Blog posts by Rajveer Singh2025-08-04T06:22:37.0000000Zhttps://world.optimizely.com/blogs/rajveer-singh/Optimizely WorldUnderstanding Query Execution in Optimizely Search & Navigation<p> </p>
<h3>A Real-Life Example of How Search Works Behind the Scenes</h3>
<p>In today’s digital landscape, delivering accurate and relevant search results is critical to user experience. Optimizely Search & Navigation (formerly Episerver Find) offers a powerful search engine that enhances queries using features like <strong>synonyms</strong>, <strong>fuzzy matching</strong>, <strong>boosting</strong>, and <strong>best bets</strong>.</p>
<p>Let’s walk through how a search query is executed using a real-life example.</p>
<hr />
<h2>Real-Life Scenario</h2>
<p>Imagine a user visits an e-commerce site and searches for:</p>
<blockquote>
<p><strong>"iphon"</strong> (a typo for "iPhone")</p>
</blockquote>
<p>Despite the typo, the user expects to see relevant results like "iPhone 15 Pro", "Apple iPhone Case", etc. Here's how Optimizely processes this query:</p>
<hr />
<h2>Query Execution Flow</h2>
<h3>1️⃣ <strong>Synonym Expansion</strong></h3>
<p>Before the query hits the search engine, Optimizely checks for synonyms. If "iPhone" has synonyms like "smartphone" or "mobile", the query is expanded:</p>
<blockquote>
<p><strong>Query becomes:</strong> <code>"iphon OR iPhone OR smartphone OR mobile"</code></p>
</blockquote>
<p>This increases the chances of matching relevant content.</p>
<hr />
<h3>2️⃣ <strong>Fuzzy Matching</strong></h3>
<p>Next, fuzzy logic is applied to handle typos. The term <code>"iphon"</code> is recognized as a close match to <code>"iPhone"</code> using Levenshtein distance.</p>
<blockquote>
<p><strong>Match found:</strong> <code>"iPhone"</code> despite the typo.</p>
</blockquote>
<hr />
<h3>3️⃣ <strong>Document Retrieval</strong></h3>
<p>The search engine now retrieves all documents that match the expanded and fuzzy-corrected query terms.</p>
<blockquote>
<p><strong>Results include:</strong> iPhone 15 Pro, iPhone Cases, Apple Accessories, etc.</p>
</blockquote>
<hr />
<h3>4️⃣ <strong>Boosting</strong></h3>
<p>Boosting adjusts the relevance score of results. For example, if Apple products are boosted:</p>
<div>
<div class="scriptor-paragraph">
<pre><code class="language-CSharp">.BoostMatching(x => x.Brand, "Apple", 2.0)
</code></pre>
</div>
<!--ScriptorEndFragment--></div>
<blockquote>
<p><strong>Apple products appear higher</strong> in the results than others.</p>
</blockquote>
<hr />
<h3>5️⃣ <strong>Best Bets</strong></h3>
<p>If the admin has configured a best bet for the term "iPhone", such as the "iPhone 15 Pro" product page, it is promoted to the top of the results.</p>
<blockquote>
<p><strong>Best Bet Result:</strong> iPhone 15 Pro appears first, regardless of score.</p>
</blockquote>
<hr />
<h3>6️⃣ <strong>Final Results</strong></h3>
<p>The final result list is assembled, combining all the above enhancements and User sees accurate, relevant, and prioritized results:</p>
<ul>
<li>Synonym-expanded matches</li>
<li>Fuzzy-corrected terms</li>
<li>Boosted relevance scores</li>
<li>Best bet promotions</li>
</ul>
<p> </p>
<hr />
<p>This layered approach ensures that:</p>
<ul>
<li>Users get results even with typos.</li>
<li>Related terms are intelligently matched.</li>
<li>Important content is prioritized.</li>
<li>Admins can manually promote key pages.</li>
</ul>
<hr />
<h2>Summary Diagram</h2>
<pre><code>User Query → Synonym Expansion → Fuzzy Matching → Document Retrieval → Boosting → Best Bets → Final Results
</code></pre>
<hr />
<p> </p>2025-08-04T06:22:37.0000000ZBlog postEnvironment-specific badges in the Optimizely CMS <p><img src="/link/d8254a55289b4471880079eb203cccf8.aspx" /></p>
<h3>Step 1: Add Environment Info to the UI</h3>
<p>Create a custom <code>UIDescriptor</code> and a <code>ComponentDefinition</code> to inject a badge into the CMS UI.</p>
<pre><code class="csharp language-csharp">using EPiServer.Shell;
using EPiServer.Shell.ViewComposition;
[UIDescriptorRegistration]
public class EnvironmentBadgeUIDescriptor : UIDescriptor
{
public EnvironmentBadgeUIDescriptor()
: base("environment-badge")
{
DisabledViews = new string[] { };
}
}
</code></pre>
<h3> </h3>
<h3>Step 2: Add a Client Resource</h3>
<p>Create a JavaScript file that will render the badge in the CMS UI.</p>
<pre><code class="javascript language-javascript">define([
"dojo/_base/declare",
"epi/_Module",
"dojo/dom-construct",
"dojo/domReady!"
