GITNUXSOFTWARE ADVICE
Data Science AnalyticsTop 10 Best Cd Library Software of 2026
Top 10 Cd Library Software ranked with technical criteria, including BigQuery, Azure Synapse Analytics, and Databricks for data teams.
How we ranked these tools
Core product claims cross-referenced against official documentation, changelogs, and independent technical reviews.
Analyzed video reviews and hundreds of written evaluations to capture real-world user experiences with each tool.
AI persona simulations modeled how different user types would experience each tool across common use cases and workflows.
Final rankings reviewed and approved by our editorial team with authority to override AI-generated scores based on domain expertise.
Score: Features 40% · Ease 30% · Value 30%
Gitnux may earn a commission through links on this page — this does not influence rankings. Editorial policy
Editor’s top 3 picks
Three quick recommendations before you dive into the full comparison below — each one leads on a different dimension.
Google BigQuery
BigQuery ML for training and predicting models using SQL within BigQuery
Built for teams building CD library analytics and experiment reporting on large datasets.
Azure Synapse Analytics
Editor pickBuilt-in pipeline orchestration for deploying and running integrated data workflows
Built for enterprises standardizing analytics delivery with reusable pipelines and notebooks.
Databricks
Editor pickWorkflows with job orchestration tied to Git-based notebook and artifact deployment
Built for teams delivering governed data pipelines with Git-based promotion and automated jobs.
Related reading
Comparison Table
This comparison table ranks the top CD library software options by integration depth, data model constraints, and how automation and the API surface support repeatable schema and provisioning workflows. It also contrasts admin and governance controls such as RBAC coverage, audit log visibility, and configuration and extensibility patterns across tools including BigQuery, Azure Synapse Analytics, and Databricks. Each row highlights concrete tradeoffs in schema management, throughput considerations, and operational boundaries for CD library usage.
Google BigQuery
cloud sql analyticsRuns fast SQL analytics on large datasets and integrates with ML and data visualization for scalable data science workloads.
BigQuery ML for training and predicting models using SQL within BigQuery
Google BigQuery stands out for serverless, SQL-first analytics that runs across massive datasets without managing infrastructure. It supports columnar storage, fast analytic queries, and data integration from common Google Cloud services plus external sources.
Its ML and geospatial functions help deliver analytics and modeling directly inside the warehouse. For a CD library software use case, it also acts as a reliable analytics backbone for dashboards, experiments, and content performance reporting.
- +Serverless SQL engine handles high-volume analytics without cluster management
- +Columnar storage and partitioning options improve scan efficiency for large datasets
- +Built-in BI integration and export formats support dashboard and reporting workflows
- +Data governance controls and auditability support enterprise compliance needs
- +In-warehouse machine learning functions speed analytics-to-model pipelines
- –Cost sensitivity requires careful query optimization and partitioning discipline
- –Complex modeling and large workflows can require nontrivial schema planning
- –Interactive debugging across multi-job pipelines is harder than in purpose-built apps
CD librarians and metadata stewards
Normalize and enrich catalog metadata at scale
Higher metadata accuracy and consistency
Content operations and publishing analytics
Measure enrichment impact on publication performance
Clear enrichment ROI by asset type
Show 2 more scenarios
Data engineers for CD platforms
Ingest enrichment sources into BigQuery pipelines
Faster enrichment data preparation
Batch and streaming ingestion loads enrichment feeds from cloud storage and external services for joins.
Experimentation and personalization analysts
Run feature experiments using enrichment fields
Improved targeting quality from data
Built-in ML and feature engineering support training and scoring with enriched attributes.
Best for: Teams building CD library analytics and experiment reporting on large datasets
More related reading
Azure Synapse Analytics
data warehouseCombines data warehousing, big data analytics, and orchestration to analyze and transform data for data science projects.
Built-in pipeline orchestration for deploying and running integrated data workflows
Azure Synapse Analytics distinguishes itself by combining large-scale data integration, streaming ingest, and enterprise analytics in a single workspace. It supports Spark-based processing, SQL-based querying, and serverless patterns to run workloads across structured and semi-structured data.
Built-in pipelines and data movement features help connect sources into analytics stores with repeatable orchestration. For CD library use, it can serve as a central analytics deployment target where reusable notebooks and pipeline definitions act as standardized components.
