Launch day traffic looks like a vertical line. Half of it may never be a human with a credit card. Product Hunt, press pickups, sitemap pings, and AI answer bots all show up in the same visitor count if you let them.
Part of marketing pulse guides. Start with detecting campaign spikes; pair with Product Hunt baseline pinning.
Two curves stacked on one day
Bot swarm: sharp, often minutes after listing goes live or sitemap updates; high request rate; common user-agents (Googlebot, preview fetchers, GPTBot, generic curl); near-zero checkout events.
Human spike: broader over hours; mixed referrers (producthunt.com, Twitter, newsletter); checkout or signup events correlate (with lag).
Your marketing pulse pin for launch should capture both; your interpretation splits them.
What triggers bot swarms on indie launches
- Product Hunt: scrapers, rank trackers, link unfurlers.
- Sitemap resubmit after launch, Googlebot bursts.
- Press syndication: aggregators fetch OG tags.
- AI crawlers: docs and marketing pages hit after mentions; see AI crawler guides.
- Security scanners: random, but spike on new subdomains.
None of these are “bad” necessarily. They are not campaign ROI.
Filtering without lying about reach
Reasonable founder approach:
- Exclude known bot user-agents in tracker config (KiboData and peers usually ship lists).
- Watch crawler dashboard separately from human visitors when docs are central.
- Do not subtract all unknown agents, mobile in-app browsers look weird too.
Goal: human line moves with launches; crawler line moves with indexing and AI, read both.
Pin windows for launch
Standard triple pin (from PH guide, generalized):
| Pin | Dates | Purpose |
|---|---|---|
| Pre-baseline | −14 to −1 | Normal noise |
| Launch | −1 to +6 | Full stress week |
| Post-decay | +14 to +27 | Hangover traffic |
Day 0 inside launch pin: note bot ratio if tooling exposes it. Day 2–5: often more human-shaped for digital products.
Humans on PH vs bots on PH
Tagged utm_source=producthunt visitors are still mixed, makers browsing vs buyers. Use sales webhooks inside the pin: Gumroad traffic-to-sales ratio beats raw visits.
If visits 5× and sales 1.2×, swarm or curiosity dominated. If visits 5× and sales 4×, launch worked, email the list harder week two.
Stripe SaaS launches
Trials and checkouts lag PH day, Stripe checkout on landing pages. Bot traffic does not create trials; humans do. Compare trial starts days 1–7 inside launch pin, not hour 1.
When the whole spike is bots
Signals:
- Revenue flat, signups flat, support inbox quiet.
- Traffic from datacenter ASNs or identical paths
/only. - Spike ends in <2 hours with no tail.
Action: verify DNS and firewall, check if you posted a raw API URL, confirm tracker not double-counting health checks. Do not run paid ads to “recreate” the spike.
Press launch vs PH launch
Press can trigger more crawler volume relative to humans than PH. Pin press-2026-02-techcrunch separately from ph-v2-2026. Blending them teaches wrong channel lessons.
Newsletter + launch same week
Common stack: PH Tuesday, email Wednesday. One 7-day pin blends channels, acceptable for “launch week” narrative. For attribution, split pins or rely on UTMs inside the shared week per newsletter send week.
Training crawlers vs real indexing
Training crawler noise can inflate week-one docs traffic without search impressions moving. Do not confuse with PH human spike on /.
Live dashboard adrenaline
Realtime globes count events fast, bots included. Use live view for morale; use filtered daily pulse + pins for decisions. Live launch guides is optional comfort, not accounting.
Worked split
Launch pin day 0: 8,200 “visitors” unfiltered.
After bot filter: 2,100 human-shaped, 6,100 crawler/preview (illustrative, your ratios vary).
Webhook sales day 0–3: 28 units vs baseline week 12 units.
Interpretation: Real lift; bots made the raw chart unreadable without filter.
Pin label note: ph-v2-filtered-humans.
Screenshot trap on launch day
Founders screenshot Plausible “1.2k online.” Archive pinned range with visitors, filtered count if available, sales, ratio. Screenshots miss the filter and the refund week after.
