You start posting about what you are building. Three weeks in, your posts get a handful of likes, mostly from people you already know. You conclude it is not working and stop.
That is the single most common way this fails, and it is a misreading of what the early period is.
What the first month actually is
The first month is not distribution. It is inventory.
Nobody is going to find your seventh post because of your seventh post. They are going to find one post, somehow, and then look at the other twenty to decide whether you are worth following. The early posts are not broadcasts. They are the thing a visitor reads after the one that got their attention.
Which means the question to ask at week three is not "how many people saw this" but "if someone landed here cold, would they follow". Those have completely different answers and only one of them is under your control.
The realistic shape
Roughly, and with enormous variance:
Weeks 1 to 4. Almost no reach. Engagement is friends and the occasional stranger. Your job is to establish what the account is about and build up enough posts that a profile visit is worth something.
Weeks 5 to 8. Something gets picked up. Usually not the post you expected. Often an older one that someone found and shared. The first sign that anything exists outside your immediate circle.
Weeks 9 to 12. If you have kept going, a small number of people are now actually following the work rather than passing through. This is where replies start coming from the same names and the account starts to have a shape.
Plenty of accounts take longer. Very few take less.
What to measure instead of likes
Likes at this stage are noise. They correlate with time of day and who happened to be scrolling far more than with quality.
More useful:
Profile visits. Someone read a post and wanted to know who wrote it. That is the actual funnel.
Replies with substance. One reply engaging with the content is worth fifty likes. It means someone read it properly.
Saves or bookmarks, where visible. Indicates something worth returning to, which is the strongest quality signal a platform exposes.
Whether you would follow you. Read your own last ten posts as a stranger. This is unscientific and it is the most useful check available in month one.
The compounding thing
Building in public compounds, which has an uncomfortable implication: the early returns are genuinely poor and there is no way to skip that.
A post today reaches your current audience. A post in six months reaches an audience partly built by today's post. The returns arrive late and they arrive for work you did early.
This is why consistency matters more than quality in the first stretch, within reason. Ten decent posts beat three excellent ones, because the ten create more surface area for the one that eventually lands.
The mistake that wastes the 90 days
Posting only when you have something impressive.
The gaps are what kill accounts. Someone finds you, likes a post, checks the profile, sees the last entry was five weeks ago, and does not follow. Not because the content was bad but because there is no evidence of ongoing anything.
A steady stream of ordinary work beats sporadic excellence, in this specific window, for this specific reason. After you have an audience the calculus changes. Before you have one, showing up is most of it.