Improving the information architecture within a blog – a small project on Ahoy Designs

Issue: GA shows a low % of users who visit an article, visit another article. The top source on Ahoy Designs is organic, with over 90% of the total traffic visiting an article.

Hypothesis: would users appreciate having relevant articles at the bottom of the current article? Would it improve ‘engagement’ metrics, such as pages per session & bounce rate?

How do users interact with articles? Why do we accept a high bounce from blogs? When I personally visit a blog, it’s for something very specific. Within seconds I know if that article is for me or not, therefore if I gets to the end and click on another, wow – that UX deserves a medal!

What are we working with?

The blog on Ahoy likely resonates with a few different personas. It’s an e-com site, selling personalised wedding items. The blog also talks about cruelty-free beauty products. The user is able to filter the blogs on the main hub page. There’s only 4, to help reduce cognitive load (“argh how many categories??”)

Whilst this could get confusing should a user land on the main blog page, I’ve tried to make it as efficient and tailored as possible when on the article pages.

When a user is on an article page tagged with ‘weddings’, they only see other articles at the bottom also tagged with ‘weddings’. This is to encourage the user to stay on the site… we’re presuming the user wants to read more about a particular subject. Hmmm. This is an assumption I haven’t seen proven stats for.

I made this change and now have some results after a month.


  • Bounce rate has reduced from 48.47% to 46.57% (-3.92%)
  • Pages per session (PPS) has increased by 0.56%


  • Bounce rate has reduced from 78.76% vs 65.52%
  • PPS has increased by 5.81%

I’m pretty chuffed with these results. Not only do they show users are bouncing less (blogs are usually the top landing pages in GA so give an accurate reading), they have increased PPS. I don’t look at time on page for pages with high bounce or exit rates – it’s not possible to obtain an accurate time because GA doesn’t know when the user actually left. GA can only tell you when a user gets to a page, from another page (e.g. page 2 from page 1).

Of course these stats could be a blip – a random occurrence. I should have A/B tested it but I don’t have the software to make a change like this from anywhere other than in the code unfortunately.

Improving the architecture further

What about those users who land on the page but want a nosy around for something other than related articles? I added the following to the very bottom of each article, reiterating to the user exactly what type of article they’re reading.

Not only does this give them the chance to visit the ‘Wedding related’ category page (improving my credibility that I have enough articles to create a category), ‘Go back to articles’ will show them all should the user want to see the full range.

Yes this is a tiny change, but showing the users other places to visit in a subtle but helpful way is going to be appreciated. Especially those users who don’t want to continue reading similar articles – it gives them somewhere else to visit.

Leave a Reply

Your e-mail address will not be published. Required fields are marked *