Skip to main content

Release notes

Last updated 10/05/2026

v6.1​

Key features​

Experiment suite (value-added feature)​

The experiment suite is ThinkingEngine's newly developed suite of product features. With A/B testing, you can make better decisions faster based on reliable data instead of guesswork.

The experiment suite supports the following capabilities:

  • Traffic splitting by IDs such as user, account, and device IDs
  • Orthogonal traffic splitting across experiments to prevent them from affecting each other
  • Confidence statistics on metrics to ensure scientifically sound conclusions

Analytics module​

  1. Supports configuring non-integer time zones

Regions with non-integer time zones, including India (5.5), are supported, and you can switch to them during analysis

  1. Added a scenario analysis model: Revenue analysis

Shows revenue metrics such as cumulative payer rate, LTV, LTV multiple, and ROI side by side

  1. Behavioral cohorts support complex selection conditions

When you create a cohort, you can combine behavior conditions and user property conditions, and reference other tags or cohorts

  1. Added Session Replay (value-added feature)

Session Replay reconstructs a user's visit to a web page and shows all information related to that visit in one place

BI suite​

  1. Supports configuring and pushing summary pages

Present data performance as dynamic text and use it together with visual charts

  1. Supports one-click AI beautification of BI dashboard layouts and styles

Generate a custom theme through multiple rounds of follow-up questions, and fine-tune it manually as well

Improvements​

Analytics module​

  1. Exchange rates support custom properties
  2. Analysis models support sharing temporary links
  3. Weekly and monthly aggregation supports turning off Complete incomplete periods to count only the data within the selected date range
  4. Supports configuring date shortcuts
  5. Date comparison supports the same period last year
  6. Funnel Analysis supports multiple group-by items
  7. Funnel Analysis supports grouping by any step
  8. The Funnel Analysis trend table supports showing data by step
  9. Attribution Analysis supports drilling down to user lists
  10. Leaderboards support showing additional properties

BI suite​

  1. Drop-down filters support adding options manually
  2. Tables support non-numeric metrics
  3. The number of rows a chart queries supports a custom top N for each chart

v6.0​

Key features​

Agent​

This update introduces the Agent intelligent conversation system, which supports two modes: custom conversations and Agent conversations.

1. Provides three-level management of models, MCP, and Skills, with system-level, company-level, and personal configuration.

2. MCP and Skill capabilities connect Agent conversations with data in the Analytics & Engage platform, enabling "data analysis through natural language".

3. Added Agent management, which lets you flexibly combine models, skills, and external services. Administrators can manage members, query usage, manage sandboxes, and configure channels in the System Admin module, making AI capability integration and management much more efficient.

Analytics module​

1. Project members can be assigned multiple roles

Analytics roles and Engage roles are removed. You can customize the feature permissions within a role and assign multiple roles to a single member, making member permission management more flexible and efficient

2. Assets usage governance supports batch operations (value-added feature)

After you filter to get the list of assets that need governance, you can perform batch operations on them as needed, such as freezing and deleting

3. Report versions can be rolled back

You can load historical versions of report content, so changes are traceable

4. Pre-computed dashboards now have task logs

View pre-computation duration and task error information to help optimize dashboard performance

5. Tags/cohorts support more scheduled update frequencies

Supports weekly, monthly, or multiple daily updates

Engage module​

1. Supports custom Engage role permissions

Bundle Engage feature permissions however you need and grant feature access flexibly

2. Journeys support node-level effect metrics

Measure how behavioral metrics perform after operational interventions on users

3. When you run push experiments with operation tasks, you can set an activation event

Accounts for the traffic loss between successful pushes and experiment exposures, eliminating the statistical problem of diluted enrolled users

BI suite​

1. Enhanced metric cards

Configure secondary metrics and set conditional formatting, prefixes and suffixes, and decorative icons

2. New chart types: progress bar, circular progress bar, liquid fill chart, and gauge chart

Better visualization of progress data

3. Supports configuring parameter controllers

Adjust sheet content with parameter controllers

4. Supports configuring common metrics

Common metrics can be reused across multiple charts on the same sheet. Change a metric once, and the change applies everywhere

DataOps Platform​

1. Data source integration supports more types

Added support for reading from Feishu Bitable as a data source, and for SelectDB / MongoDB / StarRocks as write targets

2. Workspace tables support cloud storage

Workspace tables support cloud object storage (such as OSS and S3)

3. Warehouse Management supports data authorization

Warehouse Management in System Settings can authorize the data assets in a warehouse for use by spaces and projects

4. DataOps spaces can authorize data to other spaces/projects

Data authorization across projects or spaces is supported. Authorized projects and spaces can use the authorized assets for further development

5. Supports backfill

The Ops module supports whole-flow backfill, meeting the need to update data for specific baseline dates

6. Added the Table Management module

The DataOps Platform adds the Catalog module for standardized management of the database and table resources in a space

7. Supports AI-assisted creation of tables with the same structure

Create tables with the same structure in target data sources (StarRocks, SelectDB) and write the data at the same time

Omni Insights Platform​

1. Visual Analysis

Gain multidimensional, configurable, in-depth data insights with visualization tools

2. Daily Summary

Daily Summary focuses on each day's key developments, quickly presenting the day's popular topics and AI-distilled opinions from player discussions, with drill-down support.

3. Community Risk Management rule editing

You can add and delete risk management categories and edit their names and content

4. Keyword system upgrade

Import and export, bulk show/hide, and sorting by frequency

5. Support for review data

The system will connect player review data from app stores and game communities (such as TapTap and Steam), and will gradually improve the display, analysis, and application features for this data to expand high-value insight dimensions.

Component updates​

  1. zookeeper upgraded to 3.7.2; mysql, kafka, and hdfs can optionally be updated

Improvements​

Analytics module​

  1. Engage monitoring in System Settings adds a Test feature
  2. Analytics projects now include Description information
  3. Batch configuration Excel files in Model Mapping Assets Configuration support the Upload and update configuration logic
  4. Supports modifying event descriptions via OpenAPI
  5. Dashboards, folders, and spaces can be added to favorites
  6. Dashboards and reports show data governance anomalies
  7. The date picker supports viewing history & saving frequently used date ranges as favorites

Engage module​

  1. Content templates support setting default values
  2. Content templates support required settings at the object child property level
  3. Operation tasks/journeys show an execution status progress bar
  4. Journey node details support exited-user statistics and drill-down
  5. Journeys support delay waits at second-level granularity

BI suite​

  1. Supports copying a single component and pasting it on any page
  2. Filter conditions for text fields support Empty string and Not empty string conditions
  3. Drop-down filters support two display modes: Compact and Show all
  4. Cross-page query containers support two display modes: Tile and Collapse
  5. Tables/pivot tables show the total row count

