Search

How to Make Search Your Site’s Greatest Asset

By zebethcontrolPosted on November 7, 2022Posted in Articles, SearchTagged site searchNo Comments

What makes a site truly brilliant? Impressive content? Sophisticated design? User-friendly interface? An effective support system for users, old and new alike? All this and more, my friend. No matter what you choose to build your site around, it can’t exist without a great search solution that helps guide every visitor to what they’re looking for – quickly, efficiently, and with as little effort on the site owner’s part as possible. In this article, we’ll go through the crème de la crème of the coolest features you can implement on your search with the help of Site Search 360, an easy-to-install and easier-to-maintain app fit for any site builder. Whether you have a HubSpot blog, a knowledge base maintained via Zendesk, a Shopify store, or all three at once, as long as your site’s content is searchable, this app is just what the doctor ordered! Top 5 search features for your site Search Result Categorization It’s highly likely that your site has tons of content that your users might be asked to search through. Depending on the number of pages you have accumulated over the years, that could require herculean patience. So, the first thing needed for your new search are separate tabs to neatly organize all the types of content you offer. Say you sell a million types of products. You wouldn’t want your users to scroll through all product categories mixed together as they search for their dream pair of shoes. Non-commercial sites can use this nifty technique, too – for instance, to put articles, YouTube videos, and blog posts in their own dedicated tabs. Or, as we call them in the search biz, Result Groups. Categorization via Result Groups is by no means limited to good old content types. Your search results might constitute pages from more than just a singular site – you could, for instance, have several interconnected domains for your primary content, FAQ knowledge base, news, etc. All of them have unique subsets of pages that you’d need your users to be able to search through, and as long as all these sites are included in your Data Sources, you can not only enable extensive cross-domain search, but also separate pages from these sites into dedicated groups for easier navigation. And the best part? You can even manually order these tabs to guide your site visitors to the categories you deem most important. So, how do you set this up? Easy – just enter the URL patterns of the page subsets you’d like to include in the same tab (or XPaths to specific elements found across all of these pages), give your brand-new Result Group a name, and you’re done: Repeat until all categories are in place. And here’s what your Result Groups can look like once implemented: Pro tip: If you ever feel like adding multiple search boxes to your site, you can limit each of them to specific Result Groups. You’ll then have, say, only products in the search results for the commercial part of the site, FAQ entries on the “About Us” page, etc. Configuration options are close to infinite! Filters and Ranking Your search is now organized into tabs. But that’s not the only thing you can do to make navigating your site’s content a piece of cake. Filters are a must-have when you want your users to be able to narrow down their search to instantly find exactly what they had in mind. Say someone’s looking for articles written by a specific author within a specific date range. With just a few clicks, you can create filters for both of these criteria (or anything in the world really – from prices to locations and beyond). These bad boys are configured differently for projects whose search results were generated either with a sitemap or through website crawling (low-touch integrations where the only thing we need to index your content is your site domain) and for those where a product feed was involved, turning each product into its own search result (best integrated over our API or through our extensions for various e-commerce platforms such as Shopify, Shopware, Adobe Commerce, and so on). For crawler-based integrations, filters are configured with Data Points, tidbits of information found across numerous pages that the crawler is pointed to via XPaths, URL patterns, linked and meta data, or even regular expressions. Data Points can be added to search result descriptions (across all pages as well as in specific Result Groups), used to automatically boost certain pages in your search results’ hierarchy, and, of course, they can direct the crawler to your future filter values. All of this can be configured right when a Data Point is created with a simple tick in the box of your choice. Here are the settings you can tinker with for each of your filters: And here’s your data point used simultaneously in the description of the product and as a filter: For e-com, things get even more exciting. Instead of Data Points, we extract and then use Product Facts, aka the various product characteristics (like color, material, vendor, etc.) available in your feed. The process is fully automated – no need to experiment with XPaths and regexes. It also comes with some ecom-exclusive perks such as HEX-coded circles next to “Color” filter values. An e-com filter configuration could look like this: Another pro tip for you: e-com and regular filters alike (as well as their values) can be reordered, and there’s even an option to exclude specific values from any filter. But the coolest part is that you get to choose how many pages should bear the values of a specific filter before that filter is triggered to pop up in the search. There really isn’t much of a point in showing the filter if it can only be applied to a singular page, now is there? In action, these filters are impressive to say the least. Filters are tightly connected to Ranking Strategies. Crawler-based integrations come with the option to sort results in ascending or descending order by any numeric Data Point such as “Price”. Sorting Options are configured in a very

Needl wants to become the search engine for your accounts • ZebethMedia

By zebethcontrolPosted on October 26, 2022Posted in Apps, productivity, Search, Startups, Y CombinatorNo Comments

