SlideShare a Scribd company logo

Engagement is the New Anchor Text - Stephen Kenwright - SMX London May 2017

•
•
Branded3
Branded3

Branded3's Director of Search Stephen Kenwright, looks at whether 'Engagement is the New Anchor Text' and how you can future-proof your SEO. Delivered at London's SMX, May 2017.

1 of 36
Download to read offline
#SMX #15A @stekenwright Future-proofing your SEO ENGAGEMENT IS THE NEW ANCHOR TEXT
#SMX #15A @stekenwright CTR + dwell time = rankings money
#SMX #15A @stekenwright Redesigning a blog post EXPERIMENT #1
We chose a really long blog post @stekenwright#SMX #15A
#SMX #15A @stekenwright 2,955 words No images apart from Searchmetrics charts Ranks for some things so must be doing its job (source: SEMrush) https://www.branded3.com/blog/website-migration- guide-building-an-seo-checklist-for-moving-your-site/
…and reimagined what it could look like @stekenwright#SMX #15A
#SMX #15A @stekenwright No distractions: removed opportunities to sign up to our email list, follow us/me on social or click on related posts We reduced the clutter New Old
We added a scrolling parallax effect @stekenwright#SMX #15A
#SMX #15A @stekenwright The post consists of several checklists so we made them easier to follow We designed new lists New Old
#SMX #15A @stekenwright We added some birds for some reason
#SMX #15A @stekenwright We added a splash of colour @stekenwright#SMX #15A
#SMX #15A @stekenwright StumbleUpon Twitter We used paid social advertising to drive traffic to both
#SMX #15A @stekenwright We wanted to track dwell time We don’t trust Google Analytics …so we set up two tests in Google Tag Manager: we tracked scroll depth and set a timer We tracked performance in Tag Manager
#SMX #15A @stekenwright A user spending a long time on a page doesn’t necessarily imply they like it We fired events at 0%, 25%, 50%, 75% and 100% of the way down the page We can see which bits of pages aren’t interesting or cause problems Why track scroll depth?
#SMX #15A @stekenwright http://scrolldepth.parsnip.io/ @robflaherty → 👍 To track scroll depth you want this:
#SMX #15A @stekenwright advancedwebranking.com/blog/how-to- track-scroll-depth-with-google-tag- manager @matthew_jkay → 🔥 Struggling? Read this:
#SMX #15A @stekenwright How Google Analytics calculates time on site 1st page 11:00:00 2nd page 11:00:43 3rd page 11:01:55 Another website 11:02:12 Google Analytics subtracts times to calculate the following: Time on 1st page 43 sec Time on 2nd page 1:12 min Time on 3rd page Unknown
#SMX #15A @stekenwright If a visitor lands on a page and doesn’t click anything else on the site that counts as a bounce in Google Analytics… …even if they spent 10 minutes reading the content (Plus Google doesn’t use bounce rate to rank pages) …and there’s a problem with bounce rate
#SMX #15A @stekenwright We fired an event every 5 seconds (interval 5000) to check the visitor was still there How: support.google.com/tagmanag er/answer/6106961?hl=en Our timer "event": "gtm.timer", "gtm.timerId": 21, "gtm.timerEventNumber": 1, "gtm.timerInterval": 5000, "gtm.timerLimit": 360, "gtm.timerStartTime": 1493651345330, "gtm.timerCurrentTime": 1493651350331, "gtm.timerElapsedTime": 5001, "gtm.triggers": "281204_81", "gtm.uniqueEventId": 7