DataOps Platform​

  1. Flow release packages show the published content items
  2. Supports creating space task tables of the View type. You can create views based on mounted Link source tables and trim rows and columns for greater flexibility
  3. Added the preset parameter ${env}, which uses the value of the corresponding environment (product/dev) and can be used in Flow node scripts and table creation DDL
  4. Supports modifying the field types of workspace tables
  5. Supports modifying the partition field and logical primary key settings of workspace tables

Omni Insights Platform​

  1. Tags add a Task Scope section for limiting the execution scope of tasks
  2. Improved the quality of AI topic generation for more accurate content matching
  3. Added preset channels: forums (posts & videos): WeChat Game Circle; live streaming: Huya and Douyin

v5.0​

Analytics module​

Key features

1. Team Space supports multi-level structures

Team Space folders can be nested at multiple levels, and you can set up collaboration on any folder or dashboard

2. Set up Public Links for dashboards

Choose which dashboards allow public access, generate public links, and strictly control how they are used

3. Metadata creation and connection management

Upgraded metadata system: you can create metadata such as events and event properties and build dashboards and reports before the actual data is connected, and then manage the connections between metadata and the underlying data after the data is reported

4. Supports verifying assets for easier use

Verify assets in a project that have clear, well-defined definitions to make them easier to use

5. Model mapping asset integration (value-added feature)

Works with the DataOps Platform to convert external data sources into AE standard assets that can be used directly in analysis

6. Assets usage governance (value-added feature)

Lists all types of assets in the AE system and shows their lineage dependencies and query usage to facilitate asset governance

Improvements

  1. Dashboards and folders add the Admin role, who can share dashboards and folders
  2. Batch-manage all of your personal dashboards
  3. Expanded the customizable scope of role permissions

Engage module​

Key features

1. Added Journeys

Customize end-to-end, multi-touchpoint, differentiated automated operation strategies through visual orchestration

2. Added a data analysis module to Config Center

Generate data reports on the effect of delivered configurations directly through event tracking and metric settings.

3. Supports batch operations on config strategies

Manage multiple strategies quickly with batch operations.

Improvements

  1. Operation task effect metrics support formulas
  2. Operation task touch funnels support custom events
  3. Push notifications support adding channels for multiple apps
  4. Delivery timing supports updatable events and first events
  5. Client-side triggered tasks support custom pull timing
  6. Enhanced trigger calculation for client-side triggered tasks
  7. Client-side triggered tasks and Config Center add SDK integration for Douyin, Kuaishou, and Alipay mini games/mini programs
  8. Config templates add a long text input method

[New] DataOps Platform​

Contact your Customer Success Manager to try this module

Key features

1. Integrate data between various data sources and AE

Supports data sync between many common data sources, such as MySQL, OSS, S3, Redis, and ClickHouse

2. Flow management and scheduling

Build batch processing Flows visually with Workflow, with cron and interval scheduling policies for Flows, custom exception handling strategies, and task parameters

3. Dual-environment isolation

Isolated production and development environments with release management improve development efficiency and keep production data flows stable

4. Task instance monitoring and management

Real-time monitoring and management of task instances throughout their lifecycle, from creation to termination

5. Application warehouse with higher concurrent query performance

Sync table data to the application warehouse with one click to support BI dashboard scenarios that need higher query performance

6. SQL IDE tool connected to all AE lakehouses

In the IDE module, query and extract data in AE's built-in lakehouse and built-in application warehouse, manage database and table resources in linked catalogs, and manage frequently used SQL scripts

[New] BI dashboard​

Contact your Customer Success Manager to try this module

Key features

1. Diverse data sources

Create sheets with SQL statements based on different data sources, such as behavioral data and business data. Workspace tables from the DataOps Platform are supported

2. Personalized chart components

Add many types of charts, such as tables, pivot tables, metric cards, line charts, and bar charts, and customize their content and style to suit your business needs

3. Filter components

Add text and date filter components, specify where they apply, and switch between different views of the data

4. Multimedia components

Embed images or websites in BI dashboards to help explain business data

5. Comprehensive access control

Apply row-level filtering to sheet data to limit the data each member can see and keep your data secure

[New] Community Insights​

Contact your Customer Success Manager to try this module

Key features

  1. Covers mainstream media channels in China as well as vertical gaming communities, supports collecting live stream data from live-streaming platforms, and supports connecting private-domain data and first-party data (such as in-game chat and data from your own community)
  2. Search community content details efficiently and view the content, comments, and LLM processing results such as sentiment analysis and keywords to understand what players are saying at a granular level.
  3. Supports LLM post-processing and visual analysis of community data for fine-grained analysis of community content, helping you see the full picture of the community from multiple perspectives.
  4. Collects and analyzes live stream data of game streamers in real time and records the engagement of every live stream, giving you insight into how streamers' live streams perform.
  5. Automatically discovers trending topics in game communities and tracks the issues players care about, helping you better understand the current community ecosystem.
  6. A highly real-time community risk monitoring and alerting system with customizable rules helps operations staff quickly pinpoint and stay on top of risks and avoid community sentiment risks.

OpenAPI changes​

Background: In V5.0, role permissions are fully customizable and no longer need to be based on the preset role baseRole (Member, Analyst, Analytics Admin)

[Deprecated]

/open/get-base-role-functions is no longer supported. Use /open/get-all-functions instead

[Added]

/open/get-all-functions Get the list of all available functions

[Changed]

  • /open/create-project-role

Removed parameter: baseRole

  • /open/update-project-role

Removed parameter: baseRole

  • /open/delete-project-role

Added parameter (required): newRoleName

When you delete a role, you must specify the new role to hand over to

Component updates​

Receiver

A single reported data record can't exceed 1M. Larger records can't be stored

StarRocks Connector

  • Supports reading materialized views
  • Supports LDAP authentication
  • Supports writing via StreamLoad
  • Provides table statistics to the CBO optimizer

Hive Connector

  • Allows specifying the hive table write path (OSS/S3/HDFS) via location

Iceberg connector

  • Added the $partitions hidden column
  • Added the system.bucket() function

Kudu Connector

  • Allows users to configure the Kudu Per Scan size
  • Relaxed the Kudu Statistic limit

Exchange

Allows customers to configure exchange.hdfs.replication to reduce the number of FTE replicas on HDFS

v4.5​

Improvements​

1. Analytics module - Attribution Analysis supports selecting the same event

The same tracked event can be selected as both the conversion event and a touchpoint event and distinguished by filter conditions, meeting attribution needs in more reporting scenarios

2. Analytics module - Dashboard data push supports leaderboards

When a dashboard is pushed, if CSV File is selected, leaderboard reports also generate corresponding CSV files, just like reports from other general models

v4.4​

Key features​

Analytics module​

  1. Dashboards support weekly/monthly scheduled pre-computation

    In addition to daily pre-computation, options to update weekly, monthly, or every N days have been added to better support lower-frequency update scenarios such as weekly and monthly reports. Dashboard pre-computation can now also run at any time instead of only on the hour, making configuration more flexible.

