Internal Search Thoughts Part 2

In my first post about internal search, I talked about making the search engine more artificially intelligent by incorporating analytics and user behaviour as indicators to provide the best search results. This next part will be about programmatic indicators to help manage the search results. These indicators aren’t based off of analytics, but are based off of content and semantic content.¬†

The programmatic angle in this post encompasses 4 ideas:

  • Date and time
  • Content and meta keywords
  • Meta description
  • Semantic content

Each of these indicators are a backup for the user behaviour to bring relevant results up. I’ll explain each one of them.

  • Date and Time – Depending on what type of site you have, either one or both of these indicators will be important to bring up relevant and up-to-date content. Date and time is relevant in that you want the most current content available, especially if your site provides news or content that can be considered out-dated in medical terms. While working with a large consumer medical site, I would find out-dated information at the top of the search results while more current information was buried. For e-commerce sites, having the most current content for Do-It-Yourself projects will help you sell the most modern products. However, if someone is looking for pre-dated items, having this as a part of your faceted search will allow your users to drill through the content to more easily find the items that they are looking for.
  • Content and Meta Keywords – While these are important aspects to provide the most relevant content, they have been abused. From my experience, too many sites use these indicators as a be-all-end-all for the most relevant content. Too many content creators use the keyword and description meta tags as indicators of relevant content. The keywords meta tag needs to be used sparingly. There is no reason to be over generalized if you are using this as an indicator in your search engine.
  • Meta Description – This should be used as the description that is shown in the search results and should have a small indicator for the search engine. Just like Google will take your meta description and serve it up in their search results, this should also be a part of your search engine so as to offer strong, relevant results without truncated content.
  • Semantic Content/Microdata – since search engines are algorithmic animals that don’t have the capacity to think like us humans do, adding microdata to your content can help really push forward the relevant content that your visitors and users are looking for. Microdata can also help facilitate better results when using faceted search where appropriate. By using microdata, you can mark up your content where it’s needed so that the SERPs are relevant and full of rich content like reviews, stars, and images. Adding this type of markup makes the results more user friendly and can make the search engine a more trusted source.

Each of these are important indicators, but need to be used properly. I have worked on search engines that were word match only, and this spurred my interest into making search engines better. Follow the leader, right? Google is the leader in semantic search, well created algorithms, and user behaviour. By implementing what we see them doing into an internal search engine and continuously making it better, you are creating a user-friendly environment that will create trust, better buy cycles, and more relevant information.

If you work on an internal search engine on an e-commerce site or consumer information site, what have you done to make your search engine better?



No comments so far.

Leave a Reply
  (will not be published)