#SMX #15A @stekenwright New Old 03:55 Average time on site 95.00% Bounce rate What does Google Analytics say about the two versions? 05:36 Average time on site 91.07% Bounce rate 13k visits later “blog post” style wins…
#SMX #15A @stekenwright New Old 2:18 Average time on site 9.8% Made it 75% of the way down What does Tag Manager say? 14:42 Average time on site 28% Made it 75% of the way down Actually “new” style was clearly a disaster
So. Much. White. Space. @stekenwright#SMX #15A % of visitors probably didn’t make it past 25% because most of the first 25% has no content. Hindsight.
#SMX #15A @stekenwright Rankings disappear. Completely. …then we swapped back to the “blog” version and rankings came back. What happens to rankings when the posts are swapped?
#SMX #15A @stekenwright The designed page loads 28% slower on mobile and 26% slower on desktop – could this be the cause of the drop? The blog version has 5 comments – we disabled comments on the designed version Visitors could just leave the tab open and skew timer results Maybe not so fair…
#SMX #15A @stekenwright Spending 10x as long on content doesn’t make it “10x content” Employing a designer isn’t a tick box for engagement – always consider what the page should really look like Bounce rate is a bad metric – people can find what they want really quickly and it looks like a bounce Some reminders
#SMX #15A @stekenwright New forms of rich media EXPERIMENT #2
#SMX #15A @stekenwright I loved Digiday’s TLDR feature @stekenwright#SMX #15A
#SMX #15A @stekenwright We’ve been using tl;dr on posts for a long time @stekenwright#SMX #15A
#SMX #15A @stekenwright tl;dr reduces dwell time and engagement How do we give our audience a useful feature without reducing the time they spend with us? Problem
#SMX #15A @stekenwright Our solution: anchor.fm @stekenwright#SMX #15A
#SMX #15A @stekenwright …or, y’know, podcasts that only last 2 minutes What is Anchor?
#SMX #15A @stekenwright How does it work? @stekenwright#SMX #15A
#SMX #15A @stekenwright Embed a 2 minute audio version of a blog post for people who only want a summary Drive traffic from the Anchor app to our blog posts…only a small user base so it isn’t much Increase time on page What are we using it for?
#SMX #15A @stekenwright We added 10 “tl;dr” summaries to blog posts that were ranking and tracked over a month: 8 improved (all by more than 5 places) 1 dropped (by more than 5 places) 1 stayed the same So it seems to be helping. Does it make a difference to rankings?
#SMX #15A @stekenwright Podcasts and video have proven similar but it’s so easy to sound like a pro on a podcast (no budget) As in the previous test – it’s a blog post so don’t take up too much space, let people read if they want to Some reminders
#SMX #15A @stekenwright slideshare.net/branded3 THANK YOU

