Marcus runs a small landscaping company. On Monday he launched his first Facebook ad: a photo of a freshly finished backyard, a line about fall cleanups, and a link to the free quote form on his website.
On Thursday he opens Ads Manager to see how it’s going. What he finds is a wall of columns and numbers he doesn’t recognize, a status that says “Learning,” and just 2 leads. His stomach drops. He’s already wondering if he should switch it off and try something else.
(Marcus is an example we made up, and so are all the numbers in this post.)
If you’ve been in Marcus’s shoes, this post is for you. Three days in is one of the hardest moments in running an ad, because the numbers are real but they don’t yet mean what you think. Let’s look at what’s actually happening behind the scenes first, then at the few numbers worth your attention.
What Happens in the First Few Days: The Learning Phase
When you launch a new ad, Meta doesn’t know yet who is most likely to respond to it. So it experiments. It shows your ad to different groups of people, in different places across Facebook and Instagram, and watches what happens.
Meta calls this the learning phase. In its help article About the learning phase, Meta describes it as the period when the delivery system still needs to learn how your ad set may deliver and perform. While it’s happening, the Delivery column in Ads Manager reads “Learning.”
Think of it like a new employee’s first week. They’re figuring out where things are and who to ask. They’ll get faster, but judging them on day three wouldn’t be fair.
Here’s what Meta says about this period:
- Ad sets usually leave the learning phase after about 50 results in the week after the last significant edit. A “result” is whatever your ad is set to get, such as leads.
- During learning, performance is less stable, and the cost per result is usually higher.
- Editing your ad, ad set or campaign during learning resets it, which delays the system’s ability to improve.
That last point is the important one. If Marcus panics on Thursday and changes his photo, his audience or his budget, he sends the ad back to the start of learning. He’d be throwing away the three days of experimenting he’s already paid for.
You might also see a status called “Learning limited.” That means the ad set isn’t getting enough results to finish the learning phase. It’s common for small local budgets and doesn’t mean the ad is broken. It’s a sign to look at the bigger picture over a longer time, rather than to make quick changes.
You can see each ad’s status in the Delivery column; Meta explains every status in View campaign, ad set or ad delivery status.
The Few Numbers That Matter Early
Ads Manager can show dozens of columns. In the first few days, you only need a handful. Meta explains each one in its help center (see, for example, its pages on reach and Clicks (all)), but here they are in plain words.
Reach and impressions: is the ad being shown?
Reach is the number of different people who saw your ad. Impressions is the total number of times it was shown, counting repeats. If one person sees your ad twice, that’s a reach of 1 and 2 impressions.
In the first few days, you just want to see that these numbers are growing. If reach is near zero, your ad isn’t really running yet, and that’s a different problem (we’ll come back to it).
Frequency: how often the same person sees it
Frequency is impressions divided by reach: the average number of times each person has seen your ad. Early on it’s usually low, and that’s fine. Frequency matters more after a few weeks, when the same people may start seeing the ad so often they tune it out.
Link clicks and CTR (link): are people interested?
Link clicks counts the clicks that took people through to your form or website. CTR (link) is the share of impressions that led to a link click.
Be careful with a similar-looking column called Clicks (all). It also counts likes, comments, shares and taps on your Page name, so it can make an ad look busier than it is. For “is anyone interested enough to find out more?”, link clicks is the honest number.
Results and cost per result: is it doing its job?
Results is the number of times your ad got the thing you asked for, such as leads or messages. What counts as a result depends on the goal you chose when you created the campaign, so if Marcus had chosen “traffic” instead of “leads,” his Results column would count website visits instead. Cost per result is what you’ve spent divided by that number.
This is the number that matters most in the long run. It’s also the one that looks worst in the first few days, because of the learning phase.
