I’d noticed something different about the NY Times search over the last few weeks, but just realized today that it’s because they’re split testing a new interface. I have to say I really like it.
Google launched a new hosted site search solution on June 11. Though aimed mostly at SMBs, it wouldn’t surprise me if a lot of large enterprises give Google’s Site Search a serious look.
NYTimes.com is splitting traffic, so to see the beta search you’re going to need to click the links below. In addition to the differences mentioned below, the algorithm is tuned differently in the new search. There is a different result set returned, possibly with less emphasis on AP stories, and I notice a lot more photos in the new results.
- No multimedia included in results. Toggle result to match on “Closest Match”, “Newest First” or “Oldest first”
- Date range radio buttons are difficult to discern and include a “custom date range” function
- Complicated search filters and link to advanced search
- Numbered results set
- Separate blog results included in a right hand box
- Refine results by NY Times section (Arts, Books, Business, etc). Pagination suggests an overwhelming results set.
- As a whole, the interface is much, much cleaner. Results are no longer numbered. Fonts in the results set are slightly larger, snippets are more legible, bylines and dates are diminutive yet easy to find. Multimedia (video, interactives) are now displayed intermixed with other result (not shown in my screencap)
- There is a separate tab for just Multimedia, in addition to one for just articles
- No more advanced search or complex filters. Date filtering is reduced to 5 links, allowing search scope as recent as “Today” and as comprehensive as “Since 1851”, a nod at the Times efforts to make their whole archive available
- The ad column has much more room to breathe, and text ads are used more sparingly. In this case, there wasn’t one
- They’ve killed “topic” search links and greatly simplified pagination
- They’ve moved the “Also Try” links to the bottom and call them related searches. Why put them up top if you’re confident you got it right the first time?
With 3,000,000+ pages in the Google main index, the Times is no small site. However, there are a few compelling reasons for them to stick with Google for hosted site search.
According to Hitwise (and not surprisingly), Google is the number one search engine referrer to NYTimes.com. Since it’s the engine their readers prefer, it does make sense to give site visitors a continuous search experience. Beyond that, I’d guess that using Google site search allows the Times to consolidate a lot of duplicative work. Google Site Search has its own index but also draws results from the main Google index. The Google Site Search crawlers rely on many of the same structural basics (clean and valid code, XML sitemaps) to populate its index. This means that a lot of the efforts by Marshall Simmonds to improve organic search will also help make site search more relevant. Focus on a few key areas, and win in both places. Pretty smart.
Anyone else like the Times new site search? What do you think?