Changelog

Follow up on the latest improvements and updates.

RSS

This update focuses on improving usability, navigation, and overall consistency across the platform, making everyday workflows smoother and more intuitive.
1. Improved Cluster Filtering:
Added additional cluster-specific filters for more precise data exploration, including some long awaiting Filters like Clusters with OR without active redirect.
special_filters
2. Faster Query Navigation:
Introduced a quick link from Intent Details directly to the corresponding query in the table. !
query-highlight-chrome-capture-2026-11-19
3. Consistent Cluster Creation:
The "Create New Cluster" dialog now automatically converts input to lowercase for consistent naming.
4. Better Query Visibility:
Added a Valid Queries Marker to help users quickly identify valid queries.
fully-validated - fully validated vs. partially validated partially-validated
5. Easier Access to Updates:
Added a direct link to the SmartQuery changelog in the system drawer.
system-status
6. Safer Redirect Configuration:
Added feedback and confirmation prompts when configuring redirects that might lead to an incorrect shop or domain. redirect-url-checker
💡 Overall Impact
A more intuitive, consistent, and reliable user experience, with improved navigation, clearer feedback, and greater visibility across key workflows.
Last night, we performed scheduled maintenance on the searchHub UI, introducing several improvements to enhance usability and streamline workflows.
While most of the changes happened behind the scenes and may not be immediately noticeable, we've also made significant improvements to two key views based on your valuable feedback.
1. Introducing the New "Intent Cluster" View
The former "Cluster View" has been completely redesigned and replaced by the new "Intent Cluster" View. Based on your feedback, we've reworked the view with a strong focus on task workflows, clarity, and usability.
Bildschirmfoto 2026-10-15 um 10
Our main objectives were to:
  1. Simplify task selection: Make it easier to identify and select the right task for a specific problem.
  2. Reduce errors: Introduce more intuitive and less error-prone workflows.
  3. Increase transparency: Provide a clearer overview of intent clusters, available actions, and their implications.
The result is a more intuitive and streamlined experience designed to help you work more efficiently and confidently.
2. Enhancements to the "Redirects" View
We've also introduced a few improvements to the "Redirects" View to make managing and analyzing redirects easier.
  1. Predefined Filters: Quickly access the most commonly used filters without having to configure them manually.
Bildschirmfoto 2026-10-15 um 10
  1. New "Cluster" Column: Easily identify and group redirects associated with the same Intent Cluster. This is particularly useful for spotting cases where multiple redirects exist within a single cluster.
Bildschirmfoto 2026-10-15 um 10
These enhancements provide better visibility into your redirects and make it easier to identify potential overlaps.
Your Feedback Makes the Difference
Both updates were directly inspired by your feedback and are part of our ongoing efforts to make searchHub more intuitive, transparent, and efficient.
We're looking forward to hearing what you think about the improvements!
As more and more customers use Info Markers in the Performance Charts view to add context about their systems such as configuration changes, A/B tests, and feature rollouts, we noticed that multiple markers within a short time frame could be difficult to identify and read due to text overflow.
To improve clarity, we have updated the UI so that Info Markers are now displayed in the detail view, providing significantly better readability and greater flexibility.
chart-marker-2026-07-14
We hope you enjoy this improvement and continue adding markers to your charts.
We've improved trend visualization for KPIs.
Based on user feedback, we introduced a trend line that better reflects how key metrics evolve over time. Instead of only comparing the first and last value in a time series, the new trend line uses linear regression to calculate the overall trajectory over the defined period.
This makes it easier to identify whether a KPI is generally increasing, decreasing, or remaining stable. By visualizing the true underlying trend rather than just connecting the endpoints, this update helps teams quickly assess metric development, ignore short-term outliers, and understand whether actions are having the desired impact.
Expected Range_Development of Suggest-CTR_2026-01-01-2026-07-08
Today, we’re excited to announce that we have quietly launched our new
0.5B DE-UserQueryLLM
a model we have been working on for the past 12 months.
Why we built it:
In ecommerce, product data usually follows a clear structure. Whether for SEO, GEO, PIM systems, or catalog management, products are organized, enriched, and standardized. User queries, however, are not.
This gap creates a major opportunity: aligning user queries with a structured representation can unlock significant value for ecommerce and retail players and their users.
Structure improves navigation:
Structured queries are easier and faster for humans to scan. This is especially visible in autosuggestions: when suggestions follow a recognizable structure, users understand them more quickly and learn the underlying query patterns over time. As a result, they can formulate, refine, and expand their searches more effectively. This led to significant uplifts across the board in user engagement.
autosuggest_ctr_uplift
Structure accelerates aggregated knowledge:
