So, your site needs an e-commerce site search solution.
Right now, you’re probably imagining the daunting task of researching, comparing, testing, and ultimately choosing a site search solution. Before you drown in side-by-side comparisons or give in to that dull headache from yet another generic demo video, ask yourself this:
What top needs or challenges should the new e-commerce site search solution address?
- Does it need to increase customer engagement?
- Do you need a more simplified yet intuitive search interface?
- Do you want to drive more sales and conversions from search activities on your site?
- Are you looking for a better ROI on your search tool?
With your top priorities in mind, you can focus in on relevant features that will drive results. Your business and customer needs can form the basis for your analysis and also help you define questions for vendors. After all, why have a site search tool if it can’t handle everything your e-commerce site currently does—as well as what it needs to do in the future?
Once you confirm the site search tool you’re eyeing covers the basics, focusing on your needs will ensure you find the best e-commerce site search partner.
To be effective, all e-commerce site search tools need a few basic things:
The basic function of an e-commerce tool is to find and rank results that best match a customer’s query. While this seems simple enough, all e-commerce site search solutions search algorithms are not created equal.
Some algorithms rely on simple keyword matching. Others rely on how frequently the term shows up on a page, the number of likes or reviews, or some combination of these factors. However, these methods alone won’t surface satisfactory results—for a robust search function, you need much more.
The algorithm should take into account the context of each search and present results based on textual and business relevance. To offer top-notch relevance, the tool should use ranking criteria, including:
- Tolerate user typos
- Distinguish between the importance of different product attributes
- Allow the business to set custom ranking criteria
- Rank results by distance, if relevant to your business
- Dynamically filter results
The search engine should present these results in order of relevance to the user.
In order to deliver the most value to users, e-commerce site search should also surface relevant content alongside products. For example, product guides or blog posts about relevant current trends can appear in a search along with products. This promotes search and discovery, increases user engagement with the site, and adds value for site searchers.
Out-of-the-box solutions entice businesses with the promise of quick implementation. Yet, money saved on setup can compound on maintenance and upkeep if the solutions fails to deliver reliability, coverage, and speed.
- Speed. It’s easy to create a fast site search, and it’s relatively easy to make a relevant one. But mastering both is difficult. Seconds of delay waiting for relevant results can add up and hurt the user experience. The site search vendor should maintain extensive infrastructure to support search-as-you-type and instant search results for all of your customers without latency.
- Global Coverage. Every website, but large sites in particular, needs the guarantee that all customers will have a stellar search experience—regardless of their location. The company should have infrastructure, such as a global distributed search network, to ensure each customer enjoys top-notch site search performance.
- Reliability. Downtime can ruin your customer’s perception of your brand and their experience. If the vendor can’t guarantee data centers around the world to ensure maximum uptime, you might want to look elsewhere.
Now that you have some basic requirements in mind, you need to consider some additional and advanced capabilities that can elevate the search experience for you and your customers. Let’s consider some main areas:
- Merchandising tools
- AI and machine learning
Pay close attention to how indexing works with each site search tool you consider.
For some solutions, the frequency of indexing and volume of items indexed depend on your pay tier. If your site frequently adds products on a daily or weekly basis, a solution that only indexes monthly won’t fly. New products need to appear in user’s searches as soon as they are on the website—not a month later.
If your site includes highly specialized content, look for tools that support the ability to index non-product content such as blog posts, guides, or customer reviews.
Many site search tools are ready to go out-of-the-box, which can be ideal for a quick implementation. The trade-off, however, is that these tools are difficult or impossible to customize. Open source solutions boast the most customization options, but are resource-intensive to implement, maintain, and update.
Search as a service solutions generally sit right in the sweet spot with quick implementation, a high degree of customization, and lower long-term costs for maintenance and innovation.
Consider the various configurations you might need for your business:
- Can you enhance UI/UX functionality to match your business case?
- Can you preserve customizations implemented before the solution?
- Can you apply custom business logic?
Your Search solution should be perfectly tailored to your business case. If the solution lacks the ability to meet your vision, it’s a good idea to drop it from your list.
Personalizing search results allows you to bring together textual relevance, business relevance, and merchandising to meet your customers’ intent. It guides what users see in the future based on past user behavior.