Google, DuckDuckGo, and other search engines help you find information from the web. But it’s hard to find documents, messages, meetings, and emails from your own accounts. You need to go to different applications to find things that might be related to one project. A Y-Combinator-backed app called Needl is helping users with that. Needl is a cross-platform application that lets you search across your local filesystem and accounts like Gmail, Google Drive, Google Calendar, Notion, and Slack. The free version — available on the web, Windows, and Mac — lets you connect a single account per integration. If you need more account connections and integrations like Jira and Linear, you will need to pay $10 per month. The application is simple to set up and use: once you install it on your system, it will ask you to connect your Google, Slack, and Notion accounts. Once that’s done, you can search for files, events, emails, and other things across all these accounts and your local filesystem. You can filter these results by files, messages, events, tasks, and emails. Image Credits: Needl If you’re a keyboard ninja, the app has handy shortcuts for you to launch the interface and navigate around. Users can customize shortcuts to launch the app and jump to the home view. The default view on the app shows the Activity Feed, which will show you contextual information on different apps such as your upcoming meeting. Needl founders Max Keenan, Angela Liu, and James Liu are all Chicago university alums and met at a hackathon. They worked on a few side projects like a tool to write essays using GPT-2 and a TikTok for blog posts. After university, MaxKeenan in investment banking at Moelis while Angela Liu and James Liu joined Microsoft. The trio said that they had to become organized once they joined their jobs and meticulously follow naming systems and folder structures to easily find info. They wanted to solve this problem of constantly and manually reorganizing information through search. “We were looking for a problem that historically had never been solved, but improvements in language models would be able to solve. As we were onboarding virtually during the pandemic, it hit us right in the face — information was siloed across all of these different platforms and we could improve the search and discovery of info,” Keenan said in an email conversation with ZebethMedia. Image Credits: Needl Needl team wrote the first line of code in June when it was in the Y-combinator’s summer 2022 cohort. The company has raised $2.5 million from various investors including Fuse, Y Combinator, Palm Drive Capital, Liquid 2 Ventures, Collin Wallace and Nathan Wenzel. The company rolled out the product under a closed beta to around 200 users in August. Now the company is making it available to everyone under public beta. Keenan said the company wants to focus on improving its contextual and semantic search through large language models (LLM) over the next 12 months. Plus, the startup wants to add more premium integrations like Asana, Hubspot, and Salesforce. The startup considers Glean, a startup powering enterprise search across apps, as one of its major competitors. In May, Glean raised $100 million in its series C funding round led by Sequoia with participation from Lightspeed, General Catalyst, Kleiner Perkins, and the Slack Fund at a $1 billion valuation post-money. Keenan said that a major differentiation between Glean and Needl is the shorter setup time for the latter. “Biggest difference from Glean is that our product is self-serve and can be set up in under 2 mins by anyone, regardless of company size. Glean sells through sales-led processes that require full company adoption, can take months, and are inaccessible to individuals or small teams,” he said. Neeva, a search engine built by a former Google ad exec, also offers search features through app integrations. However, it is available only in the U.S. with European expansion underway. Keenan said in long term, Needl wants to pre-empt the need for search and present information through its own recommendation engine.

Meilisearch lands $15M investment to grow its ‘search-as-a-service’ business • ZebethMedia

By zebethcontrolPosted on October 10, 2022Posted in Apps, felicis, Meilisearch, Search, StartupsNo Comments

Meilisearch, the creator behind the open source search engine project of the same name, today closed a $15 million Series A round led by Felicis, with participation from CRV, LocalGlobe, ESOP, Mango Capital, Seedcamp and Vercel CEO Guillermo Rauch. CEO Quentin de Quelen tells ZebethMedia that the new cash will help to expand Paris-based Meilisearch’s marketing and sales teams as the company transitions to an “enterprise-focused” strategy. “For three years, we have created a product that brings a lot of value to developers, which has allowed us to form a strong community,” Quelen said via email. “The new money is to focus on the development of Meilisearch Cloud, our fully managed offering of Meilisearch instances. We will also continue to invest in our open source offering by releasing an ‘enterprise-ready’ version of Meilisearch by the beginning of 2023.” Quelen co-founded Meilisearch alongside Clément Renault and Thomas Payet, two friends from college, in 2018. The trio worked together on search tech at e-commerce startup Veepee and then at Louis Vuitton, where they quickly realized the intractable problem that building a search engine presented. “Building great search experiences has historically only been possible for companies with large tech resources,” Quelen said. “[Search is often] very hard and expensive for a team to maintain and tune.” In 2020, Quelen, Renault and Payet released Meilisearch, a search API based on their professional learnings and experiences. Available on GitHub, the project grew to over 10 million downloads, making it among the most popular open source search projects. Image Credits: Meilisearch Quelen asserts that, unlike Elasticsearch, and other freely available search engine frameworks, Meilisearch is designed for frontend applications across a broad swath of domains — not just narrow use cases like e-commerce discovery. Leveraging natural language processing, Meilisearch attempts to gain a better understanding of the queries that users make on whatever app, service or website a developer builds it into. Meilisearch supports major languages and ships with search filters, like price and date, as well as customizable ranking rules. It also corrects for typos and mistakes, ensuring errors in queries don’t adversely impact the search experience. Quelen claims that more than 10,000 apps today rely on Meilisearch. That’s impressive when considering the growing competition in the “search-as-a-service” space, which includes CommandBar, Algolia and Chameleon. “[W]e quickly proved that Meilisearch was long-awaited by developers who could not find simple and powerful solutions to improve the search experience in their applications,” he said. “The open source project shows a huge adoption from the developer community and [we’re] actively working on monetization around the open source project.” To that end, as Quelen alluded to, Meilisearch is upping its investment in Meilisearch Cloud, which is scheduled to launch in late November. In development over the past few months, Quelen says that Meilisearch Cloud — which offers the same experience as the open source Meilisearch but hosted on the public cloud, with prebuilt integrations — onboarded over 50 companies during a private beta. When asked about runway and revenue, Quelen declined to comment. But he said that Meilisearch will take a disciplined approach to burn, spending the capital it raised from the Series A over the course of the next two to three years. To date, Meilisearch has raised $22 million. It plans to expand its 25-person headcount to 30 by the end of the year and 50 by year-end 2023.