- +Unified workspace for pipelines, SQL, and Spark development
- +Native integration of orchestration with notebooks and SQL scripts
- +Scalable analytics compute options for batch and streaming workloads
- +Strong connectivity to Azure data services for repeatable deployments
- –Deployment versioning for artifacts requires disciplined release practices
- –Managing workspace sprawl and environment drift can be time-consuming
- –Advanced performance tuning for Spark and SQL increases operational effort
Data engineering teams
Orchestrate batch and streaming ingest pipelines
Faster production data onboarding
Analytics engineers
Standardize SQL and Spark workload components
Repeatable analytics releases
Show 1 more scenario
Platform and DevOps teams
Centralize deployment for workspace assets
Controlled environment updates
Synapse workspace resources group linked services, pipelines, and notebooks as CD-managed libraries.
Best for: Enterprises standardizing analytics delivery with reusable pipelines and notebooks
Databricks
lakehouseOffers a unified data platform with notebooks, collaborative workspaces, and scalable processing for analytics and ML.
Workflows with job orchestration tied to Git-based notebook and artifact deployment
Databricks stands out with a unified lakehouse that combines data engineering, data science, and analytics under one execution engine. For a CD library software workflow, it supports versioned notebooks, Git-backed repositories, job orchestration, and reusable pipeline patterns.
It also integrates with CI triggers so code changes can launch parameterized data pipelines and validation steps. Strong governance features like Unity Catalog help apply consistent access controls across datasets and pipeline artifacts.
- +Unified lakehouse execution for pipelines, notebooks, and analytics artifacts
- +Git-integrated notebooks and workspace repositories enable repeatable CD promotions
- +Job orchestration supports scheduled and event-driven pipeline runs
- +Unity Catalog centralizes permissions for datasets across environments
- –CD workflows can require substantial platform setup and operational knowledge
- –Debugging distributed jobs often needs deeper knowledge of Spark execution
Data engineering teams
Deploy validated pipelines from Git changes
Fewer broken releases
MLOps and data scientists
Version models and training datasets together
Reproducible model iterations
Show 2 more scenarios
Platform and governance owners
Enforce access controls across artifacts
Controlled data sharing
Governance teams apply Unity Catalog permissions across tables, notebooks, and pipeline outputs consistently.
Analytics engineering teams
Schedule CDC or batch rebuilds reliably
More stable data refreshes
Analytics teams orchestrate repeatable ingestion and transformation jobs using reusable workflow patterns.
Best for: Teams delivering governed data pipelines with Git-based promotion and automated jobs
More related reading
Snowflake
cloud data platformProvides a cloud data platform that supports analytics workloads with SQL, data sharing, and native integrations.
Time Travel for querying and restoring prior data versions
Snowflake stands out for separating compute from storage and scaling workloads through virtual warehouses. It supports data warehousing and analytics by ingesting data from many sources, organizing it into structured schemas, and sharing data across teams with governed permissions. For CD library use, it can serve as a governed repository for versioned datasets, release artifacts, and metadata used by CI and deployment pipelines.
- +Virtual warehouses scale compute for parallel build and test workloads
- +Strong governance with role-based access control and secure data sharing
- +Flexible ingestion from multiple sources supports automated release pipelines
- +Time travel enables rollback and audit-friendly dataset versioning
- +Native support for semi-structured data reduces ETL overhead
- –Modeling choices for performance can require tuning and expertise
- –Large-scale CD workflows need careful warehouse and cost management
- –Artifact storage is not as purpose-built as dedicated DevOps tools
Best for: Data-driven CD libraries needing governed datasets and scalable analytics
dbt
analytics engineeringTransforms data in warehouses using SQL-based version control and dependency-managed models for analytics engineering.
dbt packages for sharing models, macros, and tests across projects as a library
dbt stands out for turning data transformation into a version-controlled, test-driven workflow centered on reusable packages and models. It supports documentation generation from code, lineage views, and automated data quality checks tied to the transformation logic. Core capabilities include SQL-based modeling, incremental builds, macros, and environment-aware deployments for repeatable library-style analytics assets.
- +Code-first modeling builds a reusable transformation library with consistent patterns
- +Automated tests and documentation generation reduce drift between logic and published knowledge
- +Macro and package reuse accelerates standardization across teams and projects
- –Learning curve remains steep for modular modeling, macros, and deployment concepts
- –Usability depends heavily on warehouse setup and CI wiring for reliable library governance
- –Complex lineage and large projects can make debugging slow without strong conventions
Best for: Analytics and data engineering teams building reusable transformation libraries with governance
Apache Airflow
pipeline orchestrationOrchestrates data pipelines with scheduled and event-driven workflows for analytics ETL and ELT processes.
