Is Your In-Site Search Good Enough?

There was an excellent post in 2013 by Econsultancy called “Four Reasons why Site Search is Vital for Online Retailers”. The post included a wealth of research into in-site search and became incredibly well quoted by consultants advising ecommerce retailers. It talked about the benefits of site search in terms of increased conversion (up to 50% higher) and improved customer retention.

Shopping with intent and a customer’s place in the purchase journey were big reasons why this was the case and are still accurate today. What has changed is the expectation of a standard search experience.

Functionality and expectations move fast online so although the research is only 3 years old, many of the nice to have functions such as Autosuggest/Complete are now expected functions. As are the use of synonyms, suggestions and corrections. Many of the features that led to significant benefits are now required to compete in a user focused market.

At the same time, some actions are still not being taken to improve outcomes. One item that is still rarely used to its full advantage is the analysis of results, particularly number of “0” result pages and terms that generate them. Search Analytics are probably still a weak area for most retailers looking to optimise conversion through search. They can be a great source of synonyms or non-stocked item requests.

Aside from functionality, search has seen benefits from increasing the user experience. Great search experiences do still lead to increased conversion and tend to go hand in hand with lower abandoned baskets (seems an odd one but clearer/ faster/ better search helps those intent buyers move through the journey faster, so less stalling at basket).

With in-site search still vital there are some expected functions, or available functions, which make all the difference.

Key features of search

Autosuggest / Search / Complete Suggestions

Arguably the only basic/expected function on the list. The most common and well seen applications being the “did you mean” search result alongside autocomplete.

This is utilised by Google so when you start typing “Search” it displays a drop down with “search”, “search engines”, “search by image”, “search history” as the most regular searches starting that way.

On your ecommerce site this helps users find commonly searched products without typing the whole product. It also cuts down on failed searches as you can enter the first 3+ characters and returns the most likely options. It’s particularly useful when you’re not sure of the spelling of a certain item or brand.

Product Suggestions

Provided with title and thumbnail images as the user types they are given a snapshot of images that match their query. A bit like auto suggest but much more powerful since it usually returns the exact product they want rather than taking the user to a category.

This creates a great conversion opportunity as the user doesn’t even need to know what they want in too much detail or to even finish a search.

Relevance

Returning a search results page where the most appropriate suggestions are at the top.

Nothing is more frustrating than getting a returned result set with items which seem to have very little to do with the query at the top.

Spell Correction

Like an autosuggest feature, spell correction is utilised once a search is completed with inaccurate information, e.g. the user has typed “shoos” they get a result set for “shoes”. This considers synonyms and fuzzy search to help people who misspell or do not know the brand name find what they are searching for.

An example from one of our clients where a brand ended in a “z”. People were searching for the brand with an “s” and weren’t getting the correct results.

On average, only 50% of searches are successful so we must help each inaccurate search become successful by limiting exit points due to incomplete results.

Speed

The last expectation is that all of this will be done in a timely manner. Limited delays between keystroke and results help avoid users from exiting the search.

Returning results quickly and accurately are significant trust signals that payment and checkout will be much the same. Slow search is an easy way to lose converting traffic.

Why is basic Magento search so… well… basic?

The default Magento search technology utilises a MySQL search. It is quite poor out of the box in terms of general flexibility and the product matching accuracy. You can modify the results by specifying “like” and “full-text” terms but that’s about it.

The reason this probably hasn’t been taken any further is the richness of even the free extensions available by providers.

One of Magento’s key advantages is its open source community who are frequently innovating and improving on the functionality in the platform. With so many excellent search options it’s probably not worth Magento’s time creating something that’s so well catered for.

What are the options?

Far too many to do justice in recommending really. We use one of 3 nearly every time but have implemented nearer to 10 depending on client’s appetite for cost, function and return on investment.

Some of the most popular that we implement are:

Klevu

Klevu is very popular for Magento and we’ve seen good success from it. It’s most impressive feature is its improvement over time. Klevu learns and improves the more it’s used.

Key features:

  • Autocorrect
  • Product Suggestions
  • Advanced learning
  • Dynamic filters and query reporting

Costs:

  • Free for sites up to 1000 SKUs.
  • 2 paid options Pro (£6 per month) and Premium (£399 per month)

Our rating:

8/10

Algolia

Algolia we are quite new to, but we have found it to be excellent. There’s a couple of ways to implement on the front end. Either a conventional search experience or a lightning fast return to just a results page as you type.

It has a few strengths such as its intuitive user interface which is quite the people pleaser. Its key selling point is its speed. It returns results before you’ve even finished typing. No latency and almost no wait time.

Key features:

  • Excellent user interface
  • Incredible “real time” results
  • Advanced query and product matching

Costs:

  • Monthly subscription of $49, or around £39, per month

Our rating:

9/10

Elastic

This is our go to search solution based on the configuration options. It’s a truly enterprise solution yet is readily useable out the box. You can make it work for you very quickly with only a small learning curve.
Key Features:

  • Fast
  • Simple to use
  • Phenomenal scalability
  • Fuzzy search
  • Great analytics

Costs:

  • Can require a few days to set up correctly
  • Ongoing server costs

Our rating:

9/10

Conclusion

Search is now a requirement for every ecommerce site and basic features such as autosuggest, product suggestions and drill down analytics are expected on your site.

Sales and conversions will be left on the table if you don’t offer the best solution possible and the costs associated with a great search experience are not high enough to put off serious retailers.

If you would like to talk over possible search solutions that fit your budget and will help improve your user’s buying experience, get in touch.


LATEST POSTS

Vortex Shortlisted in Two Categories in the ‘eCommerce Awards London’ 2017!Vortex Shortlisted in Two Categories in the...

1 month ago READ

Vortex Builds Walk the Walk’s New Online Shop!Vortex Builds Walk the Walk’s New Online...

2 months ago READ

WIN a Two-Day Intro to Web Design and WordPress Course!WIN a Two-Day Intro to Web Design and WordPress...

3 months ago READ
all posts