  2. Custom dashboard cache duration

    For dashboards that are not pre-computed, this release adds a custom cache time for calculation results. You can set the cache time by minute, hour, or day, so that queries with the same conditions within the cache time reuse the same results, which avoids repeated calculations and reduces waiting. For dashboards that don't need to be updated, you can set the cache to Long Term, so the same data is always shown as long as the dashboard conditions don't change.

  3. Supports handing over dashboards/reports

    You can now hand over dashboards and reports to others in the same project for a quick handoff. The recipient becomes the new owner and can share, move, edit, and otherwise manage them.

Engage module (contact your Customer Success Manager to try this module)​

  1. Supports client-side triggered tasks

    Meets millisecond-level trigger needs in in-app scenarios such as new user registration and character creation, and supports real-time A/B traffic splitting

  2. Supports user behavior sequence triggers

    Reach users immediately after they complete a specific combination of actions in a behavior sequence

  3. Supports exporting tasks and campaigns across projects

    Export operation tasks and campaigns across projects to avoid inefficient repeated configuration

  4. Push content supports trigger values

    Assemble push content from trigger-related statistics or property values of the trigger event to dynamically push more targeted content

  5. Triggered tasks support dimension table properties

    Triggered tasks support real-time calculation of dimension table properties, using dimension table data processing to cover more audience selection scenarios

  6. Enhanced Custom Audience for scheduled tasks

    In Audience for scheduled tasks, custom audience conditions support more asset types and real-time estimation

Improvements​

  1. Analytics module - View when report results were calculated
  2. Analytics module - Scheduled pre-computation supports setting start and end dates
  3. Analytics module - Supports subscribing to cluster error data and data backlog status
  4. Analytics module - Filter conditions on time properties support selecting future times
  5. Engage module - The operation task list shows push data from the most recent task run
  6. Engage module - Operation tasks support downloading push error details
  7. Engage module - Send Test supports including channel custom parameters
  8. Engage module - Operation tasks support automatic content sync in Multi-language configuration
  9. Engage module - Improved interaction for Audience in triggered tasks
  10. Engage module - A/B test data can be aggregated by experiment group

v4.3​

Key features​

Analytics module​

  1. Added Heatmap

Analyze how player behavior is distributed across game maps, validate how well the map design works, and find areas for improvement.

  1. Added Leaderboard

Use any property as the ranking subject and set sorting rules to easily get a Top N leaderboard.

  1. Comprehensive improvements to Retention Analysis

Easily build metrics such as LT (user lifetime) and cumulative payer rate to better evaluate users' long-term performance.

  1. Dashboards support fixed dates

Set a single time range for all reports in a dashboard, which is useful for scenarios such as viewing overall campaign data.

  1. Data management supports data interception

Intercept event reporting by event name to avoid dirty data, and stop data receiving for projects that no longer need it to save costs.

  1. Added project-level chart color schemes

Choose a chart color theme and apply it to all charts in the current project.

Engage module​

  1. Added A/B horse-racing tests

Triggered tasks can test multiple variants on a small scale in experiment groups, and then push the best variant to the entire audience.

  1. Added custom triggers as a push type

Reach users with pushes based on your own custom trigger scenarios (for example, when a player's stamina is full).

  1. Improved audience capabilities for triggered tasks

The audience of triggered tasks can use custom properties or cohorts.

  1. Added operation effect metrics

Operation tasks show metric performance after any touchpoint node, helping you assess operational effectiveness.

Improvements​

  1. Analytics module - Trend charts support reference lines
  2. Analytics module - Trend charts support configuring upper and lower Y-axis limits
  3. Analytics module - Pie charts support showing a custom top N
  4. Analytics module - Interval Analysis supports custom interval settings
  5. Analytics module - Interval Analysis can show more percentile metrics
  6. Analytics module - Fields in the visualization module support multi-select drag and drop
  7. Analytics module - Dashboard notes support image uploads
  8. Engage module - Operation task triggers support custom properties
  9. AWS Graviton optimization: Supports Amazon Graviton instances and Amazon Linux2023, providing more cost-effective compute resources and a secure, stable operating system.

v4.2​

Key features​

Analytics module​

  1. Added the Attribution Analysis model

Helps you explore how much ad placements contribute to payment conversions, so you can measure the return on each placement and optimize ad placement.

  1. Events Analysis supports dimension drill-down

With dimension drill-down, you can move from a high-level overview of the data to a more specific, detailed view to better understand the data.

  1. Flows Analysis supports custom events

View custom events in user behavior paths to better find users' golden paths in your product.

  1. Improved display of dashboard notes

Notes support rich text and resizing, making it easier to annotate the data performance of dashboard reports.

  1. Enhanced asset management for Project Owners

Project Owners and Root users can centrally manage any asset in a project and set editors for dashboards and spaces.

  1. Rebuilt the Third-party Integration module, so you can configure integration methods yourself in the UI

Configure all supported integration plans on the Third-party Integration page and fine-tune parameters in more detail.

  1. Added Tiki Assistant, powered by a large language model

Enter a question to get a detailed answer, and quickly complete command tasks.

Engage module​

  1. Added A/B tests

Configure multiple content split strategies in a single operation task to compare their effects, helping operations staff choose the best strategy scientifically and improve conversions.

  1. Added Campaigns

A campaign is a combination of multiple operation tasks that share the same operation goal. You can batch create operation tasks under the same theme, which simplifies the operators' workflow and makes it easier to evaluate campaign effectiveness.

  1. Added the Engage workspace

Provides global monitoring of the Engage module, helping operations staff keep track of key operation metrics and the operation task schedule in the current project.

  1. App push now supports APNs

Provide your APNs certificate information to connect this channel quickly.

