TikTok algorithm ghosting me after 6 months off?

chloeest

Just came back from a 6-month break and the difference is staggering. Test A: my old posts (pre-break) - 5% like rate, 1% save rate would sail past 100k views. Test B: now - 20%+ like rate, yet I'm lucky to crack 400 views. That's a 4x improvement in engagement quality for basically zero reach. Something's shifted.

I'm thinking the algorithm is punishing the account freshness penalty after that long hiatus, or they're heavily weighting watch time and completion rate above likes now. Someone pointed out that TikTok's official docs list watch time, completion rate, shares, and comments as dominant ranking signals - likes sit way lower. So maybe my warm audience is inflating likes while cold traffic never gets a chance because the retention metrics suck.

Going to pull the analytics on my last few posts: watch time %, completion rate, and compare against a pre-break control. If those are tanking, that's the bottleneck. If they're solid too... then either the algorithm is broken or my account got throttled. Anyone else seen similar after a long break?

zoecreative

Been digging into TikTok's ranking signals for a while now. Watch time, completion rate, shares and comments are the heavy hitters - likes are secondary. That pattern of a few hundred views with a high like percentage? Classic warm audience trap. The bottleneck's almost certainly in watch time or completion rate. Dive into those analytics instead.

growthgeek

Six months break? Lost optimisation momentum.

davidsearch

The algorithm shake-up is real. Likes are basically vanity metrics now-TikTok's optimisation model clearly weights watch-through rate, replays, and shares above anything else. A 20% like rate doesn't matter if the CTR-to-completion is poor, you're bleeding CPA on cold traffic. Taking a long break resets your account's trust score with the algo-it treats you like a fresh campaign with zero historical data. Need to prove consistency again. Focus on hooks that drive deep retention, not just double taps. Views will recover after 10-14 days of daily posting.

growthgeek

Everyone's obsessed with likes, but that's not what moves the needle. TikTok's algorithm is prioritising shares and watch time now - if your content isn't getting passed around, reach stalls no matter how many hearts you rack up.

tiktokguru

TikTok's algorithm has definitely pivoted hard towards retention and watch time over likes or shares. From what I've seen, a high like ratio on low views usually means your hook is spot-on for the tiny sample that saw it, but that initial test pool isn't expanding because the completion rate or rewatch percentage is dragging.

A few things I'd focus on after a six-month gap (yes, your account signals do reset - you're basically a new user again with smaller test batches):

🎯 Nail the first 3 seconds - that's where everything lives or dies
📈 Aim for 70%+ completion rate on shorter clips
⏰ Post consistently for 2-3 weeks to rebuild momentum

Likes barely matter now. Watch time is the only metric that counts.

ryanseo

People keep saying the algorithm isn't random, but I'm not convinced after running two ASMR audio accounts myself. We're talking over a hundred videos in a niche where the content is almost identical each time - same structure, same audio quality, same length. The metrics that supposedly matter, like completion rate, shares, and comments, don't line up with what actually gets pushed. I've analysed the data point by point. The video with the most views is rarely the one with the strongest engagement signals.

Then there's the real kicker: you can post the exact same clip twice and get 400 views on one, 100k on the other. There's no logical threading through it. It's not even about testing different hooks or thumbnails because ASMR audio is pretty stripped back visually. It genuinely behaves like a lottery draw more than a merit-based system.

chloeest

Test A: Post consistently for 3 months, go cold turkey on holiday. test B: Schedule posts in advance to maintain momentum. I'd bet the heatmap of that account tanked the second you stopped. would love to see the engagement decay curve - probably a sharp drop-off after day 5.

ryanseo

i've been running tests in a similar vein-though mine are in the programmatic SEO space rather than ASMR audio. The randomness you're describing isn't just a TikTok-specific quirk, it's a feature of any platform trying to balance exploration vs. exploitation. that said, your ASMR niche is a clean lab environment because the content is virtually identical each time, so any variance in performance is pure signal.

i've seen the same pattern across three accounts in the "lo-fi study beats" corner. two pieces of content that are functionally the same (same length, same hook, same audio source) will see drastically different reach. The "winning" video rarely aligns with the highest completion rate or share count. it's like the algorithm throws darts at a board and then amplifies whichever one sticks first.