Security and DDoS masquerading as marketing
Rare on micro-SaaS, but traffic spikes without referrer and without bot UA patterns deserve ops review before marketing retrospective.
Post-launch email tail
Humans buy days 3–10 via newsletter after PH curiosity. Extend revenue pin to 14 days for digital products; keep traffic pin at 7 for “launch burst” only.
Comparing year-two relaunch
New pin slug every launch, ph-v3-2027. Bots do not remember your brand; your baseline should still be pre-launch quiet week, not last year’s pin merged.
UTM hygiene under swarm
UTMs on PH listing and first comment still help human slice. Bots often omit query strings. Low tagged % on launch day is normal; rising tagged % day 2+ is humans.
RPV under bot noise
RPV computed on unfiltered launch visitors looks terrible forever. Recompute inside filtered human visitors or inside sales ratio pins, honest denominator.
Where KiboData fits (no oversell)
KiboData provides marketing pulse (visitors + webhook revenue), pinned ranges, and an AI/crawler visibility slice for docs-heavy sites. Bot filtering depends on maintained UA lists, not perfect against bespoke scrapers. It does not run WAF rules or replace Cloudflare analytics.
Tracker script on marketing site; webhooks from Stripe/Gumroad/etc. Free tier + badge; lifetime Pro removes badge. Crawler charts help explain spikes; they do not block bots.
Checklist
- Launch pin + baseline pin created before listing goes live
- Bot filter enabled and reviewed day 0
- Sales webhook verified on production
- Separate press vs PH pins if both fire same week
- Revenue window longer than traffic window when needed
- One-sentence postmortem: human lift vs bot noise vs sales
Hub links
Main guide: spikes without screenshots · Twitter lag · /blog/topics/marketing-pulse
HN, Reddit, and forum scrapers
Show HN and subreddit launches trigger link expanders and mod bots. Traffic may arrive with referrer from news.ycombinator.com or reddit.com while a parallel crawler line hits / and /docs without referrer. Pin hn-2026-02 separately from PH when both happen, forum humans often convert differently from PH browsers.
Affiliate and deal-site bots
Coupon aggregators and “deal” bots fetch prices on launch. They inflate visits on /pricing without trials. If pricing page views 10× and signups 1.5×, check path breakdown inside the launch pin before you discount.
Monitoring and uptime checks
Founders add Better Stack or UptimeRobot on launch day. Misconfigured monitors pinging every minute look like a traffic spike. Exclude monitor IPs or paths if your host allows; otherwise note uptime-added in pin label so you do not confuse ops with marketing.
Geographic skew on launch day
PH and US press skew US traffic day 0; EU and APAC humans often peak day 1–2 in their local evenings. Bot swarms are less timezone-shaped, they follow the listing going live in UTC. A pin that ends at US midnight can chop off international human tail; use full 7-day launch pin for global products.
Webhook ordering vs visitor spikes
Pageview after Stripe webhook ordering rarely matters on launch hour, but same-day revenue attribution can look “early” if checkout webhooks arrive before the session is counted. Trust pin totals over hour-by-hour overlay when sample size is small.
Teaching co-founders and investors
When someone asks “how did launch go,” open marketing pulse with launch pin and baseline, filtered humans if available, sales line, one ratio. Avoid raw unfiltered visitor totals; you will defend bot traffic in a board update and lose credibility.
After the swarm: what to optimize
If humans and sales moved: fix onboarding, mail non-buyers, raise price later. If only bots moved: improve robots.txt and crawler policy for docs, not hero copy. If humans moved without sales: offer, checkout, or audience mismatch, bots were a red herring.
Year-one vs year-five product launches
First launch: almost everything is new, bots and humans both spike hard. Fifth launch on same domain: crawlers already know you; human lift may be smaller relative to bots because indexing is warm. Compare pin to your prior launch pin, not someone else’s PH screenshot on Twitter.
Launch day is supposed to look chaotic. Pins, filters, and sales in the same view turn chaos into a week you can learn from, without saving another screenshot.