], function (declare, _Module, domConstruct) {
return declare([_Module], {
initialize: function () {
this.inherited(arguments);
var env = window.cmsEnvironment || "Production";
var color = {
Development: "#007bff",
Staging: "#ffc107",
Production: "#28a745"
}[env] || "#6c757d";
var badge = domConstruct.create("div", {
innerHTML: env,
style: `
position: fixed;
top: 10px;
right: 10px;
background-color: ${color};
color: white;
padding: 5px 10px;
font-weight: bold;
border-radius: 4px;
z-index: 9999;
`
}, document.body);
}
});
});
</code></pre>
<h3> </h3>
<h3>Step 3: Register the Module in <code>module.config</code></h3>
<pre><code class="xml language-xml"><module>
<clientModule initializer="environment-badge">
<requiredResources>
<add name="epi.shell" />
</requiredResources>
</clientModule>
</module>
</code></pre>
<h3> </h3>
<h3>Step 4: Set the Environment in Razor or Layout</h3>
<p>In your <code>_Layout.cshtml</code> or a shared Razor view:</p>
<pre><code class="csharp language-csharp"><script>
window.cmsEnvironment = '@Environment.GetEnvironmentVariable("ASPNETCORE_ENVIRONMENT")';
</script>
</code></pre>
<h3> </h3>
<p> </p>
<p>When editors log into the CMS, they’ll see a badge in the top-right corner showing the current environment, color-coded for clarity.</p>
<p> </p>2025-07-01T07:09:44.0000000ZBlog postBoosting by published date with Relevance<p><strong>Goal:</strong> To ensure that the latest published or last updated content is ranked higher in your search results, you can modify your query to include a <strong>boost based on recency</strong>. This typically involves boosting content based on a date field such as <code>PublishedDate</code> or <code>LastUpdated</code>.</p>
<div>
<p>Here’s how you can update your query to prioritize recent content while still considering relevance:</p>
<pre><code class="csharp language-csharp">var result = query.ApplyBestBets()
.BoostMatching(x => ((IContent)x).SearchTitle().MatchCaseInsensitive(contentFilter.SearchQuery), 1.5)
.OrderByDescending(x => ((IContent)x).LastUpdated) // Boost by recency
.Take(pageSize)
.Skip((pageNo - 1) * pageSize)
.StaticallyCacheFor(TimeSpan.FromSeconds(CacheSpan))
.GetResult(hitSpec);
</code></pre>
<h3>Explanation:</h3>
<ul>
<li><code>OrderByDescending(x => ((IContent)x).LastUpdated)</code> ensures that the most recently updated content appears first.</li>
<li>If you want to balance <strong>relevance and recency</strong>, you might consider a <strong>custom scoring function</strong> or a <strong>composite boost</strong> that combines both.</li>
</ul>
<p> </p>
<p>------------</p>
<p> </p>
<p>Also, we can implementing a more advanced scoring strategy that balances relevance and freshness as below:<br /><br /></p>
<div>
<p>To implement a <strong>balanced scoring strategy</strong> that considers both <strong>relevance</strong> and <strong>recency</strong>, you can use a <strong>custom boost function</strong>. This approach allows you to:</p>
<ul>
<li>Boost content that matches the search query (relevance).</li>
<li>Boost content that is recently published or updated (recency).</li>
<li>Combine both into a composite score.</li>
</ul>
<p>Here’s how you can update your query using a <strong>custom boost expression</strong>:</p>
<pre><code class="csharp language-csharp">var result = query.ApplyBestBets()
.BoostMatching(x => ((IContent)x).SearchTitle().MatchCaseInsensitive(contentFilter.SearchQuery), 1.5)
.OrderByDescending(x =>
((IContent)x).SearchTitle().MatchCaseInsensitive(contentFilter.SearchQuery) ? 1 : 0 +
GetRecencyScore(((IContent)x).LastUpdated)
)
.Take(pageSize)
.Skip((pageNo - 1) * pageSize)
.StaticallyCacheFor(TimeSpan.FromSeconds(CacheSpan))
.GetResult(hitSpec);
</code></pre>
<h3>You’ll need to define <code>GetRecencyScore</code> like this:</h3>
<pre><code class="csharp language-csharp">private double GetRecencyScore(DateTime lastUpdated)
{
var daysOld = (DateTime.UtcNow - lastUpdated).TotalDays;
// Example scoring: newer content gets higher score
if (daysOld <= 1) return 1.0;
if (daysOld <= 7) return 0.8;
if (daysOld <= 30) return 0.5;
if (daysOld <= 90) return 0.3;
return 0.1;
}
</code></pre>
<h3>Why this works:</h3>
<ul>
<li>It gives a <strong>higher score to recent content</strong>, but still considers <strong>query match</strong>.</li>
<li>You can <strong>tune the weights</strong> in <code>GetRecencyScore</code> to fit your needs.</li>
</ul>
</div>
</div>
<p><br /><br /><br /></p>2025-07-01T07:04:27.0000000ZBlog post