Build the Ultimate Search for Your Business, Just Like Google and Amazon

Improving the website search experience: speed, relevance and design are indispensable

Core points:

• An excellent website search experience is crucial and can be achieved without the need for a large team of search engineers. The focus of future searches is on user experience, which depends on UX designers and front-end engineers.
• The three pillars of an excellent search experience are: speed, relevance and design. Speed ​​refers to providing search results in milliseconds; correlation refers to matching queries with recorded searchable terms and business metrics; design refers to creating user-friendly search bars and results pages.
• The managed search engine Algolia allows any engineer to build an excellent search experience. It provides extremely fast search results (under 35 milliseconds) worldwide, ensures results relevance using optimized algorithms, and provides an analytical dashboard for fine-tuning and iterative search experiences.
• No matter how you build your search experience, the standards are the same. Users expect the search experience to work properly, whether on Google, Amazon or any other website.

(This article is sponsored by Algolia. Thanks to the sponsors who support us who made SitePoint possible.)

The first thing that comes to mind when many companies think about search is Google. They invest heavily in SEO and SEM to attract traffic, while their own website search experience is often overlooked. The default search options on platforms like Shopify and WordPress pale in comparison to Google and Amazon. The problem is, when most people think of improving their website search, they think of hiring a team of search engineers to create a hyper-personalized, machine-learning, predictive search experience—either invest heavily now or stick to default searches. ,Is it right? wrong. If your website has a bad customer experience inside it, investing in attracting traffic to your website is a bad idea – if you think “my search traffic is only a small part of my website, so why should I worry about it? ”, then your website search may need improvement. Remember: In 2016, 93% of Internet experiences began with search. The biggest mistake you can make is not thinking that “search is not important to my client’s digital experience”, but thinking that it takes a team of search engineers to get a good search experience. If you only remember one thing from this article, remember this: In 2017, search was no longer just a backend thing, it was an experience. The future of an excellent search experience is in the hands of UX designers and front-end engineers. Since 2012, Algolia has been committed to enabling any engineer to build a great search experience on top of our managed search engines, powering thousands of websites and products around the world. During this time, we educated our clients about the necessary requirements for a great search experience – great search is more than just fancy algorithms or hyper-personalization. Search is a very intuitive process. Users don't analyze the search experiences one by one - in fact, most users of your product may say that the search bar itself is a feature, rather than the individual components that make up the search bar - meaning your users are on the first time You can intuitively know whether your search is excellent or average after trying it. The difference between a general search and a great search is the three pillars of search - speed, relevance, and design.

Milliseconds are important

Algolia offers 99% of search results worldwide with speeds of less than 35 milliseconds - just ask our customers. Answering "How fast is it to be fast enough?" This question takes some "how" and some "why", but let's start with the latter: Why milliseconds matter:

• A 100 millisecond delay will cost Amazon 1% of sales
• A 500 millisecond delay will cause Google to lose 10% of its traffic
• 10 seconds: The attention duration of digital consumers

It's a race against time and you need to keep your audience engaged. In any step of the user experience, if they feel stalled, especially search, they have subconsciously compared your search experience to Google and Amazon, they will churn. You either get as fast as the user expects that they won't even notice; or you get slower and they will notice. In a conversation, if someone answers your question quickly, the conversation will continue smoothly; if someone stares at you blankly and asks you "How long will it take?", the conversation will Interrupt, the topic of the conversation will also change from what you are interested in to "Why does this take so long?"

"How" focuses more on the operational level. When the search query returns a speed of the result, three things are important:

1. The distance that the query must propagate
2. The size of the record that must be queried to determine the correlation
3. Complexity of algorithms used to determine correlation

For internal solutions, the distance to query the user depends on where you save the hosted website server. It doesn't make sense to deploy servers around the world to optimize website loading speed, so you need to copy your search solution in major markets to optimize search speed. Blink is about 180msCreating instant searches—an instant search experience that displays results immediately after typing—needs to deliver results faster than blinking. Your goal should be that 99% of queries are displayed in 35 milliseconds. For hosting solutions like Algolia, you need to make sure you have a Distributed Search Network (DSN)—the CDN of your search bar—this allows you to replicate you on up to 50 servers in every corner of the world based on the source of the user, depending on the source of the user, to replicate you on up to 50 servers in every corner of the world. search index.

Keep the record streamlined

The smaller the record, the faster it will pass through the engine. For search engines like Algolia, searching for 1 million single-segment blocks is faster than searching for 10,000 100-segment blocks—it can also improve relevance, pointing out where the query matches in a given text. Keeping a streamlined search machine means having a streamlined search algorithm. Algolia uses an algorithm that breaks the draw - a series of simple nested functions that will only be triggered in turn when the ranking output of the current function is tied. This means that if your first function (for example, for GeoSearch, distance user) returns a unique ranking for each result and there is no draw, the function stops there, fast and does not affect the correlation. Search engine frameworks such as Elastic Search and Solr use floating point scoring algorithms and are written by developers themselves. These algorithms may be simple at first, but with each iteration and improvement – ​​there is a little spelling tolerance here and a little synonym processing there – the algorithms will get bigger and slower, and not necessarily more Related.

Relevance

Relevance means finding what you are looking for. Relevance includes two aspects:

• Text Relevance: The degree of matching between searchable words of the query and record
• Business relevance (i.e. your business indicator): sales/views/likes, special objects, price, date...

