It is a few minutes past seven. Coffee is on, the shop does not open for another hour, and you have the ad account open on your phone. Forty-seven dollars spent yesterday. Two clicks. No leads.
You feel something in your chest. Not panic exactly, but the specific low-grade dread of watching money leave and nothing arrive. So you tap into the campaign, look at the keywords, and turn one off. Or you lower the budget by ten dollars. Or you swap the headline you were never sure about anyway.
Then you close the phone and get on with your day, and the thing you just did will cost you more than the forty-seven dollars did.
Nobody ever tells an owner how often they are supposed to look at their ad account. It is one of the most common questions I get and one of the least written about, probably because the honest answer is not a number. It is a schedule, and it is different for different questions.
The Problem Was Never the Looking
Let me be precise about what costs money here, because "stop checking your account" is bad advice and I am not going to give it.
Looking is free. Looking is good. An owner who knows what their account is doing makes better decisions than one who does not, and I would much rather work with someone who opens the dashboard than someone who has not logged in since March.
The problem is that almost nobody opens an ad account and then closes it.
Watch what actually happens. You look, you see a number you do not like, and you feel the pull to do something about it. Doing something is the only available way to discharge the feeling. Waiting does not discharge it. Understanding does not discharge it. Changing something does, immediately, because now you have acted and you can tell yourself the situation is handled.
So the discipline is not about looking less. It is about separating the act of observing from the act of intervening, and putting them on different clocks. Most owners have them welded together, which is why a bad Tuesday morning turns into a campaign edit by 7:15.
And the reason that specific edit is expensive is not a matter of opinion. Both major platforms publish exactly what it does.
Meta Publishes the List. Almost Nobody Reads It.
Meta's delivery system goes through what it calls a learning phase, and its own documentation describes it plainly: it is the period when the system is still working out how best to deliver your ad set. During it, performance is less stable and the cost per result is usually higher. An ad set leaves that phase once it can deliver stably, which Meta says usually happens after about fifty results in the week following the ad set's last significant edit.
Read that last part again, because the whole argument sits inside it. The clock does not run from when you launched. It runs from the last time you changed something significant.
And Meta publishes the list of what counts as significant. These reset the learning phase:
- Any change to targeting
- Any change to ad creative
- Any change to the optimization event
- Adding a new ad to your ad set
- Pausing the ad set for seven days or longer
- Changing the bid strategy
A second group may or may not reset it, depending on how big the change is: the ad set spending limit, the bid control or cost per result goal, and the budget amount. Meta's own example is the useful one. Moving a budget from one hundred dollars to one hundred and one will probably do nothing. Moving it from one hundred to a thousand may well send the ad set back to the start.
Now hold that list up against the things an owner actually does at 7am on a bad morning. Swap the headline. That is a creative change. Narrow the audience because the leads felt unqualified. That is a targeting change. Drop the budget because yesterday looked expensive. That is possibly a budget change large enough to matter. Add a new ad because you finally got the photos back. That is on the list too.
Every one of those is a defensible decision on its own. Made on a Tuesday because the numbers looked bad on a Monday, they are a reset, and you have just bought yourself another week of the expensive, unstable period you were trying to escape.
Meta's own guidance on this is one sentence long and I think it is the single most useful sentence the company has ever published for a small advertiser: only edit your ads or ad set when you have reason to believe that the edit should improve performance.
Not when you are worried. Not when the number is ugly. When you have a reason.
Do this today, it takes two minutes. Open Ads Manager, click Columns, and add the column called "Last significant edit." Then add "Results" next to it. You will be able to see, for every ad set, when its clock last got reset and how many results it has managed since. For a lot of owners this is genuinely uncomfortable viewing, because it turns out the account has been restarted every four or five days for months and has therefore never once been allowed to finish learning.
Google's Clock Is Different, and It Is Longer
Google runs the same idea on a different mechanism, and the difference matters if you are running both.
When you change an automated bid strategy, Google shows a "Learning" status, and it will tell you which of three things caused it: a new strategy, a setting change, or a composition change, meaning campaigns, ad groups or keywords were added to or removed from the strategy. On Shopping campaigns, an ad group target change can trigger it too.
How long it lasts depends on three things: how many conversions the campaign is getting, how long your conversion cycle is, and which bid strategy you are using. Google's stated figure is up to three weeks, or one to two conversion cycles.
That second measure is the one small businesses consistently get wrong, and it is worth sitting with. A conversion cycle is how long it takes a click to turn into a conversion. If you sell something people buy the same afternoon, your cycle is short and three weeks is a fair estimate. If you sell a roof, a legal matter, or anything where the customer talks to their spouse first and calls you back in a fortnight, your conversion cycle is two or three weeks on its own. One to two of those is a month or more. Judging that campaign at day nine is not impatience, it is a category error: the data you are judging has not physically finished happening yet.
Two more things from Google worth knowing. The learning period does not apply to Manual CPC, because there is no algorithm calibrating anything. And even after the "Learning" label disappears, Google says its algorithms carry on learning. The label going away is not a starting pistol.
None of this means the platforms are always right or that you should sit on your hands for a month. It means the cost of an edit is real and documented, so an edit should be worth paying for. Which brings us to what you should actually be doing, and when.
What Genuinely Deserves a Daily Look
There is a daily list. It is short, and everything on it has one thing in common: it is about something being broken, not about something being disappointing.
- Is it spending at all? A campaign that spent zero yesterday is not underperforming, it is off, or disapproved, or out of budget, or the card was declined. That is a same-day problem.
