Social Media Analytics Checklist

List the platforms in scope this cycle — typically Meta (Facebook + Instagram), LinkedIn, TikTok, X, and YouTube. Note any campaign-specific channels (Reddit, Pinterest) that were active. The social media manager owns this list; the demand-gen lead confirms paid channels are included.

Spot-check 10 published posts from the period: source / medium / campaign should match the convention doc. Drift here is the #1 reason monthly numbers don't reconcile to GA4. Flag any rogue UTMs to the channel owner before pulling reports.

Export from Meta Business Suite, LinkedIn Campaign Manager, TikTok Analytics, and X Analytics for the reporting window. Native exports are the source of truth for impressions, reach, and follower growth — third-party tools sometimes lag platform API changes by a week.

Use the saved GA4 exploration filtered to Source/Medium = social. Confirm the conversion event mapping matches what the lifecycle team uses — a misfiring sign-up event will distort attributed conversions and push budget to the wrong channel.

Sprout (or Hootsuite / Sprinklr if that's your stack) gives the consolidated engagement and inbound-message view that native exports don't. Pull the same date range; mismatched windows are a recurring reconciliation headache.

One creative posted to Instagram, Facebook, and LinkedIn shows up as three rows. Group by creative ID or canonical post URL so engagement totals don't get triple-counted in the cross-channel view.

Remove repeat-emoji-only comments and known spam handles from engagement totals. A single bot wave can inflate engagement rate by 5-10% on a niche post and skew the top-performer list.

Boosted posts inflate organic numbers if not split out. Tag each row as paid, boosted-organic, or pure organic — the demand-gen team needs the split to compute organic CPM and ROAS independently.

Platform-reported conversions almost never match GA4 — different attribution windows (7-day click vs. last non-direct), different conversion definitions. Document the gap; flag any single-channel discrepancy over 25% to marketing ops.

Apply the brand's content pillar taxonomy (e.g., thought leadership, product, customer story, recruiting). Pillar-level performance is what brand and PMM actually use to plan next quarter's calendar.

Use the Sprout listening view or Brandwatch / Talkwalker pull, filtered to brand and product mentions. Compare positive / neutral / negative ratios to the trailing 3-month baseline — a drop of 10+ points in positive sentiment is the trigger for the brand-safety branch below.

Sentiment dropped — pull the negative-mention cluster, identify whether it's a single PR incident, an ad-creative complaint, or a sustained drift. Loop in PR and the brand manager same-day; do not wait for the monthly review meeting.

Use engagement-rate-by-reach (not by followers) — followers count is a vanity denominator on platforms with algorithmic feeds. Break out by Reels, carousels, single image, and link posts; the format mix is usually the biggest driver of month-over-month variance.

Top 10% and bottom 10% by engagement rate, tagged by pillar and format. The patterns in these two lists drive the calendar adjustments in the monitoring section — what to repeat, what to retire.

Net follower change broken out by acquisition source (organic, paid, profile-visit-from-post) and by audience demographic where the platform exposes it. Drift away from the ICP's demographic profile is a leading indicator that creative is targeting wrong.

Use the standing template — exec summary, KPI scorecard vs. targets, channel breakdown, top / bottom posts, sentiment trend, recommendations. Every chart caption should answer 'so what' for the reader, not just label the axes.

Update the Looker (or Tableau / GDS) dashboard so leadership can self-serve between reviews. Confirm filters default to current-month and that the comparison column is set to prior month, not prior year.

Distribute to the CMO, demand-gen lead, PMM, and the brand manager 24 hours before the review meeting so attendees come prepared with questions, not first reactions.

45-minute working session: walk the scorecard, debate the recommendations, leave with a list of decisions. The social media manager runs the meeting; the demand-gen lead and PMM are required attendees.

Document each decision, the owner, and the due date. Anything left in Slack DMs gets dropped — paste into the running review log so next month's report can show whether last month's adjustments worked.

Walk anything that moved more than two standard deviations from the trailing 3-month mean. Common root causes: a viral post skewing engagement rate, a tracking-pixel misfire, a paid campaign budget change, a platform algorithm update.

A significant anomaly needs a written brief to the CMO and demand-gen lead within 24 hours: what moved, suspected cause, blast radius, and the corrective action being taken. Don't wait for next month's deck.

For each paid campaign, log spend, ROAS, CPA, and payback period against the target set at campaign launch. Campaigns under target for two consecutive months get reviewed for reallocation.

Translate the top / bottom decile patterns into next month's calendar — increase formats and pillars that are working, retire what isn't. The content manager owns the edit; the social media manager confirms the rationale ties to the report.

Pull the named competitor set from Sprout / Brandwatch listening — share of voice, engagement rate, post cadence, content-pillar mix. Quarterly is fine for the deep cut; the monthly version is just SoV trend.

Confirm the next run's date range, the review meeting slot on calendars, and any methodology changes (e.g., new pillar taxonomy, added channel) so they're applied from day one of the next pull.