Baton is hiring a Head of Performance Marketing to spend "a multi-million-dollar, always-on daily budget", with a data scientist to run "landing-page strategies". But Google's Chrome data shows pump.fun's homepage failing all three Core Web Vitals for real visitors: the main content takes 4.2 s, taps take 467 ms to respond and the layout shifts by 0.32 (the limits are 2.5 s, 200 ms and 0.1), and loading has got 20% slower in a year. On a mid-range phone in Google's lab test the main image arrives after 28.7 s, behind 3.4 MB of JavaScript. Paid clicks that land on that page will bounce before they see a coin.
First action
Reserve fixed heights for feed cards and re-run PageSpeed.
How to measure it
Paid landing bounce rate
- Now
- –
- Target
- -25%
- By
- 1 quarter
Steps
Baseline
Core Web Vitals by device and country, plus paid bounce rate.
Staff Data Scientist - Growth & Marketing 2 days
Stop the shifts
Fixed-size slots for live feed cards and images.
Staff Frontend Engineer 3 days
Server-render the first screen
Top coins and images sized, no client fetch first.
Staff Frontend Engineer 1 week
Split the bundle
Trading and chart code on demand; drop about 900 KB unused.
Staff Frontend Engineer 1 week
Defer tags
Ads and analytics tags after first paint.
Staff Frontend Engineer 2 days
Paid landing route
A light page per campaign intent.
Product Designer 2 weeks
A/B test
Homepage vs landing route on cost per first trade.
Staff Data Scientist - Growth & Marketing 3 weeks
Monitor
Chrome UX Report monthly; alert on regressions.
Staff Frontend Engineer monthly
Why it works
Treat speed as part of the paid-acquisition launch. Give the live feed fixed-height card slots so new posts stop pushing the page around, render the first screen of coins from the server with sized images, split the trading and charting code so it loads only when a visitor opens a coin, and defer tags that do not need to run on first paint. Then build a lightweight landing route for paid traffic and A/B test it against the homepage on cost per first trade.
Every second of load time on mobile loses a share of paid visitors, so a faster page makes the same budget buy more active users. It also helps organic search, where Core Web Vitals are a ranking signal.
Ready to use
ChecklistPaid-traffic readiness check
[ ] LCP under 2.5 s at p75 on mobile [ ] INP under 200 ms [ ] CLS under 0.1 (feed cards sized) [ ] First screen server-rendered [ ] Trading code loaded on demand [ ] Tags deferred after first paint [ ] Paid landing route live and tested [ ] Performance budget in release checks
Risks and how to handle them
Real-time feed feels slower
Keep updates live, just in reserved space.
Speed regressions from fast shipping
Add a performance budget to the daily release checks.
Evidence
- PageSpeed testGoogle PageSpeed for the pump.fun homepage (October 22, 2026). Real Chrome visitors fail all three Core Web Vitals (75th percentile for the page: main content in 4.2 s, taps respond in 467 ms, layout shift 0.32; good is 2.5 s, 200 ms and 0.1); across the whole origin, load time is 20% slower than a year earlier. The mobile lab test scores 29 out of 100: Largest Contentful Paint 28.7 s, Total Blocking Time 1,916 ms, 5.7 MB in 233 requests, 3.4 MB of JavaScript with about 900 KB unused; the biggest layout shift comes from a live feed post card.Source
- HiringHead of Performance Marketing & Acquisition ($400k-$1m, New York, posted August 26, 2026): "build and run our paid growth engine from the ground up"; "a multi-million-dollar, always-on daily budget" across "Meta, Google Ads, TikTok Ads, TikTok Shop, X Ads, programmatic, and emerging crypto-native networks"; "UGC and clipping-style content"; "attribution and measurement infrastructure connecting spend to activation and trading volume, not just clicks"; "Partner closely with Legal and General Counsel ... non-negotiable given the regulatory environment we operate in".Source
- HiringStaff Data Scientist - Growth & Marketing ($400k-$500k, posted September 8, 2026): "frameworks for understanding incrementality, attribution, CAC, LTV, ROAS, payback periods", "optimise for genuine business impact rather than platform-reported performance", "experiments across creative, audience, channel, bidding and landing-page strategies", "predict downstream value", and "Combine advertising platform data with first-party product, behavioural and on-chain data".Source
- Site technologyThe pump.fun homepage loads Google Tag Manager, Google Analytics, Google/DoubleClick ads tags, LaunchDarkly feature flags and AppsFlyer (onelink.me and appsflyersdk.com) mobile attribution. The company blog is on Medium (medium.com/@pumpdotfun_). YouTube @pumpfunofficial: 3,850 subscribers, 215 videos, about 845,000 views; TikTok @pump.fun.Source