Grayscale audit

// briefing

Briefing

What is wrong, who must decide, and what ships first.

3

Critical

13

High

13

Medium

8

Low

37 findings, rated Critical, High, Medium or Low, across 5 Grayscale properties.

Bottom line

Fund pages rank for their own tickers. Deck v1.1 names links that need JavaScript as the main technical problem, and a new crawl does not confirm it. The main problem is pagination (the page numbers under a list): the list pages (hubs) for The Stack (a Grayscale article series), the blog, research reports and market commentary show older articles only after a button click. Those buttons have no link. 91 of 108 Stack articles sit only on pages 2 to 27 of the hub. 25 Stack articles have no link from any Grayscale page and are in no sitemap.

Two findings need a legal or security owner before any SEO work. The ETF site keeps loading analytics and ad tags (tracking scripts) after a visitor declines cookies. The Institute site shows text that matches security test markers, template placeholders, and a privacy policy line that says personal data is sold to sponsors.

Fix in this order: consent tags on etfs, the Institute placeholder and privacy text, crawlable pagination links, redirects for 975 dead URLs, canonicals (tags that name each page's preferred address), and sitemaps (lists of a site's pages for search engines). Most of these are template, tag and server changes. Changes to text that visitors or search results show (the Institute text, fund page asset names, titles, descriptions and H1s) need legal or compliance review, as the action plan states for each item.

PageCriticalHighMediumLowTotal
Trust and compliance270312
Search and AI discovery169319
Content and editorial00426
Total31313837

The three Critical findings

What is at stake

  • Trust and compliance: consent tags (A-1) and the Institute privacy and test text (C-1, A-3) affect a regulated issuer's public record.
  • Discovery: 25 Stack articles have no link from any Grayscale page and no sitemap entry, so search and AI tools are less likely to find them (A-2).
  • Lost links: 975 old URLs end in a 404 (page not found), and the backlink data shows 2,693 broken backlinks (links from other sites) to the domain (A-4).
  • AI answers: plain HTTP clients (programs that fetch pages without a browser) get no page content from five hosts (A-5). No page on www, etfs or research carries structured data (code that labels page content for search engines) (A-14). The Institute structured data names another company as publisher (A-26).

What to do first

  1. Make the ETF site tags obey a declined cookie choice (A-1).
  2. Remove the Institute test markers and placeholders, and replace the privacy line, after legal review (C-1, A-3, C-2).
  3. Give hub pagination real links (A-2).
  4. Redirect the dead URLs, starting with the 45 usable rows of the 60-row redirect map (a list of old and new addresses) (A-4).
  5. Add self-canonicals and canonicals for parameter URLs (addresses with extra tags after a question mark) (A-6).
  6. Rebuild the sitemaps from the live page list (A-8).

Who acts and who reviews

Order Finding Who Review needed
1 A-1 consent tags on etfs Engineering, legal Legal
2 C-1, A-3, C-2 Institute text Engineering, legal, security Legal, security
3 A-2 pagination links Engineering None
4 A-4 redirects Engineering None
5 A-6 canonicals Engineering None
6 A-8 sitemaps Engineering None

Decisions and access needed

  1. Legal: approve the consent fix scope for etfs and www (A-1, A-11).
  2. Legal: review the Institute privacy policy text (C-1).
  3. Security: review the Institute test markers before they are removed (A-3).
  4. Engineering: confirm who changes the firewall rules at Vercel (hosting) and Sucuri (firewall service), and share 30 days of firewall logs for AI crawlers (bots that read pages for AI tools) (A-5).
  5. Compliance: approve fixes for wrong asset names on fund pages and fund documents (C-3), and the status notes on 27 pages that still call an ETF pending (C-4).
  6. Compliance: decide whether the public webinar page marked confidential stays public (C-5), and whether the Ethereum report copy without its disclosure section is removed at once (Part G.1).
  7. Access: Google Search Console and Google Analytics 4 (GA4) were not available. First-party (the client's own) traffic numbers in this report are absent for that reason.

Where to start, by role

The five properties

What could not be checked

  • First-party search data: no Search Console or analytics access.
  • AI crawler access from the crawlers' own networks: firewall logs are needed (A-5).
  • events.grayscale.com: a bot challenge (a check for automated visitors) blocked every test request, including a real browser.
  • Login areas (Institute classes, slate.grayscale.com): out of scope.

Scope and evidence

  • 1,125 URLs fetched in a real browser on 5 hosts, with the server HTML and the rendered page saved for each of the 1,083 HTML pages.
  • 499 indexable pages (status 200, no redirect, no noindex [a tag that hides a page from search], no query string [no ? in the address]) read in full and given a verdict (Keep, Improve, Merge or Remove).
  • 1,952 URLs tested for redirect codes, including 635 historic page paths and 738 historic PDFs from the Internet Archive (a public web archive).
  • 1,503 rows in the seven deck issue lists checked one by one.
  • Speed: Lighthouse (Google's page quality and speed tool), 3 runs per template on mobile and desktop. Accessibility: axe-core (an automated accessibility checker) on 15 templates at 390 px and 1440 px.
  • Search data: a third-party keyword, ranking and backlink data provider (US market). Traffic and volume figures are modelled. None comes from Search Console.