Building a better The Economist search

Yvo Schaap
3 min readSep 25, 2016

Search. Done almost perfect for most queries using Google if you know what you are looking for. But for specific content discovery or research definitely not the best tool.

Case in point: Their website is build around their weekly print editions, which can you browse as a subscriber for each edition. Unfortunately not a very web browsing friendly UX, exemplified by a recent series regarding Economic Briefs which is super interesting (and great to refer to) but spans multiple print editions making it hard to easily access.

To try to find & access that series on their website, the current options offers are:

  • Their search feature (on the homepage) uses Google via their Custom Search Engine product. It takes only a single input field, searches through every page on, gives you a simple list of links, is pretty slow and doesn’t allow for any filtering or sorting (example) making it pretty much useless in almost all cases where you actually want to find something.
  • For special reporting, like the case of the Economics Briefs series they created a custom PDF. Also their store allows you to order and download these special reports. Actually finding these are a hit and miss, and the idea is not a very web friendly approach.
  • They host an archive of all print editions (back to 1997!). You select a year of publishing, followed by an specific edition and then you have to look up the wanted article by going through the complete articles list.
  • They have specific segment-like landing pages on their website. For the mentioned Economist Briefs they created a nice overview page, but there is no logical access to these pages.
    Also a large amount of these pages aren’t updated (Schumpeter and Asia) or don’t allow paging let alone search within these specific sections. The other sections (accessed by the menu bar) like Europe, Asia, United States only chronologically list articles (with limited details) and paging restricted to only three pages (a bug I presume).

It all feels like a confusing mismatch between their print archive and their online CMS.

The good news seems that they are trying to improve by experimenting. Their special The World If is a custom mini-website listing these specific articles. The same goes for the Technology Quarterly. Personally I find it a bit distracting from the actual content, but worse it creates (another) content silo separate from the rest of their site which doesn’t enable users to explore deeper or broader.

In summary a great opportunity for improvement. So I gave it a try.

The search rethought

My goal was to allow for quick access to the relevant articles mentioning specific entities (people, businesses), a geographic area (city, country), keywords, and/or sections, including Economists’ editorials like Schumpeter or their Obituaries (a section more interesting then you might expect).

But also a must needed sorting and filtering on publish date, popularity and/or keyword relevance.

After some night of gathering the data, analyzing the content with OpenCalais, pushing it to Algolia for instant search indexing plus queries, and with use of the Instantsearch.js library to build the search interface.

Et voila “The Economist Explorer” was born. Give it a spin!

Most popular articles about Apple

I thought ‘explorer’ in the name is more appropriate for the features, which make it a great starting point to explore and discover The Economist’ content. Check these example queries: Economics Briefs in one place, Business articles related to India between 2012 and 2014, the most popular Obituaries, Uber mentions, Schumpeter in one place.

Leaders mentioning Google

The entity filtering count even exposes who are most mentioned in specific sections of The Economist (Putin and Trump in Leaders and Nobel and Higgs in Science and Technology).

Nice to have

Todo: Fun to add would a ‘read later’ feature, a link to Blendle for instant access, and of course mobile support of the UI.

Besides those I think there are lots more opportunities with the actual content. But this is already a nice starting point