DAG-based orchestration with configurable retries, backfills, and dependency tracking in the scheduler
Apache Airflow stands out by turning data and application tasks into scheduled, dependency-aware workflows with a code-first model. It provides DAG definitions, task orchestration, and robust retry and backoff behavior across batch and streaming-adjacent pipelines.
Operators cover common integrations while the web UI and REST API expose run state, logs, and dependency status. It also supports scalable execution through separate schedulers and workers using common backends like Celery and Kubernetes.
- +Code-defined DAGs give repeatable, reviewable workflow logic
- +Rich operator ecosystem covers common data and systems integrations
- +Web UI shows run state, retries, and task logs for debugging
- +Strong scheduling and dependency management reduce manual orchestration work
- –DAG design and scheduler configuration add operational complexity
- –Large DAG sets can increase UI and scheduler load without tuning
- –Debugging distributed execution often requires multi-service observability
Best for: Teams building complex data pipelines needing dependency-aware scheduling and visibility
More related reading
Apache Superset
bi dashboardsCreates interactive dashboards and ad hoc analytics from connected data sources using SQL and semantic modeling.
Row-level security tied to authenticated identities for governed dashboard access
Apache Superset stands out with its broad support for interactive dashboards built from multiple database connections and SQL-based datasets. It provides chart building, dashboard layouts, and scheduled refresh so teams can publish reporting artifacts from governed data sources.
Built-in features like semantic layers, row-level security, and alerting help operationalize analytics across environments. Its strongest fit appears in analytics delivery workflows rather than document-centric library management.
- +Rich dashboard and visualization builder with flexible chart configuration
- +SQL lab and dataset abstraction streamline reusable metrics across dashboards
- +Role-based access and row-level security support controlled analytics publishing
- +Strong integration with common databases through SQLAlchemy and drivers
- –Advanced security and governance setup takes time and careful configuration
- –Semantic layer modeling adds complexity for teams wanting quick starts
- –Performance tuning for large datasets often requires database-side optimization
- –Complex cross-filtering can be harder to maintain as dashboards grow
Best for: Teams building governed analytics dashboards and reusable reporting metrics
RStudio
data science ideSupports data science workflows with an integrated environment for R that includes workspaces, versioned projects, and collaboration options.
R Markdown and Quarto publishing directly from package and project sources
RStudio stands out with tight, workflow-first support for R, including interactive consoles, script editing, and project organization. For CD library work, it provides reproducible documentation via R Markdown and notebook execution via Quarto.
It also integrates with Git for versioned code and supports automated builds through command-line rendering. The ecosystem includes extensive R package tooling for managing and testing library releases.
- +R-centric workflow makes library authoring and testing faster
- +Integrated R Markdown and Quarto pipelines for documentation outputs
- +Project and Git integration support consistent release practices
- –CD-style deployment automation depends on external tools beyond RStudio
- –Multi-language library builds are not as smooth as R-focused stacks
- –Complex CI orchestration often requires manual configuration
Best for: R teams needing reproducible libraries with documentation and Git workflows
More related reading
Kaggle Datasets
dataset catalogHosts curated datasets for analytics and data science and supports notebook-based exploration and collaboration.
Dataset version history with notebook references for reproducible dataset exploration
Kaggle Datasets distinguishes itself with a huge, community-curated index of downloadable datasets and strong metadata around data provenance. The site supports dataset browsing, versioning, and dataset pages that link to notebooks for reproducible exploration workflows.
It also enables dataset downloads through Kaggle tooling, making it practical for quickly bootstrapping data for CD library ingestion. Dataset licensing and update cadence vary widely across contributors, which can complicate governance for strict release pipelines.
- +Large searchable catalog with detailed dataset metadata and schema notes
- +Dataset pages connect directly to notebooks that show data preprocessing steps
- +Versioned dataset entries help track changes across dataset releases
- –Data quality and licensing clarity vary across community submissions
- –Operational CD-style governance needs extra layers for auditing and validation
Best for: Teams needing fast dataset discovery and reference notebooks for CI workflows
Observable
data visualizationBuilds and publishes interactive data visualizations and analysis notebooks for exploratory analytics.
Reactive cells that rerun automatically when upstream variables change
Observable turns JavaScript notebooks into shareable, interactive data experiences with reactive cells. It provides notebook-based building blocks, built-in visualization libraries, and publishable documents that function as a lightweight app surface.
