The short version: Facebook ads feel complicated because the platform got harder to run well after Apple’s 2021 privacy changes gutted Meta’s tracking, and Meta responded by pushing everyone toward automated campaigns that hide most of the levers you used to be able to pull. It’s not that you’re bad at this. The tool changed underneath you and nobody sent a memo.
It used to be simpler, and that’s not you misremembering it
I’ve been running Facebook ads since 2011, back when you could set a campaign to “clicks to website,” pick an audience, and get a cost per click under 20p in most UK markets. I’m not being nostalgic for the sake of it. The mechanics were plainer. You picked an objective, you picked people by age and interest, you wrote an ad, you watched the numbers, you adjusted. Three or four decisions, in a straight line.
Now the ad manager asks you to choose between six consolidated objectives (Meta merged eleven down to six in 2022, which they framed as simplification and which mostly just moved the complexity somewhere else), then wants you to pick a campaign type, then an ad set with a daily budget and an audience, then creative, then it quietly runs a “learning phase” you have almost no visibility into. Each layer used to be optional or obvious. None of them are now.
The real turning point was April 2021, and most guides skip it
Apple’s App Tracking Transparency update landed in iOS 14.5 and asked every iPhone user, point blank, whether apps could track them across other apps and websites. Somewhere between 80 and 90 percent of people said no, depending on whose numbers you trust (Meta’s own reporting put the cost at over $10 billion in lost ad revenue for 2022 alone, and that figure is on the record in their SEC filings). That single prompt broke the thing Facebook ads had always been good at: knowing that the person who saw your ad on Tuesday was the same person who bought your product on Thursday.
Meta’s answer was the Conversions API, a server-to-server way of sending Meta the same purchase and lead data your website already collects, bypassing the phone’s own tracking blockers. It works. But setting it up means either learning to use Google Tag Manager and Meta’s Events Manager, or paying a developer, or using a plugin that half does the job. That’s a new skill that didn’t exist as a requirement five years ago, and it’s the single biggest reason accounts that used to convert well in 2019 now feel broken with the exact same budget and the exact same audience.
A real example: the £30-a-day account that never left the learning phase
I worked with a life coach client in 2023 who’d set up her own campaign, budgeted £30 a day, and got almost nothing for three weeks straight. Zero leads some days, one lead at £28 cost per result on others. She’d read that you’re meant to let Facebook “learn,” so she left it alone, which is correct advice for a healthy account and terrible advice for a broken one.
What had happened: she’d edited the ad set five times in ten days, swapping images and tweaking the audience each time engagement dipped. Every edit reset the learning phase, which Meta officially says needs roughly 50 conversion events within a seven-day window before the algorithm has enough data to optimise. Her budget could only produce about 15 to 20 leads a week at her cost per result, so the account never once hit that 50-event threshold. It was permanently stuck relearning from scratch, which is expensive by design, not by accident. We stopped touching the ad set, raised the budget to £45 a day so she’d clear 50 events faster, and left it completely alone for 12 days. Cost per lead dropped from £28 to £11 and stayed there.
That’s the kind of detail that never shows up in a “10 tips for Facebook ads” list, because it requires admitting the platform punishes the exact instinct most small business owners have, which is to fiddle when something isn’t working.
The uncomfortable part nobody quite says out loud
Here it is plainly: Meta’s interface is not built to make your job easier. It’s built to make you spend more, faster, with less friction between “I have a thought” and “I have spent money.” The Advantage+ suite, which now auto-selects your audience, auto-tests your creative, and auto-allocates your budget across placements, is good technology. It’s also designed so the natural path of least resistance is to hand over control and let Meta decide, which almost always means a bigger recommended daily budget than you’d have chosen yourself.
I’m not saying don’t use Advantage+ campaigns. I use them. I’m saying go in knowing the “recommended budget” box that pops up is calibrated to Meta’s revenue, not yours, and treat every suggested figure as a starting negotiation, not an instruction.
Why the interface itself adds to the confusion
Ads Manager has changed its layout at least four times since 2020 that I can count from memory. Column names move. The place where you’d check frequency (how many times the same person has seen your ad, which matters because anything above 3 to 4 in a week usually means ad fatigue is dragging your cost up) has been buried under “customise columns” for years now, rather than sitting on the main view where it used to live. Small, constant reshuffling like this means every time you finally get comfortable, something’s moved, and that alone accounts for a chunk of the “this feels complicated” feeling that has nothing to do with strategy at all.
The same goes for how differently ads behave depending on where they show up. A video ad in the Facebook feed, in Instagram Stories, and in Reels are three different creative shapes, three different attention spans, and three different costs per thousand impressions. UK CPMs in late 2025 were running roughly £8 to £14 in feed placements and noticeably higher, sometimes £18 to £22, for Reels, because that’s where Meta is pushing engagement and where competition for the format is fiercest. Nobody tells you upfront that one campaign might need three separate pieces of creative to perform evenly.
Running a Facebook group alongside your ad account can help here, because organic engagement patterns tell you which creative styles your actual audience responds to before you spend a penny testing it in ads.
