What 11.4M TikToks Reveal About Posting Frequency - The Sweet Spot for Per-Post Views

Executive Snapshot

• Posting 2–5 times per week correlates with 17% more views per post vs 1 per week.
• 6–10 per week correlates with a 29% lift; 11+ per week correlates with 34% vs 1 per week.
• The steepest per-post gain is moving from 1 to 2–5 per week. Going from 2–5 to 6–10 adds 12 percentage points, indicating diminishing returns.
• Median views per post remain roughly flat near 500 regardless of frequency (489 at 1/week vs 459 at 11+/week).
• 90th-percentile views rise sharply with frequency: 3,722 (1/week) → 6,983 (2–5) → 10,092 (6–10) → 14,401 (11+). The p90/median ratio climbs from 7.6× to 31.4×.

Implication for marketers: Start around 2–5 TikToks per week to capture most per-post view gains. Additional volume primarily increases the chance of a breakout post rather than shifting the typical outcome.

TikTok posting frequency: per-post views vs weekly cadence

Drawing on platform-connected accounts over the past year, Buffer’s panel analysis compares each account during periods of different posting cadences. The study reveals that posting 2-5 times per week provides the largest per-post lift (+17% vs 1/week), with smaller incremental gains at 6–10 (+29%) and 11+ (+34%). Median views per post stay near ~500 across frequencies, while the 90th percentile climbs substantially, indicating a heavy-tailed distribution in which a small share of posts drives outsized reach.

A technical note by the study’s author explains the fixed-effects approach used to isolate within-account changes and reduce bias from account size, niche, or brand strength. This aligns with broader evidence that user-generated content view distributions are heavy-tailed across social platforms.

Methodology and source reliability

What was measured: Per-post TikTok views segmented by weekly posting frequency bands (1; 2–5; 6–10; 11+), plus distributional metrics (median, 90th percentile, and p90/median ratio).

Sample: 11.4 million posts from 150,000+ Buffer-connected TikTok accounts over the prior year.

Approach: Account-level fixed effects to compare each account to itself during different cadence periods, mitigating cross-sectional confounders such as follower count, niche, or brand equity.

Caveats: Results reflect Buffer-connected accounts that skew toward SMBs and may not generalize to celebrities or large media brands. The analysis is observational, not causal. Time-varying factors like creative changes, promotions, and seasonality may still influence outcomes. The metric is views per post, not total reach, watch time, engagement rate, follower growth, or conversions. Platform dynamics evolve, so findings may shift as algorithms change.

Diminishing returns as posting increases

Relative to 1 post per week, per-post views lift by +17% at 2–5/week, +29% at 6–10/week, and +34% at 11+/week. The largest efficiency gain is the first step from 1 to 2–5/week. The marginal improvement from 2–5 to 6–10 is +12 percentage points, then just +5 points from 6–10 to 11+. Because the model compares each account to itself, these lifts reflect within-account changes rather than differences between accounts.

Practically, once a moderate cadence is reached, additional posting mainly expands exposure opportunities rather than improving the typical per-post outcome. Total weekly views can still rise with more posts, but the average performance per item flattens.

Median stability and viral potential

Median views per post remain nearly constant across frequencies: 489 at 1/week vs 459 at 11+/week. In contrast, 90th-percentile performance increases materially with cadence: 3,722 (1/week), 6,983 (2–5/week), 10,092 (6–10/week), and 14,401 (11+/week). The p90/median ratio expands from 7.6× to 31.4×, quantifying that higher cadence raises the ceiling without raising the floor.

This pattern is consistent with heavy-tailed dynamics: sampling more posts increases the probability of hitting the long tail of the distribution, where a few posts attract disproportionate attention even when the central tendency does not move.

Implications for cadence, resourcing, and measurement

• Likely: Establish a baseline of 2–5 TikToks per week to capture the largest per-post efficiency gain while managing production costs.
• Likely: Track both per-post and total weekly or monthly reach. If the goal is aggregate reach or testing velocity, higher posting may still be warranted despite diminishing per-post returns. If the goal is to raise the “typical” post, frequency alone is unlikely to move the median.
• Tentative: Allocate creative cycles toward experiments that increase the chance of tail events - e.g., varied hooks, formats, and topics - and report p90 or p95 alongside median to reflect the distribution.
• Tentative: For small teams, prioritize consistent 2–5/week output with rigorous iteration rather than daily quotas that strain quality.
• Speculative: Brands with strong distribution may see different thresholds. Validate locally with controlled cadence tests, especially for large publishers or celebrity accounts.

Contradictions, edge cases, and gaps

• Metric scope: The study covers views per post only. It does not quantify total weekly reach, follower growth, watch time, engagement rate, or revenue by cadence.
• Sample bias: Buffer-connected accounts likely skew SMB. Effects may differ for large creators, celebrities, or media companies.
• Causality: Even with fixed effects, time-varying factors like creative shifts, paid boosts, and seasonality can co-move with cadence.
• Content mix: The model does not segment by vertical, region, or format. Some categories may respond differently.
• Temporal sensitivity: Results reflect the past year and may change as TikTok’s algorithm evolves.
• Tail parameters: Heavy-tailed behavior is consistent with broader research, but exact tail parameters for TikTok were not estimated here.

Sources

• Buffer - How Often Should You Post on TikTok? (11.4M posts, 150k+ accounts): buffer.com/resources/how-often-should-you-post-on-tiktok/
• Julian Winternheimer - Technical note on fixed-effects model and results: symphonious-shortbread-cd2770.netlify.app
• Search Engine Journal coverage: searchenginejournal.com
• Borghol et al., The Untold Story of Video Popularity in YouTube (WOSN 2012): dl.acm.org/doi/10.1145/2342549.2342551