For a configuration-management and delivery library workflow, it supports embedding versioned code examples, parameterized experiments, and reusable visualization patterns. The platform also has limits for heavyweight library packaging and long-running backend execution, since projects primarily run in the browser and focus on interactive exploration.
- +Reactive notebook execution updates visuals instantly from dependent code cells
- +Publishable notebooks provide a reusable documentation surface for code-driven examples
- +Tight JavaScript integration supports custom components and visualization logic
- +Forkable, shareable documents simplify iteration of interactive library patterns
- –Browser-first execution limits workflows needing persistent backend services
- –Library packaging and distribution are weaker than package-manager-first approaches
- –Large notebooks can become harder to maintain as reuse patterns proliferate
- –Collaboration and review workflows are less structured than full software repos
Best for: Data teams sharing interactive code libraries and examples in browser-run notebooks
Conclusion
After evaluating 10 data science analytics, Google BigQuery stands out as our overall top pick — it scored highest across our combined criteria of features, ease of use, and value, which is why it sits at #1 in the rankings above.
Use the comparison table and detailed reviews above to validate the fit against your own requirements before committing to a tool.
How to Choose the Right Cd Library Software
This buyer's guide covers Google BigQuery, Azure Synapse Analytics, Databricks, Snowflake, dbt, Apache Airflow, Apache Superset, RStudio, Kaggle Datasets, and Observable as CD library software candidates.
Each tool is evaluated through integration depth, data model fit, automation and API surface, and admin governance controls so selections map to how CD libraries get built, versioned, promoted, and audited.
The ranking uses practical fit to content delivery pipelines, analytics backbones, and code-library publishing patterns that appear in these tools’ workflows.
CD library delivery tooling built around versioned data, code artifacts, and controlled releases
CD library software organizes reusable content assets like datasets, transformation logic, pipeline definitions, and interactive examples into repeatable delivery flows with clear provenance.
It reduces drift by enforcing a structured data model and release mechanics that can run scheduled or event-driven jobs with visible run state and rollback paths.
Teams typically use these capabilities to power analytics reporting, transformation libraries, governed dataset releases, and interactive documentation surfaces. For example, BigQuery supports in-warehouse analytics plus BigQuery ML, while dbt turns SQL models and tests into a version-controlled transformation library.
Integration depth, data model governance, and automation surface for repeatable releases
A CD library tool must connect code, data, and execution so releases are reproducible across environments.
The decisive factors are how well each platform exposes automation via pipeline orchestration and APIs, how the platform models versioning and rollback, and how governance controls like RBAC and auditability attach to datasets and artifacts.
Google BigQuery, Azure Synapse Analytics, and Databricks lead when the library needs both execution orchestration and data governance in one platform.
In-warehouse analytics and BigQuery ML for model-ready release pipelines
Google BigQuery provides BigQuery ML training and predicting using SQL inside BigQuery, which keeps content performance reporting and modeling close to the released datasets. This matters when the CD library includes experiment logic that must publish results directly from the same data model.
Built-in pipeline orchestration tied to notebooks, SQL scripts, and scheduled runs
Azure Synapse Analytics offers built-in pipeline orchestration in a unified workspace with notebooks and SQL scripts. Databricks adds job orchestration tied to Git-backed notebook and artifact promotion, which reduces manual handoffs when releasing library updates.
Versioning mechanics for datasets and time-based rollback
Snowflake’s Time Travel allows querying and restoring prior data versions, which is a concrete rollback mechanism for released datasets. This matters when CD libraries publish curated datasets and the library needs rollback without rebuilding upstream pipelines.
SQL-first transformation library with dependency-aware models, tests, and documentation
dbt turns transformation logic into version-controlled models with automated tests and documentation generation from code. It supports incremental builds, macros, and environment-aware deployments so governance is attached to the transformation library rather than only the runtime.
DAG-based dependency-aware orchestration with a REST API for run state and logs
Apache Airflow defines workflows as code-defined DAGs and exposes run state, logs, and dependency status through its web UI and REST API. This matters when the CD library needs explicit backfills, retries, and dependency tracking across many library components.
Admin governance controls like RBAC and dataset access enforcement
Snowflake provides role-based access control and secure data sharing for governed datasets, and Apache Superset adds row-level security tied to authenticated identities. Databricks adds Unity Catalog to centralize permissions across datasets and pipeline artifacts, which supports governance across environments.
