An organic post that did well is a candidate, not a winner. What makes it a candidate is not likes, it is saves and shares, because those measure usefulness rather than goodwill. And before you put money behind it, it has to be rewritten for somebody who does not know you, which is a completely different reader from the one who saw it first.
What you are actually buying
Promoting a piece that already went out is not repeating it louder. It is changing who sees it, and that change is bigger than it looks. The organic version was seen by people who follow you, who get your references, who already know what you do and who arrived with a predisposition you never paid for. The promoted version will be seen by somebody with none of that.
That is why a good organic piece is a hypothesis and not a result. What you learned for free is that the topic is interesting, that the format holds attention and that the hook works at least on people who already know you. It is valuable information, and it is exactly half the question.
The other half, the one only money answers, is whether the same thing stands up without the context. An in joke, a reference to a previous video or a "like I was telling you last week" all work at home and collapse outside with nothing wrong with the piece. Before promoting, read it as a stranger would read it, and work almost always falls out of that.
Which signal to look at, and where the signal exists
Likes measure goodwill. They are the easiest thing to collect and the worst predictor. Saving a piece means "this is useful to me later". Sharing it means "this is useful to somebody else". Both point at usefulness, which is the only thing that survives a cold audience, and both are far more expensive to earn than a tap on a heart.
The practical problem is that those signals are not available everywhere, and it is worth knowing that before you build a rule on a metric your network never reports. This is what actually arrives from each network with connected analytics.
| Network | Saves | Reach | Impressions | Clicks |
|---|---|---|---|---|
| Yes | Yes | Yes | No | |
| No | Yes | Yes | Yes | |
| TikTok | No | No | No | No |
| YouTube | No | No | No | No |
| No | Yes | Yes | Yes |
The uncomfortable reading of that table is that on TikTok and YouTube you will be deciding on engagement and on whatever the app itself shows you, because reach and impressions do not arrive through connected analytics at all. And on Instagram, which is where this decision gets made most often, you have saves but no clicks, so the question "how many people went somewhere" has to be answered with tagged links rather than with the network's own metric. What each of those columns means, and which of them cannot be summed across days, is in the engagement rate guide.
How long to wait, and why waiting too long costs too
Two clocks are running and they point in opposite directions. One is evidence: a piece needs hours to show whether it stands on its own or whether it only got a good first push. The other is currency: the older the post, the odder it looks turning up as an ad, and the less it has to do with what you are talking about now.
Between 48 and 72 hours is usually where they cross, because by then most of a piece's engagement has already happened and the post is still current. But the number is not universal: if your account gathers engagement slowly over a week, your window is longer, and you can see that in your own data by looking at what percentage of its final total a post has collected in each window since it went out.
That figure is worth working out once and keeping, because it answers more than this question. The same curve tells you when the next piece can go out without stepping on the last one, how long to wait before declaring that something did not work, and from what point a monthly report can be considered closed without the final days distorting it.
What to rewrite before you pay
An organic piece is written for a context the ad does not carry with it. Four things almost always need changing, and none of them is the video.
- The opening. Organically, the hook competes with the rest of the feed; paid, it also competes with the feeling of watching an advert. The start usually needs reframing so it says what this is about before anybody decides to skip.
- The subtitles. Plenty of people watch without sound, and a cold audience has no patience for working out what you are on about. If the organic version had none, the promoted one needs them.
- The call to action. Organically you can leave it implicit, because people who follow you know where to find you. A stranger does not, and a piece with no exit is paid reach going nowhere.
- The missing context. Anything that depends on having seen something of yours before has to be explained in a sentence or cut.
None of that gets written for you by a tool, and it is worth saying because the opposite promise is everywhere. Reframing an opening and deciding which call to action goes at the end is the work of somebody who knows the brand, and it is exactly the work described in the ad creative guide, where a rescued organic piece is simply one more variant inside the set.
Everything in this article happens in one place in GoFeed.
Try it freeReading the result next to what you already knew
The part almost nobody sets up properly is what comes after. A promoted piece now has two lives, the organic one and the paid one, and both keep running at the same time. Look only at the ads side and you miss what the piece is still doing on its own; look only at the network's analytics and you credit the content with something you paid for.
What you can do is keep them close together. Meta, Google, LinkedIn, TikTok and Pinterest ads accounts are connected on the brand, and their spend and results are read in one list, beside that same brand's calendar and numbers: that is the ads module, and exactly what each platform brings is broken down on the ads dashboard page. Google campaigns are created in Google Ads and read here.
There is also an asymmetry worth knowing before you plan the month. Promoting a post that already went out is an operation only Meta opens up to a third party: on the other four platforms that same piece goes in as a new creative inside whichever campaign it belongs to, which is not worse, but it is different work on a different timetable.
The honest comparison is against the same account's previous period, not against the organic version of the piece itself. And resist adding organic reach to paid reach: somebody who saw you both ways is one person, not two, and that sum inflates the number in exactly the direction you would like to believe. Why reach cannot be summed across days or across sources is covered in the reach and impressions guide.
When not to promote
Some pieces should not be promoted however good their numbers look, and recognising them saves more money than fine tuning everything else.
The first is the one that worked because of context. A piece about that week's news, a reply to something that happened in your industry, or a thank you for a milestone of yours all post high numbers and hold zero interest for somebody arriving cold.
The second is the one that worked because one specific person shared it. If the spike came from a mention, what worked was the mention, and the ad does not bring it along.
The third is the one that leads nowhere. If the piece is good but there is no page, no message and no conversation waiting behind it, you are buying attention and handing it back empty. Fixing the destination takes an afternoon and improves what everything else returns, so it comes before the budget and not after.
And one last case, which is a habit rather than a piece: do not promote to paper over a weak month. If the organic side is not working, money buys reach without fixing the cause, and next month you are back in the same place with less budget. Hold the publishing first, then put money behind what already works.