Putting it together: Marcus’s first three days
Here’s what Marcus sees on Thursday:
- Amount spent: $60
- Reach: 2,400 people
- Impressions: 3,100
- Frequency: about 1.3
- Link clicks: 38
- Results (leads): 2
- Cost per result: $30
Let’s read it the way an experienced advertiser would. The ad is being shown to plenty of people, so delivery is fine. Frequency is low, so nobody’s tired of it yet. The ad got 38 link clicks, so the photo and the offer are doing their job. Two leads at $30 each feels disappointing, but it’s three days into a learning phase that usually takes about 50 results to finish. Nothing in these numbers says “turn it off.”
Why Your Leads May Look Lower Than They Really Are
There’s one more reason the early numbers can look worse than reality: Meta may not see every lead.
If your ad sends people to your own website, Meta usually learns about a lead through a small piece of code on that site called the Meta Pixel. But browser privacy settings and ad blockers can stop that code from reporting back. So a customer might fill out your form, and Meta never hears about it. Your ad did its job, but Ads Manager shows one result fewer.
That’s what Meta’s Conversions API is for. It lets your website send leads to Meta directly from your server, so fewer get lost along the way. We explain how it works, and how to tell if you need it, in Why Facebook Ads Under-Report Leads, and How the Conversions API Helps.
Even without it, there’s a simple check anyone can do: count the leads in your email inbox or phone log from the last three days, and compare them with Ads Manager. If your inbox has more, your ad is working better than the report shows.
When to Wait, and When to Change Something
So how do you know whether to leave the ad alone or step in? Here’s a simple way to think about it.
Wait if:
- The ad is in “Learning” and is being shown to people.
- People are clicking, even if leads are few so far.
- Cost per result is high but slowly coming down.
Take a closer look if:
- The ad isn’t being shown at all. Reach is near zero after a full day. Check the Delivery column for a status like “In review” or “Rejected,” and check that your payment method is working.
- Lots of clicks, almost no leads. The ad is doing its job, but something after the click isn’t. Open the form or page yourself on your phone. Is it slow, confusing or too long? Our post on Facebook lead forms vs a landing page can help you decide which works better for your business.
- Later on, frequency climbs and results drop. After a few weeks, if people are seeing the ad again and again and responding less, it’s time for fresh photos or wording.
Three Common Mistakes in the First Week
Almost everyone running their first Facebook ad is tempted by the same few moves. Here’s why it pays to resist them.
Changing the ad every day. It feels productive to tweak the headline on Tuesday and the photo on Wednesday. But each significant change can send the ad back into learning, so Meta never gets the steady run of results it needs to settle down. If you want to test a new idea, it’s usually better to let the first version run long enough to judge it, then compare.
Making big budget jumps. Doubling or tripling the budget because the first day looked good, or slashing it because the second day looked bad, can also reset learning. Meta’s page on significant edits and the learning phase explains which edits count as significant. Smaller, less frequent changes give the system a fairer chance.
Judging the ad by a single day. Ad results bounce around from day to day, especially with small budgets. A Tuesday with no leads followed by a Wednesday with three is completely normal. Look at the total over the week, not at any one day.
There’s also a quieter mistake: comparing your results to numbers you read online. A “good” cost per lead for a dentist in a big city has little to do with a “good” cost per lead for a landscaper in a small town. The most useful comparison is with your own business: what is a new customer worth to you, and how many leads does it take to win one?
Your 3-day check
When you open Ads Manager three days in, ask just these questions:
- Is the ad being shown? (Reach is growing.)
- Are people clicking through? (Link clicks, not Clicks (all).)
- Do my inbox and call log agree with the Results column?
- Is the ad still in Learning? If so, hold off on edits.
Back to Marcus
Marcus decides to leave the ad alone for another week. Before closing Ads Manager, he checks his inbox and finds a third quote request that Ads Manager hadn’t counted, from a customer who filled out the quote form on his website. The ad was doing slightly better than the report said.
By resisting the urge to change everything on day three, he gave the ad the time it needed to learn.
If you’d like someone to watch these numbers for you, we run Facebook and Instagram ads for Tacoma businesses, and you can read more about our Meta Ads service for local businesses anywhere.