Ecommerce engagement data is often extremely sparse, making it difficult to identify meaningful trends, behavioral shifts, and intent patterns. By structuring queries and especially by normalizing their formulation we can dramatically reduce the long tail of events and shift engagement mass toward the short head and mid-tail.
This significantly strengthens behavioral feedback loops such as learning-to-rank, embedding-model fine-tuning, and causal analytics.
query_short-head_stableness
Structure saves compute:
We continuously strive for higher computational efficiency. Query understanding is a key part of this.
Query Understanding combined with hybrid search based on embeddings has become the de-facto standard for product discovery, but most base models are still primarily trained on structured content. This content structure often does not align with the language and structure of real user queries, which makes it harder for models to learn, adapt, and generalize.
By aligning user-query structure with product-content structure, we can substantially reduce computational effort through better caching and more reusable representations. Having more structurally aligned labels reduced our query-understanding API calls by roughly 20% whilst increasing recall over 10%.
Closing Words:
To our knowledge, we are the first to have successfully achieved this kind of task at this scale. Therefore a big thank you to the whole team.
We’re excited to bring even more downstream capabilities to our customers, powered by our new
UserQueryLLM
.
We’re excited to introduce the new Query Explorer.
Until now, SearchInsights has focused on surfacing the most impactful data points related to search performance. However, feedback from our customers has shown that search and intent data can support many additional use cases beyond search optimization.
Topics such as seasonal demand analysis, keyword analysis, and category topic mapping help different teams across your organization work more efficiently and make better-informed decisions.
That's why we're happy to introduce the new Query Explorer view: In addition to the top 5.000 top records, you can browse in the New Query Explorer through all data to get deeper insights, without limitations. A more flexible way to access your raw query data, apply powerful filters, uncover deeper insights, and draw meaningful conclusions.
queryExplorer-capture-2026-06-29
We are very proud to officially announce the release of SmartSuggest 2.5. Our Auto-suggestions always created significant impact by modeling intent, preventing search failures, and guiding users toward high-value discovery.
With this release we push the boundary one step further on our path from Autocomplete to Intent Modeling:
1. Intent-Aware Product Suggestions
Traditional keyword based product suggestions ignore evolving intent and fail on ambiguous queries. SmartSuggest combines intent similarity with behavioral signals, continuously improving product suggestion relevance.
product suggestions
2. Scoped Suggestions
For broad intent, SmartSuggest recommends relevant categories or contextual scopes before search execution preventing unfocused result pages and bounce.
scoped suggestions
3. Contextual Pre-Suggestions
Intent often starts before typing. SmartSuggest now delivers pre-suggestions based on the current or last recent user context (e.g., category page, campaign entry, viewed product), with safe fallbacks when context is limited or unknown.
contextual pre-suggestions
Bottom Line
SmartSuggest transforms auto-suggestions into an intent-driven conversion engine, combining real-time learning, guided discovery, and merchandising control to reduce search failure and accelerate purchase decisions, effectively turning the suggestion bar into the most efficient shopping assistant in the store.
As you may already know, we’ve been working intensively behind the scenes to stay ahead of the competition in the areas of Auto-Complete and Auto-Suggestions. Very soon, we’ll unveil what this means and how it will help you further enhance user experience and product discovery.
Before introducing these bigger updates, we’ve focused on aligning this feature with our new UX & UI guidelines, already rolled out across several other views.
Our main goals were clarity and transparency. To achieve this:
  • We’ve separated traditional Query Suggestions from Pre-Suggestions (those shown before any typing begins).
  • We now clearly indicate the retrieval strategy behind each displayed suggestion.
  • The configuration view has been redesigned to better reflect the different configuration types and make it easier to manage.
Bildschirmfoto 2025-11-12 um 14
Bildschirmfoto 2025-11-12 um 14
Many of our users’ tasks start with data sourced directly from searchInsights.
Until now, most tables were limited to the top 1,000 records due to latency constraints we didn’t want to exceed.
Thanks to improvements in our underlying stack, we’ve been able to raise this limit to the top 5,000 records without significantly affecting latency.
This means you can now access and work with 5x more data directly in the searchHub UI, helping you deliver tasks with even greater effectiveness and efficiency.
Since its launch,
Queries by Category
has become one of the most valuable data sources for assortment planning and demand analysis. The feedback from our customers has been outstanding, and we wanted to give something back.
Starting today, we’re introducing
the ability to compare data over time
. This new feature allows customers to track how demand in their assortment evolves, making it easier to see whether specific marketing campaigns or category-related changes have actually driven improvements.
chrome-capture-2025-09-29
Load More
→