A reach drop feels like punishment, but it is almost always information. The way to read it is the way an auditor would: as a dataset with a before and after, not as a mystery. Here is the diagnostic sequence we run on every channel we audit.
Step 1: Establish the real baseline
Take your median views — not your average, which outlier hits distort. Now compare the median of your last 15 posts against the 15 before the drop. A fall from a 10K median to 6K is a trend worth investigating; one video doing 800 after a 20K hit is just regression to your median, and the correct response is nothing.
Step 2: Date the break
Find the exact post where the numbers changed. Then ask what else changed at that date: a posting gap? A new format? A new editing app (watermarks)? Did you start reposting Reels? Did a video get flagged around then? Reach drops nearly always have a start date, and the start date nearly always has an explanation.
Step 3: Check the exceptions
Almost no drop is uniform. Find the posts since the break that beat the new, lower median — they are your control group. What do they share? Often it is a format, a topic, a length, or a posting time. The algorithm is telling you what it still wants from you; the exceptions are the message.
Step 4: Rule out the mechanical causes
In order: account status (strikes limit reach with a visible timer), watermarked or reposted content (systematically suppressed), cadence collapse (long gaps reset your momentum), and audio issues (uncleared sounds can limit distribution on some surfaces). These are boring, checkable, and responsible for a large share of “the algorithm hates me”.
Step 5: Accept the uncomfortable possibility
Formats fatigue. A hook structure that worked for six months stops working when your audience has seen it thirty times. If steps 1–4 come up clean, the drop is usually creative, not technical — and the fix is testing new openings, not posting the old format harder.
The shortcut
This whole sequence is mechanical, which means it can be automated. It is literally what our audit engine does: computes your medians and cadence, dates the trend breaks, finds the outliers, checks every post’s audio — then writes up what it found and what to do about it.
Leave a Reply