What happened, why it happened, and what to do about it.
⚠ Feed posts only. 808 feed posts (Reels, Carousels, Photos) across 2024–2025. Stories excluded — Meta changed how story impressions are reported mid-2025, making comparisons unreliable.
The decline
2025: a year of consistent decline across every metric
Interactions −80% · saves −95% · net followers now negative · indexed from Q1 2024
Interactions fell to 20% of their 2024 peak, saves to just 5%. Both fell far faster than post volume (53%) — which means publishing less alone doesn't explain the scale. Something about the quality of each post also deteriorated.
↓
two separate forces explain it
Why
Two things happened simultaneously — and compounded each other
Posts −42% accounts for ~half · algorithm + content quality explains the rest
Factor 1 · ~50% of the decline
We posted less
10.1 posts/week in 2024 → 4.7 by end of 2025. Fewer posts = fewer distribution events. Straightforward.
Factor 2 · ~50% of the decline
Each post performed worse
→
Algorithm change (spring 2025) — Instagram raised the bar for Reel distribution. Reel reach halved from 76K → 33K per post almost overnight. The platform started rewarding early engagement signals — watch time, shares, saves — more aggressively. See top-right chart.
→
Content quality compounded it — weaker content gets penalised harder by the new algorithm. As the mix shifted toward promotional and community posts (low saves, low shares), the algorithm distributed them less. Technique posts — the format that generates saves and early shares — nearly disappeared from the mix. The result: less content, distributed less far, to less engaged audiences. See bottom charts.
Posts per quarter
Cadence fell every quarter from Q2 2024 — across all formats
Median Reel reach per post
Halved in spring 2025, then flatlined at ~33K
Saves per quarter
534 → 53 saves/post · unaffected by algorithmic reach
Content type mix per quarter
Technique posts drove 53–81% of all saves in 2024 — then fell from 36 → 8/quarter
TechniqueDrillsCommunityProduct/App
↓
both are addressable
Response
Both causes are controllable — and independent of the algorithm
Post more · shift mix toward technique · the algorithm hasn't got worse since Q2 2025
Lever 1
Post more
At current per-post quality, doubling posts/week would roughly double total interactions and reach. Getting back to 8–10 posts/week (the 2024 pace) would recover the ~50% of decline that volume explains. This is the fastest lever — no creative change needed.
Lever 2
Make it technique-led
Technique posts generated 53–81% of saves in 2024 and drive the early engagement signals the algorithm rewards most. Rebuilding saves rebuilds algorithmic distribution, which rebuilds reach. This is the quality lever — harder to sustain, but addresses the other ~50% of the decline.
⚠ Note: posting more promotional content at higher volume would not help. Product/app posts have never generated saves or strong engagement, regardless of volume. The content type matters more than the quantity alone.
2026 · W2–W11 · Jan–Mar 15
2026 so far.
11 weeks in.
↓
and in 2026?
2026 so far
11 weeks in — volume holding, engagement yet to recover
W2–W11 · 4.7 posts/week · technique content returning in W10–W11 · algorithm stable at ~30K Reel reach
Dashed line = Q4 2025 weekly average. W7 (Dyno Reel) is the outlier that shows the channel's ceiling. W11 is the second-highest week and the most recent signal that technique-led content still works.
Posts per week · 2026
W2–W11 · all formats · Q4 2025 avg = 4.7/week
Content type mix · 2026
Stacked by category · W2–W11
TechniqueDrills & TrainingCommunityProduct/App
Weekly interactions · 2026
Q4 2025 avg = 2,876/week
Weekly saves · 2026
Q4 2025 avg = 253/week
Weekly total reach · 2026
Q4 2025 avg = 139K/week
Median Reel reach per post · 2026
Q4 2025 median = 32,743
2026 annualised pace vs 2024 peak and 2025
Metric
2024 peak
2025
2026 pace
vs 2024
Posts published
462
304
~244
−47%
Total interactions
786K
357K
~173K
−78%
Total reach
25.2M
12.5M
~7.6M
−70%
Total saves
263K
79K
~45K
−83%
Net new followers
+53,055
+9,169
negative so far
net negative
Signs of improvement: W10 and W11 both had technique-led content, and both outperformed the Q4 2025 average on interactions and reach. The algorithm hasn't deteriorated further. Post volume at 4.7/week is still 47% below the 2024 peak of 8.9/week — the Spring/Summer outdoor season is the natural moment to push volume and lean into the content formats that work.
