Skip to content
Get Ranked On TopIllustrative case studySignal plan

A fictional business. The figures are a modelled scenario, not a client result.

HUSHMARK PODCASTSINDEPENDENT PODCAST NETWORK, 14 SHOWS · NEW YORK

How a 14-show podcast network spliced out 30,000 thin pages and got its episodes indexed

From 32% to 94% of episode pages indexed

9 months · North America · Signal plan · Index pruning · Canonical tags · XML sitemaps · Log-file analysis

Episode pages indexed · month 9

up from 32%

94%

THE PROBLEM

Fifteen offcuts for every take.

Auto-generated transcript, tag and player-embed pages created 30,000 thin URLs, so only a third of Hushmark's 2,000 episode pages were indexed.

Hushmark's CMS made a page for everything: each transcript chunk, each tag, each player skin, each timestamp link. Google crawled the clutter, judged most of it thin, and left two in three episode pages out of the index, new releases included, in the very week they were promoted.

The reel, as Google crawled it

32,014

32,014 crawlable URLs

  • Episode and show pages2,014
  • Transcript fragments9,600
  • Tag pages11,800
  • Player embeds6,000
  • Timestamp and share links2,600

Only 2,014 of them worth a place in search

THE AUDIT

What we heard on playback.

URL inventory

15 : 1

Fifteen thin URLs for every episode page: 32,014 crawlable addresses, and just 2,014 that deserved to rank.

Finding 1 of 4

Transcripts

9,600

fragment pages, five per transcript, near-identical to each other. None had the player or a link back, so they competed with the episode.

Finding 2 of 4

Tag pages

71%

of 11,800 tag pages held a single episode. Generated from every guest name and keyword, they read to Google as empty lists.

Finding 3 of 4

Crawl budget

6%

of Googlebot's hits reached an episode page. A new episode waited a median of 23 days for its first crawl, long after launch week.

Finding 4 of 4

THE APPROACH

Cut the noise, keep the take.

Built on our Signal plan: technical SEO for sites whose CMS writes more pages than their people do.

  1. 01

    Index pruning

    Method: 301 · 410 · noindex

    Each of the 32,014 URLs was sorted into keep, merge, noindex or remove. Transcript fragments went back into their episodes and 11,680 tag pages were retired: one URL per idea.

  2. 02

    Canonical tags

    Method: rel="canonical"

    Timestamp, share and sort links now point a canonical at the clean episode URL, and internal links never carry parameters, so 2,600 duplicates fold into the page they copy.

  3. 03

    XML sitemaps

    Method: sitemap.xml · lastmod

    The sitemap went from 32,014 URLs to 2,134: episodes, shows and the 120 topic hubs, each with a true lastmod from the publishing feed, so new episodes are found within days.

  4. 04

    Log-file analysis

    Method: access.log · Googlebot

    Ninety days of server logs showed where Googlebot spent its visits. We measured again every month to prove the crawl had moved to episode pages: from 6% of hits to 58%.

THE SPLICING BLOCK

Cut the offcuts. Join the keepers.

Here the site is one reel of tape: each segment is a kind of URL, its length the number of pages. Choose a thin segment, cut it out with its treatment, and watch the meter for episode pages in the index.

0255075100EPISODES INDEXED
Indexable URLs
32,014
Episode pages indexed
640
of 2,000
Googlebot crawl on episodes
6%

Choose a segment, then cut it. The arrow keys move along the tape.

Transcript fragments

Marked to cut

9,600 URLs

What it was
Every transcript was chopped into 5 pages of about 400 words, with no player and no link back to the episode.
What we did
Transcripts folded back into the episode page; every fragment now 301s to its episode.
Sample URL
/transcripts/ep-212/part-3
Treatment
301 → /shows/…/ep-…#transcript

Modelled figures for a fictional network.

THE EDIT

9 months on the splicing block

  1. Month 1

    Log the reel

    Crawled all 32,014 URLs and matched them against 90 days of server logs and Search Console data.

  2. Month 2

    Mark the cuts

    Every page template sorted into keep, merge, canonical, noindex or remove, and signed off with the producers.

  3. Months 3–4

    First splices

    Transcripts folded into their episode pages behind 9,600 redirects; timestamp and share links canonicalised.

  4. Month 5

    Tags and embeds

    11,800 tags consolidated into 120 topic hubs, player pages set to noindex, the sitemap rebuilt to 2,134 URLs.

  5. Months 6–9

    Playback

    The recrawl watched in the logs every week, while indexing climbed from 66% to 94% of episode pages.

THE RESULTS

A shorter reel, and every take on air.

Episode pages indexed

32% → 94%

640 → 1,880 of 2,000 episode pages

Thin URLs pruned or canonicalised

30,000

120 tag pages kept as topic hubs; everything else redirected, canonicalised or set to noindex.

Organic visits to episode pages

+87%

41,000 → 76,700 organic visits a month

Googlebot crawl on episode pages

6% → 58%

New episodes crawled within 2 days, not 23

Modelled scenario for a fictional business — not a client result. Rankings and traffic are never guaranteed; see our Terms.

Episode pages in the index, month by monthEpisode pages indexed (%)Thin URLs still indexed
0%25%50%75%100%04,0008,00012,000M0M1M2M3M4M5M6M7M8M994%32%

Where Googlebot spent its visits

Before
Month 9
  • Episodes
  • Transcripts
  • Tags
  • Embeds
  • Share links

Sample episode queries (US, Google)

QueryBeforeMonth 9
podcast about the 1977 new york blackout—#3
night ferry podcast episode 212#14#1
subway tunnel history podcast—#6
how vinyl pressing plants work podcast#38#4
interview with a jazz luthier podcast—#2
bodega economics podcast transcript#61#5

Positions from a modelled rank tracker; “—” means outside the top 100.

Is your archive burying your best work?

Our Signal plan audits every template on your site, cuts what Google should never have seen, and keeps the pages that earn listeners.

All case studies