Not all search engines are the same, and likewise, every search experience has its own unique balance of text and business relevance. Most importantly, the search framework you use is able to handle these two different types of data points. In addition to sorting by business/vanity metrics that matter most to users, you also need to make sure to deal with some more semantic challenges:

• Fault-tolerant of spelling: Don't let automatic corrections ruin the perfect search experience.
• Synonyms: UK users who search for "jumper" should be able to find your wonderful "sweaters" series.
• Long query: Users searching for "faded blue jeans" should not find the result because you only have "faded jeans".

Relevance is unique to every search experience, and understanding what your users are searching for, what they find and what they can’t find is more valuable than making any adjustments or modifications to your ranking algorithm. Algolia provides an analytics dashboard for all users, and in addition to providing critical data, it can also help you find queries with no or few results – which is often the most frustrating for users and when deciding where to focus on your website. The best starting point. Additionally, since Algolia uses a tie-breaking algorithm rather than floating point scores, it is possible to fine-tune and iterate the correlation directly in the dashboard without touching the code with every minor change.

Design

The UI/UX of the search bar may be the most important pillar, but this is often the most overlooked - mainly because searches usually focus on ranking optimization and speed rather than experience. But if the user has to press the "Enter" key every time to get the search results, then what is the use of lightning speed and related searches? Here are some of our best practices for a great search experience:

• Instant search results: Each key press should produce results. When combined with extremely fast speeds, users will feel that the search results are constantly improving every time they press the key.
• Highlight Results: Understanding the "reason" for you to get a given response may be as important as getting the correct result itself. Is the top-level result simply because the query appears in the secondary properties of the record? Display and highlight it conditionally so that the user knows why they see the given result.
• Provide context in the results: If you use business metrics (popularity, price, availability) in your ranking algorithm, make sure that users see this by showing these metrics in search results. Think of it as highlighting your ranking formula!
• Placeholder Text: Leave the placeholder text in the search bar before the user starts typing, letting the user know what they can search for—for example: “George Clooney, No Country for Old Men, Drama, 1920s”
• Multi-category search: The more ways users can succeed, the higher their chances of finding what they need. Even if you are just searching for articles in your WordPress blog, allowing users to search by author, category, title, and tag, it means you let users search the way they want, rather than letting them learn how your search works.

As with most creative industries, behind the scenes, design is a very specific series of steps. You can use this list to get beautiful search results and keep these design elements in mind when you try your favorite search bar (or even SitePoint). What is missing? What are your problems? How does design answer these questions?

Conclusion

For your end users, search is instinctive – it either works or does not work. For you, searching can be easy if you know what to look for. You don't have to be a search expert to get an excellent search experience. At Algolia, we believe the future of search lies in excellent UX/UI, which is why Algolia's integration is pre-configured with design best practices. A hosted search platform handles speed issues for you and makes tweaking relevance something anyone can do so you can focus on building a great app experience. No matter how you build your search, the criteria are the same. The pillars that define great searches—speed, relevance, and design— are not decided by Algolia, Google, or any other participant. They are determined by users, who expect the search experience to work properly, whether you are Google, Amazon, or any other app.

FAQs about search engines (FAQ)

What are the three pillars of search engines?

The three pillars of search engines are relevance, discovery and credibility. Relevance refers to the degree to which the web page content matches the searcher query. Discovery refers to how search engines find and index web pages. On the other hand, credibility refers to the credibility and authority of a web page. These three pillars jointly determine the ranking of web pages in search engine results.

How do search engines determine relevance?

Search engines determine relevance by analyzing web page content. They look at factors such as the use of keywords, the usage context of keywords, and the overall quality of content. The more relevant the web content is to searcher queries, the higher it ranks in search engine results.

What is the role of discovery in search?

Discovery is the process of search engines finding and indexing web pages. This is done through a process called crawling, where search engines send bots to browse the network and look for new or updated content. Once a web page is discovered, it can be indexed and made available in search engine results.

How do search engines evaluate credibility?

Search engines evaluate credibility by looking at factors such as the authority of a web page, the quality of content, and the quantity and quality of links to it. Web pages with high credibility are seen as trustworthy and authoritative, and are more likely to rank high in search engine results.

What are the alternative search engines other than Google?

While Google is the most popular search engine, there are several alternatives that offer different features and benefits. These include Bing, Yahoo, DuckDuckGo and Baidu. Each search engine has its own algorithm and ranking factors and may provide different results for the same search query.

How does search engine marketing work?

Search engine marketing is a strategy that promotes a website by improving the visibility of the website in search engine results. This can be achieved through paid advertising, search engine optimization, or a combination of both. The goal is to attract more traffic to the website, which can lead to increased sales and revenue.

What is Amazon SEO?

Amazon SEO is the process of optimizing product listings on Amazon to rank higher in Amazon's search results. This includes using relevant keywords, providing detailed product descriptions, and obtaining positive customer reviews. A well-optimized product list can improve visibility, click-through rates and sales.

What are the top search engines in the world?

The world's top search engines include Google, Bing, Yahoo, Baidu and Yandex. Google is the most widely used, but other search engines also have a large user base. Each search engine has its own unique features and advantages, and the search engine that best suits you depends on your specific needs and preferences.

How to improve the ranking of my website in search engines?

Improving your website's ranking in search engines includes optimizing your website for relevance, discovery, and trustworthiness. This includes using relevant keywords, creating high-quality content, obtaining high-quality backlinks, and ensuring your website is easily discovered by search engines.

What is the future of search engines?

The future of search engines may include more personalization, more integration with other services, and more use of artificial intelligence. Search engines are constantly evolving to deliver better, more relevant results and keep up with changes in technology and user behavior.

The above is the detailed content of Build the Ultimate Search for Your Business, Just Like Google and Amazon. For more information, please follow other related articles on the PHP Chinese website!