Improvements​

  1. Event details/user details pages can be opened in a new tab
  2. Supports setting a project picture
  3. Improved error detection in the ETL stage of reported data
  4. Changing settings in time zone management now requires confirming the impact
  5. Engage module - Channel content templates support date and time types
  6. Engage module - The former Webhook channel is renamed to custom channel

Integration methods​

  1. Released Logbus 2 version 2.3, which supports new data sources: pulsar and Alibaba Cloud SLS (beta); adds IAM authentication for AWS Kafka; and supports custom data delimiters through plugins
  2. Added an integration method: Seatunnel. Apache Seatunnel is a distributed, high-performance, easily extensible data integration platform for syncing and transforming massive amounts of data (offline & real-time). We have developed AE Writer, which writes data from different platforms to AE clusters.

Engine components and performance improvements​

Trino​

Made targeted performance optimizations for several different query scenarios.

  1. Decoupled the IO and compute thread queues in scheduling, so IO threads can be tuned separately
  2. Optimized file download performance for Alibaba Cloud OSS, AWS S3, Tencent Cloud COS, and Huawei Cloud OBS storage, with improvements of up to 100%
  3. Added a dynamic distinct mode and a Roaring BitMap-based distinct udf to computation, reducing the memory usage of aggregation operators in this scenario by 30%
  4. The Kudu connector adds Kudu data locality scheduling, speeding up Kudu storage scans by 20% and reducing network bandwidth overhead by 50%
  5. The Hive connector optimizes Hive dynamic partition pruning, reducing the amount of data scanned during join operations and speeding up queries

Kudu​

Fixed an issue where kudu load skew overloaded a single node and reduced write and query efficiency, improving the overall stability of kudu queries.

  1. Introduced the leader rebalance feature to distribute read load evenly
  2. Auto rebalancer supports balancing single-replica tables and range-level balancing

Data integration​

  1. Launched a Golang version of the access point component, greatly improving storage efficiency and conflict rates
  2. Added a Seatunnel integration component for more flexible ingestion of real-time or offline data

v4.1​

Key features​

Analytics module​

  1. Improved Metric Alerts

You can set the metric settlement time zone and delayed calculation for individual metric alert tasks.

  1. Behavioral tags support behavior sequences

You can now tag users based on behavior sequence patterns.

  1. Retention Analysis supports holding properties constant

You can now get more accurate retention data.

  1. Enhanced Overview for more accurate aggregated data

In Events Analysis, you can set weighted average as the aggregation method for ratio metrics.

  1. Dashboards and analyses support selecting future dates

You can now build dashboards for campaigns planned in the future.

  1. System Settings supports data usage queries

Understand each project's costs, and use usage trends to spot abnormal growth early and avoid overspending.

  1. Third-party integration supports scheduled pulls and single pulls

You can specify the time range and frequency for pulling data from third-party platforms, and trigger data pulls manually.

Engage module​

  1. Operation tasks support pushing content in multiple language versions

With a single operation task, you can now push different content to users who use different languages.

  1. Operation tasks support pushing by user time zone

With a single operation task, you can now push content according to the local time zones of users in different regions.

Improvements​

  1. Supports saving reports to dashboards from the report page

  2. Added suggested times for scheduled dashboard updates & update status notifications

  3. Added the Contains option to text expressions and the Date range option to time expressions for dynamic parameters

v4.0​

New features​

  1. Behavioral cohorts support behavior sequence conditions

  2. Events Analysis supports multi-date comparison

  3. Supports creating custom dashboard templates

  4. Visualization supports full data

  5. Generate assets for new projects with one click

  6. Batch-edit member permissions

  7. Simplified core dashboard workflows

  8. Upgraded the navigation bar

  9. Supports English/Japanese/Korean versions

Improvements​

  1. Improved the custom property creation process
  2. OpenAPI supports querying metric data
  3. Added detail pages to metadata
  4. Supports editing report names & report descriptions directly
  5. Copying dashboards supports referencing existing reports
  6. Improved display of retention groups
  7. Improved formula interaction

v3.0 - v3.8​

v3.8

  • New features:
  1. Added metrics, which let you maintain unified definitions for common metrics and create analysis reports that include retention data
  2. Added dashboard templates: get analysis ideas from preset themed dashboards and set up metric monitoring with one click
  3. Added data tables, a comprehensive upgrade of dimension tables that supports incremental uploads and associating multiple properties with one table for easier management
  4. Dashboard key data can be customized, with flexible choices of year-over-year, period-over-period, and aggregated data to highlight data characteristics
  5. Added page filters to dashboards. Add commonly used business dimensions to dashboards to give viewers a click-to-drill-down analysis experience.
  6. Dashboard page filters and global filters apply to SQL reports
  7. Added scatter plots to SQL visualization chart types to help explore whether numeric variables are correlated
  • Improvements:
  1. Improved the year-over-year and period-over-period logic for dashboard key data, which now compares based on the same time period
  2. Improved the dashboard sharing experience, making it easier to choose who to share with
  3. Daily reports support extended width mode to show more information horizontally
  4. Daily report PDFs show the content of annotations within the time range
  5. Improved the color scale visuals in retention & visualization charts
  6. Event breakdown supports selecting multiple events/multiple properties
  7. When you create custom events, you can select Any Event, and Any Event supports selecting all event properties
  8. When you create alerts, group-by items support numeric properties
  9. In SQL IDE, you can choose whether the $Part_date dynamic parameter is affected by time zone
  10. Updated the SQL data table component for smoother performance
  11. SQL data tables support sorting by clicking column headers
  12. SQL visualization supports renaming metrics for better chart display
  13. Upgraded the automatic update mechanism for tags/cohorts to get more timely tag/cohort results
  14. Tag data display supports sorting by the horizontal axis and by group, consistent with analysis models
  15. Custom roles based on Analyst can be granted permissions for entities and annotations
  16. Supports creating custom roles based on Project Admin
  17. Supports turning on password strength checks and setting password expiration to improve data security
  • Big data component updates:
  1. Added the Flink base component and the thirdparty_transfer&ods_dump streaming applications. Third-party data can now be streamed into storage and double-written to the ods layer, which helps with later data backtracking
  2. Upgraded the Trino baseline version (380->392), upgraded the runtime JDK to jdk-17.0.4, and added metastore caching, improving performance by 20% in some scenarios
  3. Upgraded Hive (1.2.10->3.1.2)
  4. jdk upgrades: installed openjdk8 (jdk-8.0.332) to replace oracle jdk1.8; upgraded jdk11 to jdk-11.0.15; installed jdk-17.0.4

