The Curation Call: Auditing Yatra's Feed Fit and Reply Engagement

Marketing Academy · Field Work●Social Media Marketing

All projects·From: Bluesky & the Open Social Web

CoreAudit· 45 minutes

Yatra Online

Objective: Given a slate of draft posts and a week of reply-engagement numbers, decide which posts would survive human-curated feed inclusion and whether the account's reply investment is healthy or a vanity-metrics trap.

You're the social strategist at Yatra Online, the Nasdaq-listed Indian online travel agency (NASDAQ: YTRA), auditing the first month of a Bluesky pilot aimed at budget and mid-market travelers.

Yatra has five draft posts queued for a 'Slow Travel India' human-curated feed and one week of engagement data. Decide which posts would actually survive curation, and whether the reply numbers show healthy community engagement or a vanity-metrics problem.

Before you start

What you'll need

Free path (everything below is enough to finish)

Google Sheets

FreeLog post drafts, curation verdicts, and weekly reply-to-like ratios

Free, sufficient for weekly manual tracking on a pilot account

The process

2 steps

Step 01 of 02

Matching content to a feed's explicit curation rules

Each custom feed has curation rules, keyword matching, human curation, or engagement signals. Promotional content rarely survives human curation; educational, opinionated, or data-driven posts do.

Given five draft posts queued for the human-curated 'Slow Travel India' feed, which ones would a human curator actually include?

Google Sheets— A five-row content queue with a 'promotional vs. educational' column and a curator-fit verdict column.

Procedure

List the five drafts: (1) '20% off Goa packages this week', (2) 'How railway sleeper class actually works for first-timers', (3) 'We just launched a new app feature', (4) 'The 6 most underrated hill stations within a night train of Delhi', (5) 'Book now, prices rising Friday'.
Tag each as promotional or educational/data-driven based on the lesson's Step 3 framing.
Mark a curator-fit verdict: survives curation, or gets filtered out as promotional noise.
Rewrite the two flagged promotional posts into an educational frame before requeuing them.

Sample output

FEED FIT AUDIT: Slow Travel India (human-curated)

SURVIVES CURATION
  2. How railway sleeper class works for first-timers (educational)
  4. 6 underrated hill stations within a night train of Delhi (data-driven)

FILTERED OUT
  1. 20% off Goa packages this week (promotional)
  3. We just launched a new app feature (promotional, product news)
  5. Book now, prices rising Friday (promotional, urgency bait)

Healthy

3 of 5 drafts rewritten or held back before queueing, only genuinely educational posts sent to a human-curated feed.

Unhealthy

All five drafts queued as written, with promotional posts silently filtered out by curators and never explained to the team.

What this means

A human-curated feed is not a paid placement, it is earned inclusion. Posts that read as ads get filtered regardless of how well-targeted the audience is.

So what do I do about it?

Symptom	Action	Effort
Promotional drafts keep getting silently excluded from curated feeds	Run every draft through a promotional-vs-educational tag before queueing, not after	30 min

YouYou can do this yourself, no engineering access required.

Step 02 of 02

Reading reply-to-like ratio as the real amplification signal

A post with 50 replies and 10 likes signals more community health than a post with 200 likes and 2 replies, because the default Following feed surfaces conversation threads. Reply time is not optional, it is the primary amplification mechanism.

Yatra's top post this week has 340 likes and 4 replies. A smaller post has 28 likes and 31 replies. Which one is the actual distribution win, and what should next week's reply-time budget look like?

Google Sheets— A weekly post-performance tab with likes, replies, and reply-time-invested columns.

Procedure

Log both posts' like and reply counts side by side.
Compute reply-to-like ratio for each: 4/340 = 0.01 vs. 31/28 = 1.1.
Cross-check reply-time-invested: the high-like post got 0 minutes of team reply engagement, the high-reply post got 35 minutes.
Decide the following week's reply-time allocation based on which pattern actually built thread surfacing, not raw like count.

Sample output

REPLY ENGAGEMENT AUDIT, week 3

Post A: 340 likes, 4 replies, 0 min team reply time -> ratio 0.01
Post B: 28 likes, 31 replies, 35 min team reply time -> ratio 1.1

VERDICT: Post B is the real amplification win despite fewer likes.
NEXT WEEK: budget 30 min reply time per post, prioritize threads already showing early replies within the first hour.

Healthy

Reply-time budget scales with early reply signals, not with which post already has the most likes.

Unhealthy

Team spends reply time on the highest-like post regardless of whether it's generating conversation.

What this means

Likes are a passive signal Bluesky's algorithm barely rewards. Replies are the mechanism that surfaces content to a repliers' entire follower graph.

So what do I do about it?

Symptom	Action	Effort
High-like posts get all the team's attention while high-reply posts are ignored	Reallocate the 30-minute reply budget to whichever post shows early reply momentum, not highest likes	30 min

YouYou can do this yourself, no engineering access required.

Final deliverable

A feed-fit audit table for five draft posts plus a reply-engagement health readout with next week's reply-time budget recommendation.

See a reference example

Sample output

AWFIS SPACE SOLUTIONS, WEEK 2 FEED + REPLY AUDIT (excerpt)

FEED FIT
  Survives curation: 'How hot-desking actually changes your commute math' (educational)
  Filtered out: 'Sign up this week for a free day pass' (promotional)

REPLY HEALTH
  Post A: 210 likes, 3 replies -> ratio 0.01, low priority next week
  Post B: 19 likes, 22 replies -> ratio 1.16, reply-time budget reallocated here

Success criteria

You're done when you can:

Correctly separates promotional from educational drafts using the lesson's curation framework
Identifies the higher reply-to-like ratio post as the real amplification win, not the higher-like post