A more personalized search experience helps customers discover more relevant content and discover new products and resources. This leads to higher conversions, larger average cart size, higher net promoter scores, and increased customer engagement.
Personalization features should:
- Be easy to implement and use
- Increase engagement
- Decrease site bounces
- Improve the customer journey
E-commerce site search tools can deliver powerful business insights. Once you have identified the goals you want to reach with your site search, look for tools that can deliver data relevant to those goals.
Your Search tool should:
- Help you track conversions, clicks, “no results” queries, bounce rate, search exit rate, average cart size, and more
- Display information in user-friendly dashboards
- Provide custom reports
- Help you drive improvements, such as highlighting certain products, creating content to fill key gaps, and improving the results page to discourage search exits
Without Search analytics, you cannot make data-driven decisions to improve your site.
Merchandising is an essential component of e-commerce strategy. The merchandising tools in your chosen site search tool should be intuitive enough for site owners to re-configure relevance and re-rank factors on their own to improve the buyer experience.
Some capabilities include, but are not limited to:
- Highlighting specific brands
- Ability to promote specific items or group of items for certain queries
- Promoting items in banners in search results
When choosing a site search solution, you should evaluate how the new tool will work with your current merchandising configurations. You’ve put time and effort into developing your merchandising campaigns, so your chosen solution should be able to support those campaigns effectively.
Vendors often point to machine learning (ML), artificial intelligence (AI), and natural language processing (NLP) to make their solution stand out. However, these buzzwords are only as important as their relevance and implementation.
Machine learning, artificial intelligence, and natural language processing are interrelated concepts. ML and NLP are technologies that help develop and improve the AI, which in turn provides intelligence to automation.
Some companies boast that AI can completely automate tweaks and refinements to your site search. While freeing up developer capacity seems beneficial, remember that you know your business best. Do you really want to AI to completely take over an important part of your online strategy?
AI should make suggestions and provide business intelligence that helps you make the best improvements. It should not make changes without your direct input.
When in doubt, request a demo and ask detailed questions about how AI, ML, and NLP are used by the solution. If the demonstrator can’t explain exactly how the technology works or tries to oversell you on its abilities, you should be wary.
As you evaluate site search solutions for your E-commerce site, consider how it will be implemented on the site itself.
If business requirements and technical needs are not well-defined, implementation can be a headache for the business and the very customers site search is meant to help. It’s essential to consider the following:
- What is your implementation timeline?
- What is required for implementation of the chosen solution?
- Do you have the appropriate developer resources to carry out implementation?
- Is the implementation supported by the vendor or is it an entirely self-serve process?
- How will the migration from your current set-up affect your current data, tracking, merchandising tools, or other customizations?
- Will there be any downtime during implementation?
Advance planning is the best way to ensure a successful implementation. Some solutions require you to buy servers or install software to get started; it’s important to know this as soon as possible so you can plan resources accordingly. Other solutions may be as simple as transferring some data and calling an API.
The more complicated the implementation process, the more crucial it is to take stock of developer capacity and skills and find out how involved the search provider’s teams will be. If you have a high level of customization or complex variables to consider, look for a site search solution that provides customized implementation services and support throughout the process.
Data drives many aspects of business, and search is no exception. Evaluate your current data needs and data state to understand how a new tool will fit in:
- What state is your data in now? Does it need to be cleaned or transformed before it can be used?
- Will the new site search solution accept a transfer of your data?
- How will the new solution incorporate or utilize your existing data?
- Does the new tool let you collect and analyze new data?
- How are important KPIs displayed in the search tool?
If you’ve collected search data with an old tool, check if the data can be integrated into your new tool. If this data is make or break for you, you’ll want to work with a solution that can integrate it.
You also want to look towards the future. Does the tool allow you to generate new data and provide you the tools to analyze it? If not, think about how this limits your ability to learn from search data and improve your search function over time.
Once the new tool is implemented, it needs ongoing management to ensure it’s performing optimally.
With a search as a service solution, maintenance and upkeep will be handled for you. New updates and innovations to the tool are immediately available thanks to the cloud-based infrastructure.
For most other tools, you will have to do this work yourself. You will be in charge of hardware costs, building your search tool to fit in your tech stack (if you build your own tool), implementing data changes and patches, and troubleshooting daily issues. In this case, you will want to consider:
- How long will it take for you and your team to learn the ins and outs of the product?
