Somebody sends you a link. The headline says to stop running ads until you watch it. The algorithm changed, the update is here, everything you were doing is now wrong, and there is a strong implication that if you finish your coffee before acting you will have already lost money.
This happens roughly every three weeks. It has happened every three weeks for years. And in almost every case, the correct response from a small business owner is to do nothing at all.
I want to be careful here, because "ignore the news" is lazy advice and some updates genuinely matter. A few of them will cost you real money if you sleep through them. So this is not a case for burying your head. It is a case for having a way to tell the difference, because right now most owners do not have one, and the absence of a filter is what makes all of it feel urgent.
Somebody Is Selling You the Urgency
Start with who is talking.
Platform updates are extremely good content. They are timely, they sound important, and they give the person making the video a reason to tell you that what you are currently doing is broken. That is a persuasive combination, and it converts. Almost every "the algorithm changed" video you will find is made by an agency, a course seller, or a software company, and the update is the hook rather than the point.
You can watch this happen in real time. Two of the more popular Facebook ads videos going around right now both open by naming the same Meta update as the reason your account needs rebuilding. One tells you the answer is a single campaign with the budget set at the campaign level and thirty creatives underneath it. The other tells you that whether the budget belongs at the campaign level or the ad set level depends entirely on how much you spend per day and whether you already have messaging that works, which is not a statement about the update at all. Both run agencies. Both are certain.
They cannot both be the response to the same change, because they are not the same response.
The Update They Told You to Rebuild Around Has No Settings
The update in question is called Andromeda. If you have never heard of it, that is worth sitting with for a second, because you are the person being told to restructure your business's advertising because of it.
Here is what it actually is, from Meta's own engineering blog, published in December 2024. Andromeda is a retrieval engine. When Meta decides which ad to show one person, it starts with millions of possible ads and has to narrow that down to a handful of candidates in a fraction of a second. Andromeda is the machinery that does the narrowing. Meta rebuilt it on new hardware and reports a "+6% recall improvement to the retrieval system" and an "+8% ads quality improvement on selected segments."
Now the part nobody puts in the thumbnail. There is no switch. Andromeda is not a feature you turn on, a setting you choose, a campaign type you select, or a checkbox you missed. It is plumbing that Meta replaced underneath every account at once, yours included, whether you watched the video or not.
So what exactly are you supposed to do about it on Monday morning?
The honest answer is that the advice being sold on the back of it, make more and more varied creative, was good advice in 2019 and will be good advice in 2029. It did not become true because of a retrieval engine. It got attached to a retrieval engine because the retrieval engine made it sound new.
Four Things "An Update" Can Actually Mean
Once you start pulling announcements apart, they land in four piles. Knowing which pile you are looking at answers the question almost every time.
A new name for something you already have. In September 2022, Google renamed ad extensions to assets. Same features, same job, new word. The sitelinks under your ad, the extra phone number, the little list of services, all of it kept working exactly as before. The only thing that changed was the vocabulary, and it changed so thoroughly that the phrase "ad extensions" no longer appears in Google's own help pages at all. That is a real headache, but not the one the announcement described. The headache is that four years of perfectly good advice written by other people now uses a word you cannot find anywhere in your account. If you want the current version, we wrote it up in the free real estate you left empty.
A box that got checked for you. Something now happens by default that used to be a decision. This is the pile that costs money, and we will come back to it.
A new place your ad can appear. A new surface, a new format, a new platform entirely. Almost never urgent. The inventory will still be there in six weeks, and it will be better documented and slightly less expensive to learn on.
A control that got taken away. Something you used to be able to set is now decided for you. Genuinely rare, and the one category that always deserves your attention, because it means a lever you were relying on is gone and you may not find out until performance moves.
The Only One That Costs You Money Arrives Without a Headline
Here is the thing that should annoy you about all of this. The category that reaches into your account and starts spending your money differently is the one nobody makes a video about.
A flipped default has no drama in it. There is no announcement, no countdown, no "act now." Somewhere in a settings panel a box is already ticked, and the money simply starts going somewhere slightly different than it went last month. Nobody can build a hook out of that, because the whole point is that it happens quietly.
Take Meta's audience settings. Per Meta's own documentation: "If you choose the sales, app promotion or leads objective, the Audience section shows Advantage+ on." Read the rest of that page and it explains that most of what you enter is treated as a suggestion rather than an instruction, and that your ads "will also be shown to other audiences when it's likely to improve performance." If you want your audience to be a boundary rather than a hint, you have to go and find an option called "Further limit the reach of your ads."
Nothing about that is hidden, exactly. It is written down, publicly, by Meta. It is just that the writing down is where the announcement effort ended, and no creator is going to build a channel out of reading settings pages.
We have watched this pattern hit client accounts twice recently and written up both. One was a location box checked on your behalf that quietly sends your ads outside the area you actually serve. The other was a set of Google campaign settings that nobody had touched since the account was built. Neither arrived as breaking news. Both were spending money.
If You Cannot Find It in Your Own Account, It Is Not About You
So here is the filter, and it takes about two minutes.
When you see an update, before you change anything, try to answer three questions. What campaign does this touch? What setting or report would I open to see it? What number in my account would move if it were true?