A few observations from tracking this over ₂₀₀ uploads:

First 60 minutes matter disproportionately. The video that gets a lucky early push (maybe from a single high-retention user in the right geo) enters a feedback loop of impressions → engagement → more impressions.
Posting time vs. audience timezone asymmetry. Two identical videos posted an hour apart can land in completely different initial feed cohorts.
The "lottery" element is real, but there are levers. Using a slightly different audio frequency or a minimal thumbnail difference can tip the initial batch of impressions.

Honestly, it feels as opaque as Google's early days before we reverse-engineered ranking signals. TikTok's reward distribution is heavily skewed toward novelty-even if the content is identical, the platform treats each upload as a fresh experiment. you're not doing anything wrong, you're just playing a game where the house edge changes every week

ryanseo

People keep saying the algorithm is broken, but it's not random - it's just impossible to reverse engineer for any single piece of content. I've run two accounts in the ASMR audio niche, 100+ videos all built around the same format. The video with the best completion rate, shares, and comments didn't get the most views, a near-identical clip hit 100K while another sat at 400. Same content, totally different outcomes. It feels like an absolute lottery.

That niche is the perfect test bed because the variables are so controlled. The algorithm isn't random, but it's processing dozens of signals we can't see - user session context, time of day, network effects, even device type. So you can do everything right and still get buried. Methodical testing helps, but there's no silver bullet.

zoecreative

Honestly, I've seen this pattern a lot in my own testing. The idea that an algorithm would be truly random doesn't hold up-there's no business incentive for that. What I have observed is that the first 200-500 impressions act as a gatekeeper. If that seed audience doesn't engage strongly (think retention rates, not just watch time), the algorithm kills distribution fast.

The 400 vs 100k view gap? That's not randomness-that's a signal that the creative hook either resonated perfectly with the seed group or fell flat. Two identical videos can land wildly different results if the initial batch of viewers has different interests or attention spans. I've had a video die at 300 views, changed literally nothing except the hook frame, and watched it climb to 50k. That's the algorithm testing against different audience clusters, not rolling dice.

So no, it's not random-but it is incredibly sensitive to the first few seconds and the platform's read on who to show it to. Treat each upload like a fresh experiment, not a guaranteed outcome.

ryanseo

Been running the same test myself. Two identical videos, same audio, same hook, posted back-to-back. One hits 100k, the other flatlines at 400. Nothing in the retention data explains it. The algorithm is just chucking darts at a board at this point.

zoecreative

Hang on-was it literally the exact same file uploaded twice, or just two nearly identical videos built from the same template? Because that distinction is everything. TikTok's algorithm punishes straight duplication harder than people realise-it'll throttle the second one unless you swap out the hook or the first three seconds. Would love to see the links if you've still got them, I've been tracking how playback patterns shift between clones and near-clones and it's maddening.

ryanseo

I've been tracking the same pattern in the ASMR intimate audio niche for a while now. Looked at about 30 creators doing that type of content - it's almost identical across the board.

The bulk of their videos hover around 700-800 views, sometimes lower.
Then you get one or two outliers hitting 100k+, but that success never replicates.
New uploads slip straight back to the 700-view ceiling.

It's like the algorithm gives you a taste of viral reach, then slaps a ceiling on your account. no amount of optimisation seems to shift it - consistency, hashtags, sound quality.. doesn't matter. the niche itself might be too narrow for sustained growth, or TikTok's pushing that content into a specific sub-feed that caps exposure.

Would be interested if anyone's found a way to break that plateau without pivoting the content type entirely