a month at minimum. Below that, the platform cannot exit its learning phase and you are paying for noise. If the budget cannot support that, put it into a channel that works at small scale.

Meta or TikTok?

Meta first for almost every pizza location. The targeting is better, the audience skews toward household food decision-makers, and the conversion tooling is more mature. TikTok is worth testing where you have a young trade area and the capacity to produce genuinely native video, which is a real production commitment rather than a repost.

Should I boost posts?

Boosting is a convenience feature with limited targeting and objectives. It is fine for local community content, and it is not a substitute for a structured campaign with a conversion objective and a polygon.

Keep reading

Related chapters

All 46 chapters

This is one chapter of the job.

marketing.pizza runs all of it, every night, across every store you have.