v3.7

  • New features:
  1. Added Query Statistics, which lets you track query tasks and performance overhead in the system and optimize system resource allocation
  2. Added exchange rates, which integrate daily exchange rate data from Currencylayer so you can convert multiple currencies into one currency for analysis
  3. Added the System Admin and Project Owner roles, which delegate permissions in System Settings and Project Settings for more flexible team management
  4. Supports report-level collaboration: dashboard editors can directly edit and update dashboard reports
  5. Supports Quick Search, which searches dashboards, reports, tags, and cohorts in a project directly so you can locate assets easily
  6. Upgraded the date picker so you can select a time range from a specific date to N days ago
  7. User tags and cohorts can use a custom time zone for calculation
  • Improvements:
  1. Added variance and standard deviation calculation methods for numeric properties
  2. Events Analysis time intervals support quarterly/yearly aggregation
  3. Retention Analysis overview values support showing the total number of users
  4. Events Analysis tables support sorting by the horizontal axis
  5. SQL IDE reports support adjusting parameters directly on dashboards
  6. Line charts and bar charts in analysis models support showing values
  7. The group-by item under the mouse pointer is highlighted on hover
  • Integration methods:
  1. Added the Cocos2d-Lua SDK
  2. Added the C++ client SDK

v3.6

  • New features:
  1. Added daily report push, which sends dashboards to email and common enterprise collaboration tools immediately or on a daily schedule.
  2. Added SQL tags/cohorts, which let you create custom tags and cohorts with SQL statements to meet more personalized tag creation needs.
  3. Added annotations, which let you annotate key dates such as version releases and campaign launches and show them in charts, making it easier to observe metric changes before and after the annotated dates.
  4. Events Analysis supports custom trend chart settings, making it easy to display dual-axis charts, time bar charts, combo charts, and more.
  5. Retention Analysis supports aggregating and comparing retention metrics such as next-day and 3-day retention by week/month, making it easier to observe how data changes over a period.
  6. Funnel Analysis supports holding properties constant. By setting a property to hold constant across funnel steps, you can view conversions in specific scenarios in more detail.
  7. Enhanced drill-down for custom entities: you can view tag information and event details for custom entities such as accounts, devices, and guilds.
  • Improvements:
  1. The Events Analysis overview supports the Total method, which is suitable for calculating deduplicated user counts.
  2. Improved how future dates are displayed in Events Analysis trend charts.
  3. Distribution Analysis shows group data directly when totals are displayed, making charts more efficient to read.
  4. You can edit display names or descriptions directly in the pop-up cards for events and properties.
  5. Funnel Analysis reports on dashboards now match the model, with aligned interactions and display styles.
  6. When you delete a member in system member management, you can hand over their data assets.
  7. Made a series of experience improvements and updated the visual style of the system's empty-state pages.
  • Integration methods:
  1. LogBus 2 is officially released. Compared with the original LogBus, it uses only one-fifth of the memory and runs 5 times faster
  2. The Unity SDK supports home console platforms (Switch, XBox, PS4&5) and mini games (WeChat, Douyin, QQ, and Toutiao mini games)

v3.5

  • New features:

1. Added two data types, object and object group, for reporting and analyzing complex structured data such as battle lineups and deployed troop types. This provides a practical way to analyze nested data.

2. Added comprehensive data permission controls, such as visibility settings for events/user properties and restricting data access with property rules, providing a permission control solution for stakeholders such as co-publishing partners.

3. Upgraded alert management. Supports alerts on group values, formula editing, quick creation from Events Analysis, and alerts based on predicted values, making alert tasks more flexible and granular to meet alerting needs in more scenarios.

4. Cohorts support custom entities. You can create all types of cohorts, including result cohorts, for different entities such as accounts and characters, greatly expanding where cohorts can be used.

5. Upgraded SQL visualization features and made a series of interaction improvements, adding cross table display for easier comparison of grouped data.

6. Upgraded User Behavior Sequence. Supports locating events by property, displaying properties in behavior sequences for quick viewing, and viewing user tags to build a 360-degree profile of a single user.

7. Events Analysis table display on dashboards is now consistent with the model. Overview values can be customized with a separate aggregation method for each metric, and tables support sorting by multiple columns.

  • Improvements:

1. Uniques, times per user, and per-user values in Events Analysis support selecting an entity, so you can count metrics such as "devices that triggered login"

2. Added Extend Date Range to Events Analysis, which lets you calculate metrics such as DAU/WAU and DAU/MAU

3. Events Analysis time intervals support 5-minute and 10-minute granularity

4. You can use Event Time and Event Name directly as filter conditions or group-by items in analysis

5. Retention Analysis table headers can be shown as Day 0, Next Day, 3 Days

6. Formulas can be displayed with 3 decimals or 4 decimals

7. Event/property selectors add pop-up card descriptions to help analysts understand the data

8. Adjusted Team Space permissions so that Space Admins/editors can collaborate on space dashboards

9. When you drill down from a saved report to a list and go back, the previous analysis conditions are kept

10. Supports setting user watermarks to improve data asset security

v3.4

  • New features:

1. Added the data integration module, which provides two major features, Data Collection Guide and Third-party Data Integration, offering guidance and visual configuration for data ingestion

2. Added the Tracking Plan and data validation modules, which let you manage tracking plans in TA and validate data against tracking plans to quickly locate data issues and continuously improve data quality

3. Redesigned the user tag UI and upgraded its features. You can create tags with custom entities and select any entity (formerly Unique User Identification) as the tag subject

4. Upgraded dashboards: you can add tags to dashboards for display, and add notes to dashboards for richer custom information

5. Upgraded Distribution Analysis with the new Also Show option, which calculates any metric for users in different distribution ranges at the same time

6. Unique User Identification is upgraded to entities. You can customize entity names, and after you switch entities in models such as Retention Analysis/Funnel Analysis, you can still drill down to view detailed data

7. Upgraded the TA mobile version: on mobile, you can switch between filter conditions saved as favorites on PC to view data under different conditions

8. Added data receiving URL management, which lets you manage data receiving URLs in TA and periodically checks them for errors

  • Improvements:

1. Supports moving dashboard folders to Team Space

2. Supports sharing dashboards with members whose accounts are not activated

3. Analysis supports common percentile metrics

4. Events Analysis supports event breakdown for Any Event

5. Improved formula editing interactions. Filtered metrics can be carried into formulas

6. The SQL IDE page shows the corresponding report name

7. User lists and event details support forced refresh

8. Improved interactions on the dimension table upload page

9. Improved error message display in Tracking

10. Improved the Project Settings page with a separate analysis configuration page

11. Added account freezing to System Settings

12. Supports binding multiple third-party applications

13. Template types support tags and cohorts

v3.3

  • New features:

1. Added the visualization module. This version supports visual configuration of SQL IDE results and displaying them in dashboards. Later versions will gradually support visual configuration for each analysis model

2. Added Interval Analysis. The Interval Analysis model analyzes the time interval between two user actions and shows how the intervals are distributed. It is suitable for analyzing interval-related metrics such as conversion time, action duration, and dwell time.