Goals & post performance · W/E Mar 15, 2026
How the channel is performing against each goal.
Four goals, four scorecards — then every 2026 post ranked against them. Click a scorecard to expand the detail.
Goal scorecards
Four goals · click any card to expand the detail
01
Reach & Audience Growth
▼
−825 followers7M reach pace
2.3/5
Net new followers
−825
First net-negative year. Was +53K in 2024, +9K in 2025.
1/5
Reel reach per post
~30K median
Stabilised after halving from 76K peak (Q4 2024). Not getting worse, but 60% below peak.
3/5
Best-ever share count
3,134 shares
Feb 9 Dyno Reel. When the content earns it, organic spread is strong.
4/5
Total reach (2026 pace)
7M/year
Down from 25M peak — fewer posts + algorithm halving Reel distribution in early 2025.
2/5
The channel is losing followers for the first time while total reach is 72% below its peak. Per-Reel reach has stabilised at ~30K — volume is now the most controllable lever. The Dyno Reel proved the channel can still distribute widely when the content earns it.
02
Keeping the Audience Engaged
▼
Saves −90% from peak100% positive sentiment
2.5/5
Saves per post (Q4 2025)
53
Down from 534 at peak — a 90% drop. Clearest quality signal; can't be distorted by reach.
1/5
Interactions per post (2026)
662 avg
Marginally better than Q4 2025 (603) but well below 2024 median of ~820.
3/5
Audience sentiment
100% positive
4,682 earned mentions with zero negative sentiment across the entire dataset.
5/5
The audience that engages is loyal — but saves collapsed 90% as technique content disappeared from the mix. Each post became less worth keeping.
03
Credibility & Trust
▼
228K followersBosi · Rainer · Le Nevé
3.7/5
Follower count
228,586
Strong for a niche climbing training account. Immediate credibility to a first-time visitor.
4/5
Ambassador quality
Bosi · Rainer · Le Nevé
World-class athletes. Anyone who recognises the names is immediately convinced.
5/5
Earned mentions
4,682 · all positive
+407% growth since 2023. Community talks about Lattice positively.
4/5
Posting cadence
4.7 posts/week
Down from 10.1/week peak. Technique posts fell to 13% of output — recent feed doesn't reflect coaching expertise.
2/5
Static signals (228K followers, world-class ambassadors, 100% positive mentions) are strong and slow to decay. Dynamic signals (recent post quality and cadence) are weaker — a first-time visitor sees a lot of app promotion rather than coaching expertise.
04
Converting Followers to Customers
▼
Cannot score · no UTM tracking
?/5
⚠ Cannot be scored. No UTM tracking on the bio link means zero data on how many app signups come from Instagram. Add ?utm_source=instagram&utm_medium=bio to the bio link URL — 30 minutes, starts generating data within days.
App customers (all-time)
8,805
Source unknown — could be IG, search, word of mouth, or any combination.
W10 trial → subscription rate
33%
Down from 45% peak in W1 — conversion is steady but not recovering.
W10 rate
IG / app timing correlation
Visible but unproven
Higher-reach weeks tend to see more signups — but no causal data to confirm.
Two separate declines. Customer volume down 64% from W3 peak (1,444 → 527). Trial rate fell from 21–23% to 14–15%. Trial→subscription holding at ~29–33%, down from 45% peak in W1.
2026 posts · scored against four goals
Scores 1–5 per goal · click any post to see the reasoning · Conversion excluded — no UTM tracking on bio link so it cannot be measured Audience = reach vs 23K median · Distribution = reach + shares vs 43-share median · Engagement = interactions + saves vs 468/85 medians · Social Proof = interactions + shares (visible credibility)
Every category is dominated by the same format — specific climbing problem, fix demonstrated on camera. The formula hasn't changed. The posting frequency has.
The pattern is consistent. Every top post by interactions is a Reel with a specific technique problem and a named coach or athlete demonstrating the fix. The top posts by saves are drill-based content people bookmark to use in training. The all-time share record is a 2026 post — the format still works when the content earns it.
Spring/Summer 2026 · W/E Mar 15
What to try over the next three months.
Spring and Summer are the highest-intent period for climbers — outdoor season, goal-setting, fitness. That makes it the best window to test the things the 2025 data suggests. These are grounded in the dataset, not guesswork.
Where 2025 ended
−83%
interactions vs Q4 2024 peak. Saves −95%. Followers net negative for the first time.
