In Google's data from real Chrome visitors, greatergood.com's homepage fails all three Core Web Vitals (LCP 2.7 s, INP 385 ms, CLS 0.12), and LCP got 11% slower than in the previous period. Google's phone test shows why. The page downloads 6.2 MB in 438 requests, and 4.5 MB of that comes from 44 third-party services. The biggest are the tag manager, the Raptive display-ad script and Google's ad tags, and ads.txt lists 43 ad networks. Ads make sense on the click-to-give pages, because advertisers pay for the free donations. On the pages where people browse and buy socks and hoodies, the same ad stack holds up every tap. The audience is mostly 65+ and spends about four minutes per visit.
First action
Ask Raptive to disable ad slots on the homepage for a 50% traffic split, and record today's revenue per session, ad RPM and field Core Web Vitals.
How to measure it
Total revenue per session, ad-free vs current
- Now
- –
- Target
- +10% with mobile INP under 200 ms
- By
- 60 days
Steps
Baseline
Revenue per session, ad RPM by template, conversion rate and field LCP/INP/CLS for the homepage, collections and products.
Analytics Lead 2 days
Map the templates
List which Shopify templates are cause pages and which are shopping pages, and where ad slots render on each.
Shopify Developer 1 day
Agree the test with the ad partner
Set a 50/50 split with ad slots off on the homepage, with Raptive's reporting matched to the same split.
Ad Operations Manager 3 days
Defer the remaining tags
In the tag manager, fire non-essential tags after the first interaction or on page idle, on the test variant only.
Shopify Developer 3 days
Run the homepage test
Two full weeks, covering at least one weekend and one email send.
Head of Ecommerce 2 weeks
Decide on money
Compare shop margin plus ad revenue per session; keep the winner.
Head of Ecommerce 1 day
Extend to collections and products
Repeat the split on collection and product templates, keeping ads on click-to-give and story pages.
Shopify Developer 2 weeks
Watch field data
Track 28-day field Core Web Vitals in Search Console and PageSpeed until all three pass.
Analytics Lead monthly
Why it works
Split the site into two kinds of page. On cause pages (click-to-give, stories, results), keep the ads that fund donations. On shopping pages (homepage, collections, products, cart), switch off the display-ad stack. Load the remaining tags after the first interaction and lazy-load any ad slot below the fold. Run it as a 50/50 test on the homepage first, then on collection and product templates. Judge it by total revenue per session, counting both shop margin and ad revenue, so the decision is made on money rather than on speed alone.
The shopping path loses sales when the page is slow to show and slow to respond to taps. The ad income from those few shopping pages is easy to measure against the shop revenue gained. GreaterGood keeps the ad model where it funds giving and removes it where it gets in the way of buying.
Ready to use
ChecklistAd-free shopping path test
[ ] Baseline revenue per session, ad RPM and field LCP/INP/CLS by template [ ] Label templates: cause pages (ads stay) vs shopping pages (test ad-free) [ ] 50/50 split agreed with the ad partner, reporting matched [ ] Non-essential tags fire after first interaction on the variant [ ] Two-week homepage test incl. a weekend and an email send [ ] Decide on shop margin + ad revenue per session [ ] Repeat on collection and product templates [ ] Field Core Web Vitals pass on all three
Risks and how to handle them
Ad revenue loss exceeds shop gains
Decide on total revenue per session, and keep below-the-fold lazy-loaded slots if they pay for themselves.
Field data lags the change by 28 days
Use lab tests and real-user monitoring for the first read, and field data to confirm it.
Ad partner terms limit page exclusions
Check the contract before the test and start with the homepage only.
Evidence
- PageSpeedReal Chrome visitors (field data): the homepage fails Core Web Vitals on all three measures, with LCP 2.7 s, INP 385 ms and CLS 0.12 at the 75th percentile, all above Google's "good" limits of 2.5 s, 200 ms and 0.1; LCP is 11% slower than in the previous period.Source
- PageSpeedGoogle PageSpeed mobile test of the homepage (November 3, 2026): performance 45/100; the main content (the hero image) appears at 5.3 s against a 2.5 s target; total blocking time 1.8 s; 6.2 MB downloaded in 438 requests, of which 4.2 MB is JavaScript and about 1.7 MB is unused JavaScript; the element that shifts the layout is the "Menu" control.Source
- PageSpeedThird-party code on the homepage: 44 third-party services sending 4.5 MB. The largest: Google Tag Manager (986 KB, 909 ms of main-thread blocking), the Raptive ad script from adthrive.com (773 KB, 406 ms), Bold Commerce (484 KB, 282 ms), Google/DoubleClick ads (406 KB), Listrak (375 KB), Facebook (228 KB, 207 ms) and Microsoft Clarity (109 ms).Source
- Your websiteads.txt lists 118 seller lines across 43 ad networks (including Google, Amazon, Magnite, PubMatic, Index Exchange and TripleLift), with cafemedia.com (Raptive) as the managing domain.Source
- News coverageHistory: The Hunger Site launched in 1999 and was bought by GreaterGood in 2000; CharityUSA.com, a for-profit Seattle company, took control in 2001 and founded the nonprofit GreaterGood.org (now Greater Good Charities) in 2007. Click-to-give donations are paid for by advertisers. Sister sites include The Animal Rescue Site, The Breast Cancer Site, The Rainforest Site, The Literacy Site and The Veterans Site; Click for Paws launched in 2023.Source
- Traffic estimateSimilarweb estimates for September 2026: about 2.5 million visits, down 10.6% on the month; direct is the largest source (41.6%), then display and email; 81% of visits from the US, 5.0% UK, 3.5% Germany; audience 74% female, with 65+ the largest age group; 3.95 pages and 4 minutes per visit; Similarweb lists Gen AI as the most underused channel.Source