3. Added historical versions to user tags. A historical version records a tag's values on a specific past day. Historical versions can be generated manually or backed up automatically when the tag is updated. History tags can be used directly in models, either by selecting the historical version for a specific day or by automatically matching the historical version for each analysis date. Historical version data of tags can also be used in custom properties.

4. Dashboard filters can be saved long term. Users can save frequently used dashboard filters as favorites, and dashboard owners and collaborators can set their personal saved filters as the dashboard's default filter, which is applied every time a user opens the dashboard

  • Improvements:

1. The summary column of Events Analysis tables supports switching the summary value between Sum and Average

2. The task list shows the size of download files

3. Events Analysis tables remember the selected layout (switching between By Time, Events, and Groups)

4. The project switcher supports pinning projects, which applies only to you

5. Added search to the Tracking → Real-time saved data module, so you can search data saved in real time

6. Dashboards support batch downloading the data of all their reports

7. Added greater than or equal to and less than or equal to to the filter logic for numeric properties

8. When you compare groups in Composition Analysis, you can set a global filter

9. Dashboard names are now shown in full in the dashboard module

10. Custom properties of the time type are precise to the second when debugging

11. When a cohort is used as a group-by item, the chart shows the full cohort name. Previously, in cohort and not in cohort were shown. Now in and not in are followed by the name of the cohort. Note that API results change accordingly

  • Template Gallery:

1. Added Interval Analysis templates to template types

v3.2.1

  • New features:

1. Added the message center. TA users can customize message channels to receive System Notification, Project Dynamic, Dashboard Dynamic, Task Warning, and other messages from TA.

2. Launched the beta of Template Gallery. TA users can use Template Gallery to learn what each feature module actually does and quickly apply templates in their projects.

3. Added Virtual MFA, which greatly improves account security. You can set up MFA for your personal account, or system admins can require all users to enable MFA

4. Added the Grafana-based O&M Data Monitoring Console. System admins can go to System Settings → Host Node Monitoring in the upper-right menu and click Monitor in Grafana in the upper-right corner of the page to open it

5. Added self-service version upgrades. System admins can go to System Settings → System Access Details in the upper-right menu to view the current cluster version and whether a new version is available, and schedule an update time to upgrade the TA system on their own

  • Improvements:

1. Improved Events Analysis chart display and added sorting settings for chart groups

2. Filter conditions support parallel filters, enabling two-level nested filters such as "(A or B) and (C or D)"

3. Added a dashboard report overview module, where you can resize reports, remove reports, or drag reports to reposition them

4. Added the Selector dynamic parameter type to SQL IDE, which lets you select multiple SQL statement fragments from a drop-down list and is suitable for dynamically adjusting complex filter conditions. Removed the restriction that dashboard caching requires a calculation time of more than 120 seconds, so all reports can now use dashboard caching. Supports With clauses

5. The maximum number of rows for data downloads can be adjusted through parameters (contact ThinkingAI staff)

6. Cohorts, user tags, and custom properties support Save as

7. Event and property selectors support pinyin search, so you can search by full pinyin or pinyin initials

8. Improved analysis model interactions with a clearer hierarchy of query settings for greater efficiency

9. Improved the experience of long-running tasks such as downloads. Task progress is shown in a floating widget, and long-running tasks can be moved to the background. When a task finishes, you can view its result or download the data from Task List in the title bar

  • Changes:

1. Renamed labels: Cohort no., Cohort name, Tag no., and Tag name are renamed to Cohort name, Display Name, Tag name, and Display Name respectively in the new version

  • Integration methods:

1. Added the Cocos2d-x SDK

v3.1

  • New features:

1. Added Team Space. Project members can use Team Space to standardize dashboard sharing and unify the folder structure and order of dashboards, making communication and collaboration more efficient and convenient.

2. Added the dashboard Editor permission to dashboard sharing. Dashboard owners can assign editors to edit dashboard reports together

3. Added third-party platform apps. You can bind TA to WeCom, DingTalk, or Feishu for features such as third-party QR code login, one-click login from the app workbench, and receiving TA updates.

4. Added the TA mobile version. Access the TA console from a mobile device such as a phone, or enter the TA console through a third-party platform app, to use the mobile version

  • Improvements:

1. Added more dynamic parameter types to SQL IDE. Parsing a statement no longer replaces everything in the query box. It appends to the end instead

2. Tags can be used in the creation expressions of custom properties

3. When you change the event of a metric, the system checks whether the filter conditions still apply and keeps the applicable ones

  • Integration methods:

1. The Unity SDK supports PC games (Windows, macOS), supports reporting data in UnityEditor, adds multi-instance support, and supports setting the default instance in code

2. Added the status command to LogBus to view the current upload speed and status

v3.0

  • New features:

1. SQL IDE is now native: SQL can be saved as reports and added to dashboards. Dynamic parameter types are expanded and can be modified when querying in the Explore module

  • Improvements:

1. Rebuilt the model UI with a brand-new interaction design, improved table performance, and a better analysis experience

2. Added the By sets option to the grouping method for list properties. When it is selected, list properties are deduplicated as sets, and grouping calculations are not affected by element order. Added Deduplication of set to the calculation logic, which calculates the deduplicated count after a list is processed as a set

3. The association between custom event properties and events can be customized. In addition to the original logic, added the Associate with specified events and Associate with all events options

4. Reports in dashboards can be dragged to reposition them, and the height of large reports can be adjusted for viewing large amounts of data

5. Added full data download to the retention and funnel models

6. In Composition Analysis, tag values can be used as analysis properties in calculations

  • Integration methods:

1. Added the Lua SDK

2. Major update to the mini game SDK: added support for Bilibili mini games, Xiaomi mini games, and H5 games; supports ts projects; the three game engine SDKs also support common mini game and quick game platforms (Huawei, Xiaomi, OPPO, vivo) and H5 games

v2.0 - v2.8​

v2.8

  • New features:

1. Added two tag creation methods: First or last and Metric value. First or last uses a property from the first/last time a user performed an action within a period as the tag value, and Metric value uses the user's metric value within a period as the tag value

2. Added the Explore module, which supports custom drill-down analysis of dashboard reports

3. Greatly improved the secondary development tool. The former data deletion tool is merged into it, and metadata editing (deleting properties or modifying property types) and data deduplication are added

  • Improvements:

1. Rebuilt the dashboard UI for clearer data and easier use

2. Added pie distribution charts to Events Analysis

3. Event search supports searching by original event name

v2.7

  • New features:

1. Added user tags. Tags divide users into multiple tiers that correspond to multiple tag values. Currently, user tags can be set by metric conditions or by uploading ID lists.

