Internal Search - The Lost Child of Web Analytics
Download as PPT, PDF•2 likes•1,295 views
The document discusses the significance of internal search in digital analytics, highlighting it as an often overlooked area despite its potential for revealing user intent and improving user experience. It outlines various tips for optimizing internal search, such as analyzing search queries, capturing failed searches, and understanding user demographics. The author emphasizes the importance of leveraging analytical tools and user observation to enhance the search functionality on websites.
Internal Search - The Lost Child of Web Analytics
- 1. Internal Search ‘The Forgotten Child of Analytics’ Charles Meaden Digital Nation
- 2. A Very Brief History • I’ve been involved in web analytics since 1995 • Run a small digital agency in Mumbles, Wales • Came to look at internal search from a usability and user experience perspective • Worked on internal search for large ecommerce clients, charities and government departments and councils
- 3. Why The ‘Lost Child’? • Some analysts don’t bother to look too deeply into internal search as at first glance it doesn’t look like search is used that often.
- 4. Understanding User Intent • “Search queries are gold: they are real data that show us exactly what users are searching for in their own words” Lou Rosenfield • It’s the closest you’ll getting to what the customer is thinking without actually talking to them • Understand this and you can start improving your – Navigation – Search results – The whole user journey
- 5. Allows For Quick And Easy Fixes • Depending on the platform, it’s a relatively simple job to tweak the platform to fine tune the results • Google Custom Search allows you to – Add Synonyms – Date Bias the results – Categorise the results
- 6. Show All The Routes To Your Content
- 7. It’s Not Just About The Search Box
- 8. Tip 1: Use Your Analytics Internal Search Tool • It makes your life easier • A lot of the tools and reports are already setup
- 9. Tip 2: How Deep Is Your Long Tail? • Which words standout? • What % of the total searches do your top 10 search phrases represent 4,752 unique phrases and 8,414 unique searches Top ten only account for 15% of searches
- 10. Tip 3: Where Are People Searching On Your Site? • Which pages are generating the most searches? • Are there areas of your site that are generating more searches than others
- 11. Tip 4: Collect Your Failed Searches • This doesn’t always come as default • You’ll need to get your developers to find a way to capture this – Via the URL – Via Javascript • Google Tag Manager makes this so much easier – Scrape the DOM – Push an event into the data layer • Once you have it, you can: – Fine tune your search / autocompletes – See how your potential customers describe your services – Find out what your visitors think you are offering – Source new products to sell
- 12. Tip 5: Discover Where People Go Next • Mine your analytics package to find out where people go after the search results page • Use Regular Expressions and / or SQL to manipulate the data • Find out whether people got a content page or did they just go back to the home page Up to 20,000 different combinations per week
- 13. Tip 6: How Many Searched and Then Left? • What proportion of users search and then leave the site?
- 14. Tip 7: What Types Of User Are Searching? • Segment, Segment and then Segment again… • Thing to look out for – New or Returning Visitors – Where did they enter the site – Where did they come from – Are they customers or prospects?
- 15. Tip 8: Cross Match Organic / PPC Phrases • What's the difference between the phrase someone entered into Google and the one they used on your site?
- 16. Tip 9: Does Search Help or Hinder Conversions? • Are people who use search more or less likely to convert? • If more, what phrases work best • If less, what are the ‘stop phrases’
- 17. Tip 10: Look For Patterns In Your Data • Are there certain types of searches that appear? • How often do advice or questions appear – “how to”, – “help on” – “how can” • Are people using abbreviations or shorthand • Catalogue numbers from printed brochures around mailings – If too catalogue numbers appear as failed searches, you may have an issue
- 18. Tip 11: What Are Your Most Common Words? • Your analytics tool will give you the most common search phrases • Are there certain words and phrases that keep on appearing? • Tools such as Textanz and Hermetic Word Frequency Counter will happily crunch 100,000 of phrases an and display the most common words and phrases used within it • Give these to your – Copywriters – SEO and PPC staff and agencies
- 19. Tip 12: How Good Are Your Search Results? • Two metrics will tell you just how good your search is – Number of results pages people searched through – How many times they refined their search
- 20. Tip 13: How Are People Refining Your Faceted Search • If you’re using filtered or faceted search, what’s the most popular refinements people make?
- 21. Tip 14: Learn SQL • Or make friends with someone who knows it • Internal search can throw up 100,000’s of rows of data • Excel has it’s limits • Processing Power is cheap – Use a cloud service such as Amazon AWS and crunch data for less than 50 per hour
- 22. Tip 15: Setup Search as a Goal • A Google Analytics tip • Setting up the search result page as a Goal lets you use the ‘Reverse Goal Path’ • You can then see the three previous pages – The page they searched from – The two previous pages before that • Hat tip to Tim Leighton-Boyce for that tip – Follow him on Twitter @timlb
- 23. Tip 16: Don’t Just Rely on Your Analytics • Use tools such as Hotjar, Decibel Insight, Sessioncam or Inspectlet to observe your users • Run some live usability tests and ask users why they searched in a particular way • Run tests on your current search results and classify the results – The excellent Lou Rosenfield “Search Analytics for Your Site” has plenty of good examples
- 24. Tip 16: Check out Algolia • Very clever SASS search tool – https://www.algolia.com/ • Builds facets on the fly based on a JSON feed • Great analytics • We’re testing with Google Tag Manager to get some really rich data
- 25. Things to Do Next • Read “Search Site Analytics for Your Site” by Lou Rosenfeld – PDF and Print versions • View all the articles I’ve tagged as site_search on Pinboard • Follow me on twitter @charlesmeaden • Join my mailing list – email me [email protected]