Search is a critical feature for any e-commerce store. It helps visitors cut through the clutter so that they can find what they are looking for and take the decision to purchase quickly. These search users have contributed close to 30% of revenues to businesses and ignoring these users would be a big mistake. Yet, that is exactly what so many Magento based e-commerce sites do.
The native search that Magento uses lacks the finesse that an e-commerce search needs. More and more e-commerce businesses are taking the call to switch to third-party hosted search providers for a permanent fix to the problem. To understand where the faults lie with native Magento 1.x search we need to understand how the search works.
How the native Magento 1.x search works
Magento 1.x search is based on MySQL for community users and if you’re using the Enterprise edition then you can base your search on Apache Solr. As far as Solr is concerned, you will need a dedicated development team, ideally with experience in Solr implementation. Then to be able to have it function at a high level will be a constant task.
This sort of time and resources is not something that a lot of e-commerce businesses can afford so they tend to stick with the default MySQL search. MySQL can return results to queries made in three ways. “Like”, “FullText” and “Combine”.
“Like” - This returns results based on a “loose” approach. Whenever a query is made, it is broken down to its individual keyword components and results for each keyword are shown. This can work for specific setups where users search only for simple keywords, but it is not accurate when it comes to queries with more than one word. Also “Like” doesn’t sort the results so don’t be surprised if the search user has to scroll down to find the product. Eg. searching for “shir” would give results for “t-shirts” and “formal shirts” both.
“Fulltext” - This matches each keyword of the query against the index and the result consists of matches found. It is able to sort the results based on best matches. Biggest roadblock “Fulltext” runs into is not being flexible enough to accommodate mistakes in user queries. The best bet to deal with these mistakes is to configure Search Terms as “Fulltext” won’t be able to find results based on partial matches, typos, or misspellings. Eg. searching for “dtess” instead of “dress” will not give any results. In other words, “Fulltext” offers no typo tolerance.
“Combine” - This matches results based on the “Like” condition and sorts them according to best match like “Fulltext”. So a user will be able to see more products when they use the search and have some sorting based on relevancy. While it may sound like a good solution, this tends to slow the search process down.
Now that we’ve established how the search functions for a native Magento search, let’s take a look at where the gaps are for the e-commerce search experience.
Trials of having native Magento 1.x search
1. Low search relevance
From the outset, you can see that whichever search configuration you choose (most likely “Fulltext” or “Combine’) there are going to be drawbacks. The most important one for an e-commerce platform is relevance. Out of the box, the native search has low relevance, but as Magento is an open source platform, you can make some changes to the logic of the search. Some e-commerce platforms say that their search was relevant only 50% of the time and 80% of search users exited the site after getting irrelevant results. Safe to say then that optimising search should be a high priority for an e-commerce store.
2. Long indexation time
When a customer uses the search bar, they expect to find what the store has to offer so what happens when a search falters at this step. There are cases where the native Magento search is returning zero results even though the product is in your inventory. The fix is simply to reindex your catalogue. The problem then arises about how long the time the indexation takes. Some stores with a catalogue of around 5000 SKUs say that the process can take days if there are server timeouts. You could also reindex individually but that would be a huge time sink and would add no real value to your customer’s user experience.
3. Slow search speed
Speed is also a concern with search speed in case of e-commerce stores with a large inventory can take very long, up to 200-300 ms in some cases. A slow search can be very frustrating to a user and can be a big reason for exits. One stopgap exists which is very counterintuitive to the search process. Some people recommend reducing searchable attributes to speed up the search. Because of the lower turnaround time, a user may get results faster but there is no guarantee that they will be accurate. You will also hinder the navigational aspect of your search by limiting which tags can and can’t be searched.
4. Poor results for long tail queries
Long tail keywords can also lead to unreliable search results for the native Magento search. When the search is breaking down the query, if attribute weights are not set, then it could lead to the results being wildly inaccurate. Eg. searching for “levis blue jeans” will give results based on brand (“levis”), colour (“blue”) and apparel (“jeans”) but which keyword gets priority over the other and how will the results get sorted? The search results page in this example would be a complete jumble with poor attribute weights. The lowest these weights can be is 1 and the highest is 5 so you can only alter these weights to some degree.
5. Poor customer support
There are some techniques and guides that you may find online, but they might not be properly documented. This could lead to much trial and error before you can get the e-commerce search to a satisfactory level. Relying on customer support is also not a viable strategy in the long term as support is rare, especially for 1.x version.
6. Updates change settings
The biggest pain point of all, however, is when you update your version of Magento, the update may wipe out the changes you’ve made and set you back to square one. Once again you will have to figure out the weaknesses of the search are and how to fix them. Is this distraction of time and resources from central business helpful? If you would like to improve your search but not take on the cost of hiring a team of developers, third-party search providers are the answer.
How to improve your e-commerce search
It’s not hard to see why so many Magento users are moving toward hosted search solutions like SearchTap.io. Not only will these search solutions offer synonyms, weighted search, etc, and they also improve the search user experience immensely by making the search faster and more accurate.
SearchTap.io is backed by an expert team that builds, implements, and maintains e-commerce search for big brands and startups. It has an unparalleled product that is already serving 15 million searches per month.
The improvements that SearchTap.io would offer to improve your e-commerce search experience are things like:
- Preserving your site design while offering a better search experience
- Dynamic filters or faceting to allow users to refine their search as they are searching
- Typo tolerance to power through minor errors made by users
- Search as you type so that your store can start showing results as soon as a user starts typing in the search bar
- Customer success team that handles implementation and troubleshooting
- Custom zero result pages to try and keep users engaged for longer
- Searchandising to promote certain products when particular keywords are searched
- Custom banners that appear when users interact with the search bar
- Customer support for implementation and troubleshooting when required
- Implementing updates without disrupting service to your customers
Outsourcing the search function allows you to focus on your core business activities because the search is being handled by experts. Reduce friction between the user and the storefront so as to shorten the user journey. The easier, the user can get to their desired product, the easier they are going to be to convert.
Read about how our Magento search extension helps optimize conversions for your e-commerce store.
Hi there! Want to ask about something? Leave a message.