2. Added Tracking, which consists of three parts: the Event Tracking module, which shows recent data receiving status; the Real-time Data module, which shows data saved in real time and samples of error data; and the Debug Data module, which works with the Debug mode of client SDKs.

3. Added event breakdown to the Events Analysis model. You can break down an event by a property and, in a formula, divide each broken-down part by the unbroken event to calculate each part's share of the total.

  • Improvements:

1. Updated the style of the event list drop-down and added pinning of frequently used events

2. Ranges for numeric properties can be set in batches

3. When calculating by week, you can customize the first day of the week

4. The time zone feature lets you select multiple project time zones

5. Added formula editing to the Also Show feature of the retention model

  • Integration methods:

1. Added the Unreal Engine SDK

2. Added data transfer tools: Filebeat + Logstash

3. Added a Windows version of Logbus

v2.6

  • New features:

1. Added the permission management module, which introduces role management, data permission groups, and member groups, so you can set system feature permissions and data viewing permissions for members in a project

2. Added the time zone management module, which lets you offset data in a project to a specified time zone for calculation and viewing

3. Added creating cohorts by uploading ID lists

4. Added the system member management page

5. Added event locating to User Behavior Sequence

  • Improvements:

1. Added viewing and exporting model SQL and model API settings in analysis models

2. Added the Median calculation keyword for numeric properties

3. Distribution Analysis supports formula editing

  • Integration methods:

1. Added the Flutter SDK

2. Added SDKs for the Egret Engine, LAYABOX engine, and CocosCreator engine to the mini game SDK

v2.5

  • New features:

1. Added the List data type, which covers scenarios such as user tags, A/B test experiment groups, and item inventories. Later models provide analysis features suited to lists

2. Added dashboard copying, import, and export

3. In metadata management, you can batch upload and download display name configurations to modify display names

4. Added viewing and downloading event details

5. Added OpenAPI for analysis models, cohorts, and dimension tables.

6. Added report management

  • Improvements:

1. Improved the funnel model window, which now supports hour- and minute-level windows.

2. Events Analysis supports date range comparison.

3. Added the Until Today option to the time picker.

v2.4

  • New features:

1. Added multiple user identification systems, so you can use custom fields other than the default ID as the user association field in Retention Analysis, Funnel Analysis, and Distribution Analysis.

2. Greatly improved the retention model.

  • Improvements:

1. Analysis objects and group-by items in analysis models can be reordered by dragging.

2. Added an events list to the event properties management page in metadata, which shows the events a property belongs to.

3. Improved Events Analysis table display to better fit real business scenarios.

  • Integration methods:

1. Added SDKs for OPPO quick games and vivo quick games.

2. Added the Node.js SDK.

v2.3

  • New features:

1. Added custom events, which combine multiple events and filter conditions into a custom event that can be used as an event in analysis models

2. Added alert management, which monitors and pushes abnormal data results based on the alert rules you configure.

3. Added folders to the dashboard list, so dashboards of the same type can be grouped in one folder.

4. Added third-party login, which currently supports DingTalk and WeCom.

  • Improvements:

1. Added global filters to Events Analysis.

2. Added copying of analysis objects and their filter conditions in Events Analysis and Funnel Analysis.

3. Custom event properties support associating user properties.

  • Integration methods:

1. Added the user property reset API user_unset, which clears the values of one or more user properties.

2. Added the preset property #zone_offset, which records the offset in hours of the event time from utc.

v2.2

  • New features:

1. Added dimension table properties. You can upload dimension tables in csv format to map content such as id values and enum values in existing properties, for example, mapping product and item ids to Chinese names.

2. Added custom properties, which generate new properties from sql expressions based on existing properties for use in analysis models.

3. Added global dashboard filters. You can filter by time and property in a dashboard, and the filter conditions apply to all reports in the current dashboard.

4. Added approximate calculation. Turning it on greatly improves calculation performance, while the standard error of the results is kept within 0.4%.

  • Improvements:

1. Increased the maximum number of group-by fields to 10 and the maximum number of funnel model steps to 30.

2. The Also Show feature of the Retention Analysis model supports adding filter conditions to the events shown alongside.

  • Integration methods:

1. Added mini program and mini game SDKs to support more mini program and mini game platforms. Currently supported: WeChat mini programs, Alipay mini programs, ByteDance mini programs, Baidu mini programs, quick apps, WeChat mini games, ByteDance mini games, and Baidu mini games.

2. Added server-side integration methods: C SDK and Golang SDK.

v2.1

  • New features:

1. Added result cohorts. When the result of an analysis model is a user count, you can create a cohort from the users in the result set, which is called a result cohort. Result cohorts can be used as filter conditions or group-by items in all analysis models

2. Cohorts are renamed to behavioral cohorts, and filter keywords are added. You can create cohorts based on conditions such as the total number of days a user did something or a metric of something they did (for example, total top-up amount > 500)

3. TA cluster deployment and updates support offline mode

  • Improvements:

1. Added cumulative trend charts to Events Analysis

2. Added client SDK reporting policy settings to project management, so you can adjust the reporting time interval or record count interval as needed

3. Added adjustment of the number of rows shown in analysis model tables, with 10, 20, 50, or 100 rows per page

4. In the User Behavior Sequence module, you can view user behavior sequences by hour and in reverse order, and refresh the data

  • Integration methods:

1. Added multi-APPID instances to the JS SDK, along with super property and dynamic super property APIs.

2. Added the C# server-side SDK

3. Refactored the Android and iOS SDKs and released version 2.0

v2.0

  • New features:

1. Added user property grouping, Average, and Also Show to Retention Analysis

2. Added copying of filter conditions

3. The data deletion tool supports custom filter conditions for deleting matching data

  • Improvements:

1. Rebuilt the web pages and improved global interactions, including search, sorting, and related interaction logic in each module

2. Optimized the front-end/back-end interaction logic for data queries and added a data loading timer and progress display, so you can see the current query progress directly on the page. When a query is interrupted, the system proactively reclaims related resources to improve the overall performance of the cluster

  • Integration methods:

1. Added multi-APP_ID instances to the Android and iOS SDKs, and added an APP crash event to auto-tracked events

2. Added the Unity SDK

v1.0 - v1.11​

v1.11

  • New features:

1. Added scheduled dashboard caching. Dashboards that don't need real-time data can be updated and cached on a schedule, which greatly speeds up later viewing.

2. Added an ip whitelist for server-side data ingestion, which you can use with your project's security policy to control the ips that send data.