Where most of the complexity is self-inflicted (and fixable in an afternoon)
A good chunk of what feels complicated is just disorganisation dressed up as difficulty. I see the same three problems constantly:
- No pixel, or a pixel installed years ago and never checked, quietly firing on the wrong pages or not firing at all. Meta’s own Events Manager has a test tool that takes about five minutes to run and will show you exactly what’s broken.
- Audiences that overlap so heavily that three “different” ad sets are all bidding against each other for the same people, driving your own costs up. Meta’s Audience Overlap tool (buried under Audiences, not Ad Sets, which is why people miss it) shows this in one click.
- Running five objectives at once with a tiny total budget spread across all of them, which starves every single ad set of the data it needs to leave the learning phase. One clear objective with a proper budget behind it will nearly always outperform five thin ones.
None of that requires an agency retainer to fix. It requires an hour, a notebook, and the willingness to turn off things that aren’t working instead of adding more on top.
Where the complexity is real, not self-inflicted
Some of it isn’t your fault and isn’t easily fixable by tidying up. Attribution windows changed after iOS14.5 (Meta now defaults to a 7-day click model instead of the old 28-day click, 1-day view combination), which means the same campaign can look like it’s performing worse purely because it’s being measured differently, not because fewer people bought. If you’re trying to compare this year’s numbers to 2021’s, you’re comparing two different measuring sticks, and almost nobody flags that when they show you a “before and after” case study.
This is also exactly where measuring your content and ad performance matters more than ever, because the platform’s own reported numbers are less reliable than they used to be. Cross-check Meta’s reported conversions against what lands in your CRM or your Shopify dashboard. I’ve seen accounts where Meta reported 40 percent more sales than the business’s own order system showed, simply because of how the attribution window counted a click from someone who bought through a different channel entirely.
Should you learn it yourself or hand it over
Genuine question worth asking honestly: how much is your time worth per hour, and how many hours a week is this eating? If you’re spending six hours a week fighting with an ad account that’s producing £200 worth of leads, the maths rarely works in favour of doing it yourself long term. A freelancer who runs ads for a living, someone who’s watched a hundred accounts go through the same learning-phase problems you’re hitting, will usually find the fix in the time it takes you to read one more blog post about it.
If you’re weighing that up, it’s worth reading through what it takes to freelance in a role like this, partly because it shows you what a competent ads freelancer’s day looks like, and that context makes it far easier to judge a quote when someone sends you one.
For a lot of small businesses now, the honest middle ground is bringing in someone for a short, structured piece of work rather than an open-ended retainer, specifically to get the account structure and the Advantage+ automation set up once, so you’re not paying ongoing fees for something a well-built account can mostly run itself. That’s the kind of focused fix an AI consultant for small business is often brought in for now, because so much of modern Meta advertising is really AI campaign management wearing an ads manager costume.
What I’d do if I were starting an ad account this week
Set one clear objective, most likely leads or sales, not “engagement,” which almost never turns into revenue on its own. Put your full weekly budget behind one or two ad sets rather than spreading it thin across five. Install the Conversions API alongside the pixel from day one rather than retrofitting it later. Leave the ad set alone for a minimum of five to seven days once it’s live, resisting every urge to tweak it on day two because the numbers look quiet. Check frequency and cost per result weekly, not daily, because daily checking is what leads to the panic-edits that reset the learning phase in the first place.
None of that removes the platform’s underlying complexity. It just means you’re not adding your own on top of it, which for most small businesses is where the real money leaks out.
And if you’re doing all of this from a laptop at your kitchen table rather than an office, the setup matters less than people assume; most of what trips people up is account structure and patience, not hardware, which is worth knowing if you’ve read guides about what running a business from home needs and wondered whether your machine was the problem. It almost never is.
Related reading: facebook ads keep playing in background.
Related reading: facebook ads cost philippines.
Frequently asked questions
Why do my Facebook ad costs keep going up even though nothing’s changed?
Ad fatigue is the most common cause; once frequency (how many times the same person sees your ad) climbs past 3 to 4 in a week, cost per result reliably rises because people start scrolling past creative they’ve already seen. Rotating in fresh creative every 10 to 14 days usually fixes it.
Is it worth learning Facebook ads myself or should I hire someone?
If you can dedicate at least three to four focused hours a week for the first month and you’re comfortable reading Meta’s Events Manager and Ads Manager reports, it’s learnable. If your time is better spent elsewhere or you’ve already burned through a few hundred pounds with no clear results, a short project with an experienced freelancer or consultant to set the account structure up is usually the faster fix.
What is the learning phase and why does it matter so much?
It’s the period where Meta’s algorithm is still gathering data on who converts, and it officially wants around 50 conversion events within seven days before it optimises reliably. Editing the ad set during this window resets the count to zero, which is why accounts that get tweaked constantly never seem to improve.
Why did Facebook ads get harder after 2021 specifically?
Apple’s iOS 14.5 update in April 2021 let iPhone users opt out of cross-app tracking, and the vast majority did, which broke a lot of Meta’s ability to follow someone from ad click to purchase. Meta’s own reporting attributed over $10 billion in lost revenue to this in 2022, and the platform has been rebuilding around server-side tracking (the Conversions API) and heavier automation ever since.
