The Weak-Signal Audit: Diagnosing Why a Post Underperformed

Marketing Academy · Field Work●Social Media Marketing

All projects·From: How Social Algorithms Work

MiniAudit· 30 minutes

Go Digit General Insurance

Objective: Given a real 5-post Instagram Reels export with completion rate, saves, shares, comments, and likes, apply the lesson's signal-weighting table to identify the algorithmically weakest post and prescribe the fix, instead of ranking posts by raw engagement.

You're the social media analyst at Go Digit General Insurance, the Bengaluru-headquartered, BSE/NSE-listed general insurer (543527), reviewing last month's Instagram Reels before the next content sprint.

Score each post against the lesson's signal-weighting table (saves highest on Instagram, likes weakest everywhere), find the post with strong likes but weak saves/completion, and recommend the specific fix.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Sheets

FreeImport, score, and re-rank the post export

Free, no account friction, sufficient for a 5-row comparison

The process

1 step

Step 01 of 01

Completion rate, saves, and shares outrank likes as ranking signals

The lesson's Content Signals table shows likes as the weakest signal on every platform, while saves are Instagram's strongest and completion rate is Instagram's second-strongest.

Post 3 has the most likes (2,400) of the 5 posts but the lowest completion rate (18%) and almost no saves (12). Is it the strongest post?

Google Sheets— Import the 5-post export, freeze the header row, add a computed 'algorithmic strength' column.

Procedure

Import the export: post_id, likes, comments, shares, saves, completion_rate
Rank posts by likes alone first, note the order
Re-rank posts weighting saves and completion rate highest, likes lowest, per the lesson's table
Compare the two rankings; flag any post where the rankings diverge sharply

Sample output

POST PERFORMANCE, RAW LIKES RANK
  1. Post 3 -- 2,400 likes, 18% completion, 12 saves
  2. Post 1 -- 1,900 likes, 61% completion, 340 saves

RE-RANKED BY ALGORITHMIC SIGNAL WEIGHT
  1. Post 1 -- 340 saves, 61% completion (STRONG)
  2. Post 3 -- 12 saves, 18% completion (WEAK, despite likes)

Healthy

A post with high saves and completion rate, even with modest likes, gets flagged as the strongest performer.

Unhealthy

The content calendar keeps repeating Post 3's format because its like count looks impressive on a screenshot.

What this means

Likes are a vanity signal the algorithm barely weighs; saves and completion predict future distribution far better.

So what do I do about it?

Symptom	Action	Effort
A high-like, low-save post gets scheduled for a repeat next month	Re-score the content calendar using saves and completion rate before reusing any format	30 min

YouYou can do this yourself, no engineering access required.

Final deliverable

A one-page signal diagnosis memo ranking all 5 posts by algorithmic strength (not raw engagement), with a specific fix recommended for the weakest post.

See a reference example

Sample output

Acko General Insurance, Reels signal audit (excerpt)

STRONGEST: 'Claim in 3 taps' demo -- 410 saves, 58% completion, 890 likes
  Why: viewers wanted to reference it later; hook lands in first 2 seconds.

WEAKEST: 'Meet our new mascot' -- 2,100 likes, 9% completion, 6 saves
  Fix: the mascot reveal has no reason to save or rewatch; rebuild around a problem the viewer can solve, not a brand moment.

Success criteria

You're done when you can:

Correctly re-ranks all 5 posts by save rate and completion rate, not likes
Identifies the specific post where the like-based and signal-based rankings diverge
Recommends a concrete content fix, not a generic 'post more' suggestion