  • Improvements:

1. Added bookmarks to the SQL custom query module. SQL bookmarks are saved to your account.

  • Integration methods:

1. Greatly improved LogBus performance and stability. Users of older versions should update to the new version

2. WeChat Mini Program supports auto-tracking

3. Improved the performance and stability of the Java SDK | Python SDK

v1.10

  • New features:

1. Added the Distribution Analysis model, which analyzes the distribution of users' action frequency or event metrics over a period

  • Improvements:

1. Improved dynamic time selection in cohort definitions, extending it from the last N days to any time window

2. In the Events Analysis model, non-formula analysis objects can be renamed

3. The username length limit of the TA system is increased from 12 to 50 characters

  • Integration methods:

1. Added the Alipay mini program SDK

2. LogBus supports the HTTPS protocol, monitoring by directory, and sending data from a Kafka data source to multiple projects

v1.9

  • New features:

1. Added default groups. In the group control of a chart, select the default group option in the control's drop-down list. The next time you open the report, the top N groups that meet the criteria are selected by default (the number is customizable)

2. All analysis models show the system fields Account ID and Distinct ID in user properties, and you can filter or group users by these two fields

  • Improvements:

1. Improved chart and dashboard display logic in Events Analysis when there are multiple analysis objects

2. Improved the UI and interactions of metadata management and the user console

3. Improved automated O&M components, fixing occasional false error alarms caused by network jitter

4. Optimized user table storage and queries, including user search, user lists, and user data ETL, greatly improving performance

v1.8

  • New features:

1. Added a user search entry, which lets you filter by user properties and go directly to the user list or user behavior sequence

  • Improvements:

1. Anonymous ID is renamed to Distinct ID throughout the system

2. Added password reset by Root users to the permission system

3. Updated IP database parsing, greatly improving the parsing accuracy of IPs outside China

4. Major UI improvements

v1.7

  • Improvements:

1. Added same-day retention analysis to the Retention Analysis model, which measures retention when users return on the same day as the initial action

2. In the Retention Analysis model, the calculation period for daily retention and churn is extended to 180 days

3. Added the Compare to current date and Compare to event time operators for time properties

4. In the Events Analysis model, the precision of the time control is improved to the hour

5. In the Events Analysis model, tables support transposing rows and columns and quickly searching groups

  • Integration methods:

1. Added the ability to set the event trigger time to the event reporting APIs of the Android and iOS SDKs

v1.6

  • New features:

1. Added the user behavior path analysis model, an exploratory model for analyzing action order, action preferences, key nodes, and conversion efficiency

  • Improvements:

1. The dashboard list supports drag and drop, so you can customize the order of dashboards

2. When editing formulas in the Events Analysis model, you can set filter conditions for individual objects in the formula

3. Optimized and updated ETL components, greatly improving data processing and throughput

4. Upgraded the query engine kernel version

  • Integration methods:

1. Added Android and iOS support for React Native

2. Added Android and iOS support for Weex

3. Added a server-side SDK: PHP SDK

4. Updated the Java SDK, Python SDK, and iOS SDK

v1.5

  • New features:

1. Added the user property analysis model, which analyzes and compares user properties through grouping, filtering, and more

2. Added a switch for the historical data import channel. When it is turned on, large volumes of historical data can be imported more efficiently

3. Added a data ingestion monitoring console where you can view data import status and error data samples

  • Improvements:

1. Added the Minute and Total options to time grouping in Events Analysis

2. Improved dashboard performance when multiple reports load large amounts of data at the same time

3. Optimized the ETL logic for importing user properties, greatly improving the efficiency of user table imports

  • Integration methods:

1. Major LogBus update: supports multiple instances, multithread safety, selectable data compression methods, and performance optimizations

v1.4

  • New features:

1. Added event grouping. You can group events and configure their display as needed, so a large number of event types no longer makes events hard to manage

  • Improvements:

1. Improved the time control to make selecting relative time easier

2. Increased the maximum number of event properties and user properties that can be received, meeting the needs of more data dimensions

3. Added monitoring of IDE tools and the NTP service to the automated O&M system

4. Added support for inconsistent server time zones

  • Integration methods:

1. Added the WeChat Mini Program SDK

2. Added the WeChat mini game SDK

v1.3

  • New features:

1. Added dashboard configuration and a small dashboard style

2. Added auto-tracking to the iOS and Android SDKs. For actions such as app start and exit, developers can complete tracking quickly with auto-tracking

3. Added a data deletion tool for on-premises environments, which lets you delete ingested data yourself

4. Added data refresh to analysis models and dashboards, which bypasses the back-end cache to refresh real-time data immediately

  • Improvements:

1. User List and User Behavior Sequence remember the selected event properties and user properties

2. Improved automated O&M tool features

3. Improved the read/write performance of the underlying storage and query engine

  • Integration methods:

1. Added the Restful Api, which supports transmitting data with the HTTP POST method

v1.2

  • New features:

1. Added User Behavior Sequence, which shows the behavior sequence and behavior distribution of a single user in the analysis results

2. Added a User List link to analysis models. When the analysis result is a user count, you can go directly to the list of matching users

  • Improvements:

1. Cohorts can classify users who have not performed any actions

2. Added calculation of users who drop off between steps to the Funnel Analysis model

v1.1

  • New features:

1. Added Funnel Analysis, which analyzes conversion and drop-off at each step across multiple key steps to optimize the conversion efficiency of each step

2. Added Cohorts, which classify user groups based on user behavior characteristics and user properties, and let you analyze specific user groups in analysis models

3. Launched the automated O&M system. On-premises users don't need to maintain clusters themselves, and the system fixes issues on its own unless they are at the server hardware level

  • Improvements:

1. Fixed known bugs and made some UI improvements

2. Greatly optimized the Retention Analysis algorithm, improving calculation performance by more than 100% under the same query conditions

v1.0

  • New features:

1. Added Events Analysis, which filters, groups, and aggregates user behavior events to analyze trends in user behavior over a specified period and compare dimensions.

2. Added Retention Analysis, which analyzes whether users come back after completing an action. For example, it shows how many users who logged in log in again over the following seven days, that is, the 7-day retention rate of active users.

3. Added Manage Reports, which lets you save frequently used analysis models as reports for later queries and add them to dashboards.

4. Added Manage Dashboards, which shows multiple reports at once so you can query and compare data dimensions of related scenarios together, and share them with other users.

5. Added Metadata, which lets users track, view, manage, and edit ingested data

6. Added the Permission module

7. Added six data ingestion tools

a. Client SDKs: Android, iOS, JavaScript

b. Server-side SDKs: Java, Python

c. Server-side log collection tool: LogBus

Was this page helpful?
On this page