If you can answer all three, you have something real and you should go look. If you cannot answer any of them, you are looking at industry news rather than an instruction, and industry news does not require a response.
Andromeda fails all three. There is no campaign, no setting, and no report where it appears. Almost every "the algorithm changed" story fails all three, which is exactly why the advice attached to them is always so general. Nobody can tell you what to click, so they tell you to make more creative and rebuild your structure, which sounds like action and is mostly just churn.
The Advantage+ audience default passes all three, easily. Campaign: any sales or leads campaign. Setting: the Audience section at the ad set level. Number: who you are actually reaching and what your cost per lead does when you constrain it. That is a Monday morning job.
Half of What Gets Announced Was Never Meant for You
There is a second reason all of this feels louder than it is, and it is worth naming because it quietly makes owners feel behind.
The big announcement events are aimed at everybody at once. At Google Marketing Live in May 2026, Google announced, among other things, that its Meridian marketing mix modeling tool is moving into Analytics 360, alongside a cross-platform assistant and expanded automated campaign types. Marketing mix modeling is a statistical technique for untangling the effect of several large advertising channels on total sales. It is genuinely useful and it is genuinely not for you, because it needs years of history and enough spend across enough channels for the math to separate them.
Analytics 360 is the paid enterprise tier. It costs six figures a year.
The same is true of most connected TV announcements, most data platform announcements, and most of the measurement suites that get significant airtime. They are announced in the same keynote, in the same excited register, as the things that genuinely apply to a business spending three thousand dollars a month. Nothing marks which is which. So an owner watching the recap comes away feeling like the ground moved, when in truth about two thirds of what was announced was addressed to companies a hundred times their size.
Being unable to use an enterprise measurement tool is not falling behind. It is being a small business, which was already true yesterday.
A Brand New Ad Platform Is Still an Ad Platform
The loudest version of all of this is a new platform launching, because that one carries hope as well as urgency.
The hope is specific and it is always the same. Somewhere out there is a channel where nobody is bidding yet, clicks are cheap, and you get a couple of good years before everyone else works it out. That did genuinely happen with Facebook once. It is the story every new platform gets sold on.
It is also worth being honest about why the interest is so high, and it is not really optimism. It is that people feel stuck between two enormous platforms and want a third option to exist. That is a completely reasonable thing to want. It is just not evidence that the third option is cheap.
We went and looked properly at the newest one rather than guessing, and wrote down exactly what you get for your money inside ChatGPT - what the targeting can and cannot do, what the reporting lag is, and the audience size floors that rule most local businesses out for now. The short version is that a new platform is not a discount. It is a platform with fewer controls, less documentation, and nobody who knows what good looks like yet, including the people who built it.
That can still be worth testing. It is not worth reorganizing your month around.
Changing Things Has a Price, and Meta Publishes It
The last piece is the one that turns all of this from a philosophical point into an arithmetic one, and it is the reason reacting to news is not free even when the news is real.
When you make a significant edit to an ad set, Meta puts it back into the learning phase. That is not a rumor; it is in Meta's own help documentation, which says the delivery system is "exploring the best way to deliver your ad set" and that ad sets typically exit "after about 50 results in the week after the ad set's last significant edit." Meta goes further and tells you plainly that by editing during the learning phase "you reset learning and delay our delivery system's ability to optimize."
Sit with the number. Fifty results in a week. If you generate ten leads a week, you do not exit the learning phase in a week. You spend a month or more in the least stable, most expensive state your ad set can be in, and then somebody publishes another update and you do it again.
This is the actual cost of chasing announcements. Not the hour it takes to make the changes. The weeks of unstable delivery you buy each time, on an account that was finally settling down. We have written before about how long ads genuinely take to start working, and every reactive rebuild starts that clock over.
The same instinct shows up on platforms that do not even have a learning phase, which we got into in the LinkedIn version of this problem. The absence of a documented penalty is not permission to fiddle.
What This Looks Like on Monday
None of this is an argument for ignoring your accounts. It is an argument for where the attention goes.
Set yourself a rule that news does not trigger changes. Data does. When an update crosses your desk, run the three questions: what campaign, what setting, what number. If it survives all three, put it on the list for your normal weekly review rather than acting on it that afternoon. If it does not survive, you are allowed to close the tab.
Then spend twenty minutes a quarter doing the thing that actually pays, which is opening your settings and reading them. Not the news about the settings. The settings. Every default in there was chosen by somebody whose revenue goes up when you leave it alone, and that is where you will find the box that got checked for you. If you want a sane rhythm for the rest of it, we laid one out in checking your account daily is a habit, not management.
The businesses that do well with paid advertising over several years are almost never the ones who adopted every feature first. They are the ones who picked something sensible, funded it properly, left it alone long enough to learn something, and read their own settings once in a while. That is a boring answer. It is also the one that survives contact with the next three updates, whatever they turn out to be.
We do this all day, across a lot of accounts, so we see which changes actually move a number and which ones just moved a lot of people to rebuild their campaigns for nothing. If you would rather spend your Monday running your business than deciding whether an announcement applies to you, that is a reasonable thing to hand off. And if you would rather know it yourself, everything above is how we do it.