- Any disapprovals? An ad or an asset can get knocked back overnight for something trivial. A disapproved ad earns nothing while you wait.
- Is tracking still firing? Conversions sitting at zero for a couple of days when they never used to is far more often a broken form, a changed thank-you page, or a tag someone removed during a site update than it is a collapse in demand.
- Did the budget get eaten by noon? Worth knowing, particularly if you are a service business and the phone only rings in business hours.
- Does the form still submit? Fill it in yourself. Once a week is fine, but if leads suddenly stopped, do it immediately.
The reason this list is safe to check daily is that every item is binary. Something is either working or it is not, and one day is enough to tell. You do not need a sample size to know a form is broken.
And notice what is not on the list: cost per lead, conversion rate, click-through rate, and every other number that needs volume behind it before it means anything. Checking those daily is not management. It is watching a kettle, except the kettle gets slower every time you lift the lid.
The other thing worth saying: almost nothing on the daily list requires you to touch the campaign. You are checking that the machine is switched on, not adjusting it.
The Weekly Pass Is Where the Actual Work Happens
Once a week, sit down properly. Not on your phone, not between jobs. This is the pass that earns its keep.
Read the search terms report. On Google, this is the receipt for what people actually typed to reach you, as opposed to the keywords you thought you were buying. Add negatives for the searches that were never going to be customers: the job seekers, the DIY crowd, the people looking for the service you deliberately do not offer. This is the highest-value weekly habit there is, and conveniently it is also among the safest edits you can make. You are removing waste rather than changing what the system is optimizing toward.
Look at frequency and creative wear on Meta. If the same people are seeing the same ad repeatedly and results are drifting down, that is fatigue, and it is a genuine reason to introduce new creative. That is an edit with a reason behind it, which is exactly what Meta asked for.
Read the actual leads. Not the count. The leads. Listen to two or three calls, read the form submissions, and ask whether these are people you want. This is the most important thing in this entire article and it is the one thing no dashboard will ever do for you. The platform can tell you that a lead cost thirty-one dollars. It has no idea whether that lead was a serious buyer or somebody who thought you were a different company. Only you know that, and if you never look, you can spend months optimizing efficiently toward the wrong people.
Check pacing. Are you on track to spend roughly what you meant to spend this month? Not down to the dollar. Roughly.
The Monthly Number Is the One That Matters
Once a month you get to ask the real question, and it is not a marketing question at all.
What did a customer cost, and what is a customer worth?
Everything upstream of that is diagnostics. Cost per click, cost per lead, click-through rate: those exist to tell you why the real number is what it is, not to be optimized in their own right. A forty-dollar lead is not expensive if a customer is worth four thousand, and a twelve-dollar lead is a disaster if nobody ever buys.
A month is also roughly the point where you have enough data to make structural decisions rather than nervous ones. This is where you move budget between campaigns, decide something has genuinely earned more money, or admit that something has not. It is also the right altitude for questions about how the account is built, which is a real lever and one that tidiness instincts tend to get exactly backwards. If you have ever split things into more campaigns because it felt more organized, that is worth understanding properly, and I wrote about what over-organizing a Google Ads account does to it separately.
One caution on the monthly read. A single month that looks bad is not automatically a signal, and a single month that looks great is not automatically a win. Small accounts are noisy, and the temptation to declare a result from not very much evidence is how most people fool themselves with their own data.
The Quarterly Questions Are Not Account Questions
Every three months, close the ad platform entirely. Genuinely close it. The questions at this level are not in there.
Is the offer still right? Is the price still right? Has the market moved? Are the competitors doing something new? Is this still the right channel for the business you have now, as opposed to the business you had when you set this up?
I put this last, but in terms of what actually moves the number it belongs first. An account can be run immaculately and still lose to a business with a better offer. No amount of daily checking will ever surface that, because the account cannot see it. The dashboard has no column for "your competitor now includes installation and you do not."
Picture how invisible this is from inside the dashboard. Say you run plumbing ads, and two competitors in your city start putting fixed prices on their websites while yours still says "call for a quote." Nothing in your ad account moves. Cost per click holds. Conversion rate holds. The number that quietly falls is the share of those leads that turn into booked jobs, because people are ringing three plumbers, getting a figure from two of them and a conversation from you. Every metric you check daily would look fine for months. That is a quarterly-level problem wearing a monthly-level disguise, and the fix is on the website, not in the campaign.
This is also the level at which adding or dropping a channel gets decided, rather than in a moment of frustration. If that question is live for you, which channel a small business should actually start with is the other half of it.
If You Cannot Act on It, Do Not Look at It
That is the whole rule, and it is worth writing on something.
The cadence is not about discipline for its own sake, and it is certainly not about caring less. It is about protecting the account from the version of you that is anxious at seven in the morning with a phone in your hand and no plan. That person is not stupid. That person is invested, which is the whole problem, because being invested and having enough information are different things and they feel identical at the time.
So: check the breakage list daily, because breakage is binary and a day is enough. Do the real pass weekly, because a week is enough to see search terms and read leads. Judge cost per customer monthly, because a month is roughly enough to mean something. Ask the business questions quarterly, because those move slowly and you cannot see them from inside the account anyway.
And when you launch something new, none of the above starts until the learning period is done. That is a different clock with different rules, and I have written about how long you should actually give a campaign before judging it.
If you want a second opinion on whether your account is being managed or merely watched, we are happy to look at it and tell you honestly. Sometimes the answer is that it is fine and the checking is the only thing that needs to change.