- What type of resources (hardware, software, developer team, etc.) are required for the new solution?
- Does this project require additional headcount or training? What about specialized skills?
- How will you continue to innovate and iterate on your search solution?
Consider the user who will be responsible for configuring controls and fine-tuning the system as needs change or fluctuate. Can this user intuitively use the system? Will the user be able to follow documentation on the product to complete unfamiliar tasks? If not, look for other options.
Even with dedicated resources, skilled team members, and experience, there always comes a time when you need additional support. Support should extend beyond implementation to cover unexpected problems.
Make sure that you have an acceptable level of support for both implementation and future needs. Common requirements to consider include:
- Will you need a dedicated expert or could you utilize product specific, detailed documentation? Keep in mind that both is an acceptable answer to this question.
- When is the support team available to you?
- How is the support team available (i.e. phone, email, chat, etc)?
- What SLAs (service level agreements) will be in place?
- Is support tiered? Is the tier basis urgency or package level?
- Does the solution have robust documentation for developers?
It’s important to note here that implementation and future needs don’t necessarily require the same type/level of support. You could use a high level of dedicated support for implementation, but only need a quick chat to guide you to the right self-service documentation in the future.
Ensure your chosen solution integrates with your current e-commerce platform and any extensions its configured with. Many site search work with common e-commerce platforms like Magento or Shopify, as well as common CMS platforms like WordPress and Zendesk—but nothing is guaranteed. Create a list of frameworks, platforms, and API requirements that must be supported by the site search tool you choose.
The question of cost is unavoidable, and may possibly greatly influence your decision. When comparing prices, keep in mind that cutting corners with essential features could cost you more in the long run. Look at price structures and decide what works best for your company—not just today, but in the future as your business and website grow.
As you evaluate site search solutions, consider the intangible benefits as well as the tangible. Great site search affects the customer experience, conversions, and ultimately the bottom line.
When deciding the best fit for your site search needs, there are three categories to consider: open-source, out-of-the-box, and search as a service:
- Open-source tools are free to use and download upfront, but will generate significant costs from personnel needs, hardware requirements, and resources devoted to designing, maintaining, and improving your search engine. Costs add up quickly for a labor-intensive search project like this. While initial costs may be lower, it will take a long time for your company to reap the benefits of your investment, assuming you get a polished tool.
- Out-of-the box tools attract users due to easy implementation and predictable pricing. Unfortunately, unexpected costs can arise from upkeep and maintenance. These tools also lack business case customizability. This means lost revenue if the tool can’t be customized to reach your customers most effectively.
- Search as a service tools offer managed implementation, continuous innovation, ongoing maintenance, and robust customization tools. Search as a service tools typically produce greater ROI over time than other approaches to site search.
So if your company has the resources to build its own custom search, why not take the plunge?
The truth is, building a custom search tool in-house is a complicated, costly project. Building your own search tool means you are responsible for:
- Building the tool from scratch.
- Implementing the search tool.
- Documenting the technical and user-facing requirements for the tool.
- Purchasing the hardware/software/add-ons required to make the search work.
- Handling all future development.
- Shouldering all maintenance and troubleshooting costs.
- Fine-tuning the relevance of the search engine.
Unless your developers have extensive experience building a search engine from scratch, this can be a daunting task, especially when things inevitable break and need troubleshooting.
While it may seem to be a less expensive solution to start, building your own tool will likely cost you more in the long run.
Still,there are certain times where building your own search is the best option, even if it’s expensive. For example, if you have complex, rigid, specialty requirements that cannot be adequately fulfilled by a vendor product. In this case, take time to speak directly to the vendor’s product teams to see if they would be an appropriate partner before you surprise your development team with a new project.
Whatever site search you choose, the vendor should work as a partner, with capabilities that drive your business goals and vision. The vendor should have reliable infrastructure and the capacity to support your future growth. They should also value innovation and constant improvement.
Overall, it’s about finding the best fit for your business case that gives you room to grow. Don’t settle until you do.
In addition to unparalleled speed and relevance, Algolia brings constant innovation to each of its search partners. With unparalleled developer resources and support, you can rest easy knowing your site search works every time, for every user. Watch our demo and read more about our features to see why e-commerce sites like LaCoste, Dior, and Birchbox trust Algolia.