Ad platforms now have vast amounts of data about what people respond to. Automated products use that data to make decisions faster than humans can. For many advertisers, this means less manual setup and often better results. For others, it means losing the ability to control where money goes. The Academy's glossary explains Advantage+.
What automation controls
| Decision | Manual campaigns | Automated campaigns |
|---|---|---|
| Audience | You define it | Platform finds people, sometimes with your suggestions |
| Placements | You choose or allow automatic | Usually automatic |
| Creative combinations | You set each ad | Platform mixes assets |
| Budget distribution | You set by ad set | Platform shifts spend |
| Optimisation | You choose the event | You choose the event |
Types of automation
- Automated audiences: broad targeting guided by conversion data.
- Automated placements: ads shown wherever they perform best.
- Dynamic or flexible creative: the platform tests combinations of images, videos and text.
- Campaign budget optimisation: spend moves between ad sets.
- Fully automated campaign types: such as Advantage+ shopping campaigns on Meta or smart performance options on TikTok.
When automation works well
- Strong tracking with reliable conversion data.
- Enough conversions for the system to learn.
- Plenty of creative variety to test.
- Clear objective such as purchases.
- Broad market where many people could buy.
The Academy's guide to the Conversions API explains how to strengthen tracking.
When manual control helps
- Very small budgets with too few conversions.
- Niche audiences, such as specific professions or locations.
- Strict brand safety or placement requirements.
- Testing a specific hypothesis, where you need clean comparisons.
- Separating new and existing customers precisely.
The trade-offs
Automation can improve results and save time, but it reduces transparency. You may not see exactly who saw ads or why decisions were made. Automated campaigns can also spend heavily on retargeting existing customers unless you set controls for existing customer budgets where available.
Keeping oversight
- Define existing customers so the platform can report and limit spend on them where the option exists.
- Feed varied creative regularly.
- Compare against your own sales data, not just platform reports.
- Run holdout or incrementality tests periodically.
- Review placements and search terms where reports allow.
The Academy's guide to incrementality testing explains how to check real impact.
Creative is the main lever
When the platform controls audiences and placements, creative becomes the main thing you influence. Provide a range of formats, angles and hooks. TikTok's documented organic signals are completion rate and rewatches, shares, comments, likes, interest signals from watch behaviour✓, and automated systems favour creative that holds attention. The Academy's guide to creative testing helps build a strong pipeline.
Testing automation against manual
Run an automated campaign alongside a well-structured manual campaign with similar budgets for a few weeks. Compare cost per result, volume and, ideally, incremental sales. Many advertisers end up using both: automation for scale, manual for tests and niche audiences. Whatever you choose, revisit the comparison every few months, because platforms update their automated products frequently and results that held last quarter may not hold now.
Automation and compliance
Automated targeting does not remove legal responsibilities. Ads must still follow UK advertising rules, and restricted products, such as alcohol or gambling, need appropriate audience controls. Check that automated placements and audiences fit your obligations, and add exclusions or age limits where the platform allows. Responsibility for compliant advertising stays with the advertiser, whatever the platform decides on your behalf. The Academy's glossary explains the CAP Code.
Automation on different platforms
Meta offers Advantage+ shopping campaigns and Advantage+ audience options. TikTok offers smart performance campaigns and automated creative tools. Pinterest and Snapchat offer automated bidding and targeting features too. LinkedIn automation is more limited, reflecting its focus on precise professional targeting. Names and features change frequently, so check each platform's current options, and read the Academy's public changelog for recent changes to platform facts.
Reporting on automated campaigns
Because automated campaigns combine audiences, reporting can blur new and returning customers. Where the platform allows, break results down by new versus existing customers and by placement. Compare these against your own order data to see whether automation is finding genuinely new buyers or mostly reaching people who already knew you.
A worked example
An online clothing retailer runs manual campaigns with interest targeting and an Advantage+ shopping campaign side by side. The automated campaign achieves lower cost per purchase but spends heavily on existing customers. The retailer defines its existing customers and limits the share of budget they receive, adds more creative variety and keeps a manual prospecting campaign for testing. Overall new customer acquisition improves.
Feeding automation good signals
Automated systems learn from the events you send. Optimise for the event that best reflects real value, send purchase values where possible and keep tracking healthy. If your business closes sales offline, send qualified leads or completed sales back through the Conversions API so automation learns which people become customers rather than which people simply fill in forms.
Common mistakes
- Using automation with broken tracking.
- Too little creative for the system to test.
- Trusting platform reports without checking sales.
- Letting automation spend mostly on existing customers.
- Abandoning manual testing entirely.
Learning automated advertising
The Academy's paid social module in the third level covers automated campaigns, structures and testing, and the AI and automation at scale module in the leader level covers automation more broadly. The free assessment shows where you would start.