More Related Content

Engagement is the New Anchor Text - Stephen Kenwright - SMX London May 2017

  • 1. #SMX #15A @stekenwright Future-proofing your SEO ENGAGEMENT IS THE NEW ANCHOR TEXT
  • 2. #SMX #15A @stekenwright CTR + dwell time = rankings money
  • 3. #SMX #15A @stekenwright Redesigning a blog post EXPERIMENT #1
  • 4. We chose a really long blog post @stekenwright#SMX #15A
  • 5. #SMX #15A @stekenwright 2,955 words No images apart from Searchmetrics charts Ranks for some things so must be doing its job (source: SEMrush) https://www.branded3.com/blog/website-migration- guide-building-an-seo-checklist-for-moving-your-site/
  • 6. …and reimagined what it could look like @stekenwright#SMX #15A
  • 7. #SMX #15A @stekenwright No distractions: removed opportunities to sign up to our email list, follow us/me on social or click on related posts We reduced the clutter New Old
  • 8. We added a scrolling parallax effect @stekenwright#SMX #15A
  • 9. #SMX #15A @stekenwright The post consists of several checklists so we made them easier to follow We designed new lists New Old
  • 10. #SMX #15A @stekenwright We added some birds for some reason
  • 11. #SMX #15A @stekenwright We added a splash of colour @stekenwright#SMX #15A
  • 12. #SMX #15A @stekenwright StumbleUpon Twitter We used paid social advertising to drive traffic to both
  • 13. #SMX #15A @stekenwright We wanted to track dwell time We don’t trust Google Analytics …so we set up two tests in Google Tag Manager: we tracked scroll depth and set a timer We tracked performance in Tag Manager
  • 14. #SMX #15A @stekenwright A user spending a long time on a page doesn’t necessarily imply they like it We fired events at 0%, 25%, 50%, 75% and 100% of the way down the page We can see which bits of pages aren’t interesting or cause problems Why track scroll depth?
  • 15. #SMX #15A @stekenwright http://scrolldepth.parsnip.io/ @robflaherty → 👍 To track scroll depth you want this:
  • 17. #SMX #15A @stekenwright How Google Analytics calculates time on site 1st page 11:00:00 2nd page 11:00:43 3rd page 11:01:55 Another website 11:02:12 Google Analytics subtracts times to calculate the following: Time on 1st page 43 sec Time on 2nd page 1:12 min Time on 3rd page Unknown
  • 18. #SMX #15A @stekenwright If a visitor lands on a page and doesn’t click anything else on the site that counts as a bounce in Google Analytics… …even if they spent 10 minutes reading the content (Plus Google doesn’t use bounce rate to rank pages) …and there’s a problem with bounce rate
  • 19. #SMX #15A @stekenwright We fired an event every 5 seconds (interval 5000) to check the visitor was still there How: support.google.com/tagmanag er/answer/6106961?hl=en Our timer "event": "gtm.timer", "gtm.timerId": 21, "gtm.timerEventNumber": 1, "gtm.timerInterval": 5000, "gtm.timerLimit": 360, "gtm.timerStartTime": 1493651345330, "gtm.timerCurrentTime": 1493651350331, "gtm.timerElapsedTime": 5001, "gtm.triggers": "281204_81", "gtm.uniqueEventId": 7
  • 20. #SMX #15A @stekenwright New Old 03:55 Average time on site 95.00% Bounce rate What does Google Analytics say about the two versions? 05:36 Average time on site 91.07% Bounce rate 13k visits later “blog post” style wins…
  • 21. #SMX #15A @stekenwright New Old 2:18 Average time on site 9.8% Made it 75% of the way down What does Tag Manager say? 14:42 Average time on site 28% Made it 75% of the way down Actually “new” style was clearly a disaster
  • 22. So. Much. White. Space. @stekenwright#SMX #15A % of visitors probably didn’t make it past 25% because most of the first 25% has no content. Hindsight.
  • 23. #SMX #15A @stekenwright Rankings disappear. Completely. …then we swapped back to the “blog” version and rankings came back. What happens to rankings when the posts are swapped?
  • 24. #SMX #15A @stekenwright The designed page loads 28% slower on mobile and 26% slower on desktop – could this be the cause of the drop? The blog version has 5 comments – we disabled comments on the designed version Visitors could just leave the tab open and skew timer results Maybe not so fair…
  • 25. #SMX #15A @stekenwright Spending 10x as long on content doesn’t make it “10x content” Employing a designer isn’t a tick box for engagement – always consider what the page should really look like Bounce rate is a bad metric – people can find what they want really quickly and it looks like a bounce Some reminders
  • 26. #SMX #15A @stekenwright New forms of rich media EXPERIMENT #2
  • 27. #SMX #15A @stekenwright I loved Digiday’s TLDR feature @stekenwright#SMX #15A
  • 28. #SMX #15A @stekenwright We’ve been using tl;dr on posts for a long time @stekenwright#SMX #15A
  • 29. #SMX #15A @stekenwright tl;dr reduces dwell time and engagement How do we give our audience a useful feature without reducing the time they spend with us? Problem
  • 30. #SMX #15A @stekenwright Our solution: anchor.fm @stekenwright#SMX #15A
  • 31. #SMX #15A @stekenwright …or, y’know, podcasts that only last 2 minutes What is Anchor?
  • 32. #SMX #15A @stekenwright How does it work? @stekenwright#SMX #15A
  • 33. #SMX #15A @stekenwright Embed a 2 minute audio version of a blog post for people who only want a summary Drive traffic from the Anchor app to our blog posts…only a small user base so it isn’t much Increase time on page What are we using it for?
  • 34. #SMX #15A @stekenwright We added 10 “tl;dr” summaries to blog posts that were ranking and tracked over a month: 8 improved (all by more than 5 places) 1 dropped (by more than 5 places) 1 stayed the same So it seems to be helping. Does it make a difference to rankings?
  • 35. #SMX #15A @stekenwright Podcasts and video have proven similar but it’s so easy to sound like a pro on a podcast (no budget) As in the previous test – it’s a blog post so don’t take up too much space, let people read if they want to Some reminders