Where 2026 is now
4.7/wk
posts per week. Technique content returning in W10–W11. Algorithm stable — ~30K Reel reach.
The opportunity
Spring/Summer
Highest-intent climbing period. Outdoor season content aligns with what the data says works.
Make technique-led outdoor content the backbone of Spring publishing
The outdoor season is the single best content opportunity of the year. Climbers transitioning from indoor to outdoor are actively searching for technique help — footwork on real rock, reading routes, managing skin, projecting. This maps directly to the content format that historically drives the most saves and shares (the Feb 9 Dyno Reel, the Mar 12 drop-knee Reel). A series of 6–8 outdoor transition Reels published across April–May would be a direct test of whether technique content at volume moves the metrics.
Why: Technique posts drove 53–81% of all saves in 2024 and fell to near-zero by Q4 2025 as the content mix shifted. W10 and W11 show saves recovering as technique content returned. The algorithm is stable — distribution isn't the problem, content type is.
02
01 Reach & Audience
Increase posting cadence toward 7–8 per week for a sustained 6-week block
The data is clear that roughly half the 2025 decline was simply fewer posts. Current pace is 4.7/week against an 8.9/week peak. The question is whether the team has capacity to sustain higher volume. Spring is the right time to test it — outdoor shooting creates natural content opportunities. Even a 6-week block at higher volume would produce enough data to see whether the volume lever actually moves the metrics, before committing to it year-round.
Why: At current per-post quality, each additional Reel adds ~30K reach and ~400 interactions based on 2026 medians. Going from 4.7 to 7 posts/week would roughly add 70–80K reach and 900 interactions per week — equivalent to the difference between a below-average week and an average one.
03
04 Conversion02 Engagement
Tie LatticePlan CTAs directly to technique Reels — not standalone app posts
The post data shows that standalone app promotion generates almost no saves or engagement, while technique Reels generate the highest save rates. Rather than publishing separate promotional content, adding a single CTA at the end of high-save technique Reels — "if you want a structured plan for this summer, link in bio" — puts the app in front of the exact audience that is already engaging with training content. No separate promotional posts needed; the CTA rides the distribution of content that already performs.
Why: The conversion goal cannot currently be measured (no UTM tracking), but technique post audiences have the strongest training intent signal in the dataset. Adding UTM tracking to the bio link at the same time would finally make it possible to measure whether this correlation is real.
Use ambassador content as amplification, not as a substitute for technique content
The 2025 content mix shifted toward athlete/ambassador posts as technique content declined. But ambassador posts (testimonials, sends, shout-outs) consistently generate lower saves and shorter-lived engagement than technique posts — even when reach is comparable. The stronger pattern is ambassador content that demonstrates technique: the Will Bosi hangboard Reel (4/5 score), the Matt Mendes bouldering Reel (4/5). Athletes performing technically interesting things outperform athletes being congratulated. Spring gives plenty of opportunity for this — outdoor sends, specific moves, real rock technique.
Why: Community/ambassador posts make up 55–70% of the current 2026 mix but generate a fraction of the saves. The channel's credibility (3.7/5 on the Expert goal) rests partly on the quality of ambassadors — but that credibility is better demonstrated through what they do than what they say about Lattice.
05
04 Conversion
Investigate the trial rate drop before the Summer acquisition window
The trial rate dropped from 21–23% in January to 14–15% by March. Summer is likely a stronger acquisition period (more people climbing, more people considering training plans), which makes it a poor time to have an unresolved onboarding problem. If something changed in the signup flow or first-use experience, identifying it now — before the Summer peak — would meaningfully increase the conversion value of whatever Instagram growth happens over Spring.
Why: RevenueCat data, W1–W10 2026. At the January trial rate, current weekly customer volumes would yield ~40% more trials. The problem is independent of Instagram performance — fixing it is a multiplier on any growth the channel drives.
06
04 Conversion
Add UTM tracking to the bio link — 30 minutes, changes everything
Currently there is no way to know whether Instagram drives any app signups at all. Adding ?utm_source=instagram&utm_medium=bio to the bio link URL takes around 30 minutes. By the end of the Spring/Summer period, there would be 12–16 weeks of data showing a direct Instagram→app conversion rate. That number would make every future conversation about the channel's commercial value a data conversation rather than a guess.
Why: 11 weeks of app data with zero source attribution. Every discussion about what Instagram contributes to the business is currently inference. This is the simplest gap to close.