- Campaign-based marketing is becoming inefficient
- Algorithms now prioritise creative quality over audience targeting
- Modern paid social requires high-volume creative production
- Human creators build trust; AI scales variations
- Always-on creator and advertising strategies outperform campaign cycles
The most costly habit in marketing right now isn’t bad creative. It relates to the campaign itself.
A campaign is a fiction. A neat little container with a start date, an end date, an audience or two, and a budget that ramps up and back down again. It’s helpful for organising PowerPoint slides. Helpful for the agency invoicing cycle.
Helpful for the marketer who needs to point at a thing and say, “That’s what we did in Q2”. Almost everything else about it is now actively working against you.
The campaign forces you to do three things that the algorithm, the platforms and your cost per acquisition (CPA) all punish:
-
You stop things that are working so that you can start new things
-
You commit to audience segments limited by your brain’s ability to choose between dropdowns, or how many coloured boxes fit on one slide
-
You hire creators for the duration of the brief, then fire them, and then rehire them six weeks later wondering why they’re harder to reach.
Meta launched Andromeda. TikTok built Symphony. Both have rebuilt the engine underneath their ad auctions. The algorithm now decides who sees your ad based on what your ad actually says, not which audience box you ticked.
If your creative is mediocre, your ‘high-intent female 25-34 Pinterest saver’ segment won’t be keen to save it. If your creative is good, the segment is mostly in the way.
The insatiable creative appetite
But the machine has an appetite. A voracious one. The platforms’ own cadence guidance, combined with industry consensus, puts the number at 60 to 80 new creatives a month to keep paid social properly fed at scale.
This is two to three new ads each working day, which must be sufficiently conceptually different that the algorithm has something to choose between. Same blue background with a different headline doesn’t count.
This volume is not something your in-house team can deliver. It’s not something your creative agency can deliver either; not with their kickoff meetings, amend rounds, and per-hour timesheets. The only group that can produce content at a velocity like this is a distributed network of creators, who are already making content for a living.
Why AI won’t solve the trust deficit
The obvious objection: what about AI? Just generate the ads. Done.
Here is where most performance teams are about to learn an expensive lesson. AI can produce content at infinite velocity. The question is not whether you can make it. The question is whether anyone wants to watch it. Remember, content is infinite. Trust is scarce.
The algorithm knows the difference; and so does the viewer scrolling past your perfectly rendered AI avatar, and reading a perfectly written script with the dead eyes of a creature that has never had a bad day.
The right division of labour is this: human creators make the hero content. The original idea, the original face, the voice the audience has chosen to trust. AI handles the iterations. New hooks. Different durations. Alternative CTAs.
Captioned, reframed, recut. The creator makes one piece of content. AI helps you turn it into the 15 that the algorithm likes and requires. That’s how you hit 60 to 80 ads a month without going broke, or without watching your hook rate collapse. AI as a tool for iteration is great. AI as a substitute for human trust is slop.
The cost of agency denial
Meta has been saying most of this for as long as two years. The industry has put its fingers in its ears and hummed loudly. Not because they don’t get it. But because their business model can’t get it.
A big chunk of the media agency world (and most large advertisers) have built their entire competitive moats on third-party data, DMPs, and the comforting story that they can find audiences better than any platform can.
Admitting Andromeda works is admitting that their main differentiator has just evaporated. So they don’t admit it. They keep selling segment plans. They keep buying inflated CPMs on narrowed audiences. They keep blaming the creative when the campaign underperforms. It’s a beautiful, fully internally consistent piece of denial.
Case study: breaking the campaign anatomy
Meanwhile, the marketers paying attention keep getting cheaper customers.
We had a front-row seat. Mukuru, UK app launch. Original setup was textbook: hyper-targeted audiences, layered exclusions, lookalikes off existing user data. The classic campaign anatomy. CPI was punishing. Growth stalled.
We killed it. Pulled almost every audience filter. Let the algorithm run broad. And then we did the part most teams skip: we took each original creator video and built three to four variations from them. New hooks. New first frames.
Different durations. Different CTAs. Some started with the punchline. Some buried the brand reveal. Some had no voiceover – at all. The point was to give Meta and TikTok a big enough variety of conceptually distinct ads, so that the algorithm had something to work with.
CPI fell by 89.5 percent. Installs went up 290 percent. Same product, same market, same budget. Different operating philosophy.
Shifting the metrics to attention
While we were at it, we added two metrics to the dashboard. Hook rate (2-second view rate on TikTok, 3-seconds on Meta) and hold rate (viewers who stay past 6 seconds). Yes, engagement is still useful: shares, saves, comments still tell you something about content quality and audience response.
But engagement was never the right proxy for attention on its own. Hook rate tells you if the ad earned attention. Hold rate tells you if it kept it. The three metrics together replace the guesswork we used to call “engagement is up so it must be working”
So what should D2C performance leads, growth marketers, and media buyers actually do differently? Two things. Both boring. Both the same shape
1. Stop closing creator contracts when campaigns end
Your creator roster is the asset. So is the content library it builds. Each month that a creator stays under contract, you keep usage rights on everything they’ve already made for you – and they keep adding to the library. Drop the contract and you lose access to all of it, then start from zero with someone new on the next brief. Most marketers don’t realise they’re throwing an important asset class away every quarter.
Briefs are just prompts you feed the roster. Measure creators, not campaigns. If a creator’s content keeps converting after the launch window, why fire them? If it stops, why wait for the next quarterly cycle to swap them out? Hire creators the way you hire growth marketers: continuously, with a measured approach, retained, or offboarded. Always-on. Not start-stop.
2. Stop turning off winning ads. Stop narrowing audiences
Same logic, different layer. The algorithm punishes stop-start. Each reset is a new learning phase, and each learning phase is a tax. Same with targeting: each filter you add raises CPMs and shrinks the canvas that the algorithm has to paint on. The two best things you can do for your CPA are also the most boring ones: leave the winners running, and let the algorithm see everyone.
Both actions encompass the same idea. Stop the stopping. The campaign model trained marketers to think of starting and finishing as their job. The job now, is keeping the system running and feeding it well.
Overcoming the gravity of the 2018 game
Here’s the honest part: none of this is really news. You’ve read versions of it. You’ve nodded at conferences. You may even have a slide about it in a deck somewhere. So why doesn’t it happen?
Because the campaign model has gravity. It fits the budget calendar. It fits the agency contract. It fits how the legal team approves spend. It fits how the CEO asks: “What did we ship this quarter?” It keeps a lot of people busy doing predictable, expensive work, and constantly reinventing what they already know worked last time.
And it keeps a lot of agencies in business – doing the same work, for the same clients, on the same schedule they’ve used since 2015.
The platforms have changed the rules. Most marketers are still playing the 2018 game, and paying more for the privilege each quarter. The campaign mindset is what’s keeping it this way.
The message? Stop running campaigns. And start running creators.
Pierre Cassuto is global chief marketing officer (CMO) at Humanz.
