Non-follower R&D layer for creative testing. Shown to strangers first, bypasses connected ranking.
No Bali competitor uses Trial Reels.
40% of creators posted Reels more often after trying Trial Reels; of those, 80% saw increased non-follower reach.[3]
Source: Instagram official stat; Bali competitor landscape audit
What Is This?
Trial Reels are a toggle-on mode when posting a Reel. The content is shown to non-followers first via unconnected recommendations, bypassing your connected audience's ranking entirely. It is designed for creative R&D — testing new ideas, aesthetics, or formats without confusing existing followers.[1][3]
Available to public accounts with 1,000+ followers (per Mosseri, July 2025)[1]
Not shown on your main profile grid or Reels tab to followers
Reported daily cap of ~20 trials/day (widely reported, not officially confirmed)[6]
Auto-shares to your full audience at ~72 hours if it performs well (threshold opaque)[1]
Can manually "share with everyone" to graduate to a normal Reel at any time
Cannot be a collab post (Metricool reports; not officially confirmed)[6]
Cannot be boosted while in trial — boost only after graduating[4]
How It's Distributed
Trial Reels are distributed exclusively through unconnected recommendations — shown to people who do not follow you. Mosseri has stated that trials typically get less reach than regular Reels, not more.[2]
The promotion threshold (when Instagram auto-shares to your followers) is opaque — no published criteria
Initial signals read at ~24 hours: views, likes, comments, shares[1]
Followers may still encounter trials via DM shares, same-audio pages, or location/filter pages[1]
Compare trial-to-trial, not trial-to-regular. Trials have a fundamentally different distribution path, so raw reach numbers are not comparable[2]
Current Benchmarks
All available benchmarks are creator-scale. No product brand or furniture brand benchmarks exist for Trial Reels.
Case study data
Source
Result
Context
Kapwing
1.4M views, 130 follows, 8.6K profile visits, 290 link taps
Single outlier Reel. Follow-up controlled test: trend Reel did better as trial, product Reel did better as regular.[4]
Buffer
Trending audio: 24% more views than voiceover, 33% more than ASMR
Same video, 3 audio treatments tested via Trial Reels[2]
All case studies are creator-scale accounts. No product brand benchmarks exist — The Medium would be generating its own baseline.
What Content Works Best
Kapwing's controlled test revealed a clear split:[4]
Content type
Better as...
Why
Trend-based content
Trial Reel
Trends are designed for cold audiences who recognize the format instantly
Product content
Regular Reel
Product context requires familiarity with the brand — followers convert better
Multi-aesthetic testing (key strategic use)
Instagram explicitly says Trial Reels are for trying new genres and topics without confusing followers.[1] For a multi-aesthetic brand, this means:
Test dark primitive vs light modern with same product, same length, same CTA, same structure — only the aesthetic changes
Judge by watch time, shares, saves, follows per 1K views — not raw views alone
This tells you which aesthetic attracts strangers, not which one current customers prefer
How The Medium Should Use It
Wide-open opportunity
No Bali competitor uses Trial Reels. This is an uncontested testing surface in the market. Use it to test aesthetic directions without confusing existing followers, and build a data-backed creative strategy before competitors even start.
Cadence
3-5 Trial Reels per week. Minimum 6-12 trials before making strategy decisions. Each round should test one hypothesis with one variable changed.[1]
What to test
Aesthetic directions: dark primitive vs light modern — which attracts cold audiences?
Hook styles: product-first vs environment-first vs maker-story-first
Audio treatments: trending audio vs voiceover vs ambient/ASMR
Content angles: trend-based (use as trial) vs product-focused (keep as regular)
Competitive gap
Recommendation: run a competitive gap analysis on Trial Reel adoption across the Bali design/furniture segment. If no competitor is using it, The Medium can build a creative testing infrastructure with zero competitive noise on this surface.
Execution Playbook
7-step testing methodology
Start with one hypothesis per round — "dark aesthetic gets more shares from strangers" or "trending audio outperforms voiceover"
Change one variable per test: hook, content type, aesthetic, CTA, or posting time
Follow the test order: hook first → content type → aesthetic → caption/CTA → timing
Score trials against other trials, never against regular Reels
Wait 72 hours for a real read (if auto-share is enabled)
Minimum sample: 3 trials per variant, or 6-12 total before making strategy decisions
Publish 3-5 Trial Reels per week to maintain testing velocity[1]
Weighted scoring scorecard
Signal
Weight
Why
Retention / average watch time
40%
Strongest signal of content quality for cold audiences
Intent signal — saves correlate with consideration
Profile visits or follows per 1,000 views
10%
Measures whether the content drives curiosity about the brand
Comment quality
10%
Qualitative — are comments from potential buyers or just engagement bait?
Scorecard designed for product brands. Score each trial, compare trial-to-trial, and look for patterns across 6-12 data points before drawing conclusions.[1]
72-hour read window
Trials show initial signals at ~24 hours but the real read comes at 72 hours, when the auto-share decision is made. Do not make creative judgments before the full window closes.[1]
Cross-Surface Flows
Trial → graduate → amplify pipeline
Trial Reel — test with non-followers, read signals over 72 hours
Graduate — auto-shared or manually shared with full audience as a regular Reel
Regular Reel — now visible on profile grid and to followers
Boost — promote the graduated Reel with ad spend (cannot boost while in trial)[4]
Stories / Highlights — reshare the graduated Reel to Stories for additional reach and DM triggers
Gap: flow not formally mapped
The trial → graduate → regular Reel → boost/Stories/Highlights pipeline is implied by the mechanics but has not been formally mapped or tested as a complete workflow. This needs validation.
Pitfalls
Pattern
What Goes Wrong
Comparing trial to regular reach
Trials get less reach by design. Comparing raw numbers to regular Reels will make every trial look like a failure. Compare trial-to-trial only.[2]
Assuming follower firewall is airtight
Followers can still see Trial Reels via DM shares, same-audio pages, and location/filter pages. It is not a complete separation.[1]
Expecting more reach than regular
Mosseri has stated trials usually get less reach, not more. The value is in testing with cold audiences, not in reach volume.[2]
Trying to collab post as trial
Trial Reels cannot be collab posts (Metricool reports). Plan collab content separately.[6]
Trying to boost a trial
Cannot boost while in trial. Graduate the Reel first, then boost the graduated version.[4]
Reading results too early
Initial signals at 24 hours are incomplete. Wait for the full 72-hour window before making creative judgments.[1]
Open Questions
UI walkthrough needed: No step-by-step walkthrough of how to toggle Trial Reels on in the app, or what the analytics screen looks like. This should be documented with screenshots.
Analytics reading needed: How to read Trial Reel analytics vs regular Reel analytics. Which metrics are available, and which are missing?
4-week test plan: A planned structured 4-week test plan for The Medium was discussed but never written. Needs to be created.
Furniture-brand examples needed: All existing case studies are creator-scale accounts. No product brand or furniture brand has published Trial Reel results. The Medium would be generating original data.