Launch Metrics

This skill covers how to measure product launches — what to set up before launch day, what to track during, and how to evaluate results afterward. Without consistent measurement, launches are guesswork. With it, each launch informs the next.

Pre-launch setup

Measurement starts before launch day. Setting up tracking the day of the launch means you miss the first-wave data when traffic is highest.

What to set up

1. UTM parameters — Tag every link in every distribution channel
2. Analytics events — Instrument the conversion funnel end-to-end
3. Baseline capture — Record current metrics before launch so you can measure lift
4. Dashboard or view — One place to watch all key metrics on launch day

UTM conventions

UTM parameters track where traffic comes from. Use a consistent naming scheme across all launches so you can compare over time.

utm_source    = the platform (twitter, linkedin, email, producthunt, hackernews)
utm_medium    = the channel type (social, email, paid, referral)
utm_campaign  = the launch name (feature-name-launch, v2-launch, etc.)
utm_content   = the specific asset (tweet-thread, announcement-email, blog-post)

Example URLs:

• Twitter announcement: ?utm_source=twitter&utm_medium=social&utm_campaign=realtime-api-launch&utm_content=launch-tweet
• Launch email: ?utm_source=email&utm_medium=email&utm_campaign=realtime-api-launch&utm_content=announcement-email
• Product Hunt: ?utm_source=producthunt&utm_medium=referral&utm_campaign=realtime-api-launch&utm_content=ph-listing

Rules:

• Always lowercase
• Use hyphens, not underscores or spaces, in values
• Campaign name should match across all channels for a single launch
• Create all UTMs from a spreadsheet before launch day so nothing is missing

Instrumented events

Before launch, verify these analytics events exist and fire correctly:

If these events don’t exist and you don’t have time to add them before launch, at minimum use UTMs to measure traffic and signups from whatever analytics platform you already have.

KPIs by tier

What “success” looks like depends on launch tier. Don’t compare a Tier 3 changelog tweet to a Tier 1 product launch.

Tier 1 KPIs (full product launch)

Tier 2 KPIs (major feature)

Tier 3 KPIs (small improvement)

Benchmark expectations

These are rough benchmarks for a B2B/developer tool with an established audience. Use them to calibrate expectations, not as hard targets.

These vary enormously by audience size, product category, and launch quality. Treat them as directional.

Launch day monitoring

On launch day, watch metrics in near real-time for the first 4-6 hours. This is when you can still adjust (reshare a post, send a follow-up tweet, extend a limited offer).

What to watch:

• Signups per hour (alert if significantly below expectation)
• Top traffic sources by signup conversion rate (not just volume)
• Email delivery issues (bounce rate, spam complaints)
• Social engagement velocity in the first 2 hours (predicts final reach)

Don’t do:

• Obsessively refresh metrics while neglecting to engage with social comments
• Make major messaging changes mid-launch based on early data
• Panic if a specific channel underperforms — wait for 24-hour data

Post-launch retro

Run the retro within 5-7 days of launch, when the data is fresh. For Tier 1, schedule this during launch prep. For Tier 2, a Slack thread async works.

Retro data to pull

Before the retro, gather:

• Traffic by source (UTM breakdown)
• Signups by source
• Conversion rates by source (signups ÷ visitors)
• Email metrics (open, click, unsubscribe)
• Social metrics by platform
• Comparison to previous comparable launch (if available)

Questions to answer

1. What was the top-performing channel (by conversion rate, not just volume)?
2. What was the worst-performing channel relative to effort invested?
3. What was surprising? (channels that over- or under-performed expectations)
4. Did messaging land? (qualitative: comments, replies, email responses)
5. Where did the funnel leak? (high traffic, low signups → landing page problem; low traffic → distribution problem)
6. What would you do differently?

Retro output

The retro should produce exactly two things:

1. Updated distribution priorities — which channels to invest more/less in next time
2. Checklist updates — additions or removals to the launch checklist based on what you learned

Document in a shared place (Notion, Linear, Confluence) so it accumulates over launches. The value compounds: after 10 launches, you have a quantitative model of what works for your product and audience.

Building a launch history

Track every launch in a single spreadsheet or database:

| Launch | Date | Tier | Top channel | Signups (D1) | Signups (W1) | Email open rate | Notes | |——–|——|——|————-|————–|————–|—————–|——-|

After 5+ launches, patterns emerge:

• Which channels consistently convert for your audience
• What tier of launch drives what magnitude of signups
• How your benchmarks trend over time (growth in list size, social following)

This history is your most valuable launch asset. It makes future launch planning faster and more accurate than any external benchmark.

Companion skills

• launch-strategy — defines the launch tier and checklist that determines what to track
• launch-distribution — each channel’s UTM is set up before launch day
• launch-email — email open/click metrics feed back into the retro
• launch-blog-post — blog post UTMs and organic search performance over time