Interactive library publishing patterns using notebook execution semantics
Observable supports reactive cells that rerun automatically when upstream variables change, which is a structured execution model for interactive examples. RStudio supports R Markdown and Quarto publishing directly from package and project sources, which matters when the CD library includes reproducible documentation outputs.
A decision framework for selecting CD library software based on release mechanics and controls
The selection starts with the release unit and the promotion path. A CD library built from datasets and transformation logic points toward Snowflake, BigQuery, and dbt, while a library built around notebooks and job execution points toward Databricks and Azure Synapse Analytics.
The second decision is whether orchestration and governance must be handled inside one platform or can be split across tools. Apache Airflow fits when orchestration needs explicit DAG governance with REST API visibility, and Apache Superset fits when the delivery target is governed dashboards with row-level security.
Define the CD library release unit and map it to the tool’s versioning model
If the library must ship versioned datasets with rollback, Snowflake’s Time Travel is a direct match because it supports querying and restoring prior data versions. If the library must ship analytics outputs tied to experiments and modeling, Google BigQuery’s BigQuery ML keeps training and prediction inside the same released dataset model.
Choose the execution orchestrator that matches how releases run
For integrated pipelines that deploy and run with notebooks plus SQL scripts, Azure Synapse Analytics provides built-in pipeline orchestration in a unified workspace. For Git-based notebook and artifact promotion with scheduled or event-driven job orchestration, Databricks ties job runs to Git-integrated notebook promotion patterns.
Validate automation and API surface for run control and observability
If programmatic access to run state, logs, and dependency status is required, Apache Airflow exposes those signals in its web UI and REST API. If orchestration lives inside a data platform workspace, Databricks and Azure Synapse Analytics reduce cross-tool integration by keeping pipeline definitions and execution in one place.
Ensure governance controls cover both datasets and library artifacts
When access control must apply to datasets and permissioning across environments, Databricks Unity Catalog centralizes permissions for datasets across environments. When the library must publish governed data sharing with controlled access, Snowflake’s role-based access control plus secure data sharing supports consistent release governance.
Use dbt or transformation-first workflows when the library is SQL logic with tests
For CD libraries centered on reusable transformations, dbt provides code-first modeling with automated tests and documentation generation, which keeps the library’s quality checks tied to the logic. This approach also supports reusable packages so transformation patterns become shareable assets across teams.
Pick publishing surfaces only if they match the delivery goal
If the library delivery target is interactive data experiences, Observable offers reactive cells that rerun automatically when upstream variables change. If the delivery target is reproducible documentation outputs for R teams, RStudio’s R Markdown and Quarto publishing connects library sources to documentation releases.
Which teams get the most value from these CD library software tools
Different tools fit different library delivery targets like datasets, transformation logic, pipelines, dashboards, or interactive examples.
The strongest matches align to each tool’s best_for profile, especially around orchestration depth, governance coverage, and how versioning and rollback are handled for released assets.
Large-scale CD library analytics and experiment reporting teams
Google BigQuery fits teams that need CD library analytics on large datasets because its serverless SQL engine handles high-volume analytics without cluster management and BigQuery ML keeps training and prediction inside BigQuery.
Enterprises standardizing repeatable analytics delivery with reusable pipelines
Azure Synapse Analytics is the match for enterprises standardizing analytics delivery since it combines a unified workspace with built-in pipeline orchestration that ties together notebooks and SQL scripts.
Teams delivering governed data pipelines with Git-based promotion and automated jobs
Databricks fits teams because job orchestration runs against versioned notebooks and Git-integrated workspace repositories, and Unity Catalog centralizes permissions for datasets across environments.
Data-driven CD libraries needing governed datasets and rollback via dataset versioning
Snowflake matches CD library needs when time-based rollback is a release requirement because Time Travel enables querying and restoring prior data versions under governed access controls.
Analytics and data engineering teams building reusable transformation libraries with tests
dbt fits teams building transformation libraries since it provides code-first modeling with dependency-managed models, automated tests, and documentation generation tied to SQL logic.
Release pitfalls that break CD library governance and automation
Common failure modes show up when release artifacts do not map to the tool’s native versioning and orchestration mechanisms.
Another recurring issue is governance gaps where access control covers dashboards but not underlying datasets, or where rollback exists for queries but not for released assets.
Treating orchestration as an afterthought and building ad hoc run logic
Teams that need dependency-aware scheduling and visibility should use Apache Airflow’s DAG-based orchestration with configurable retries, backfills, and dependency tracking instead of relying on manually triggered scripts.
Releasing datasets without rollback mechanics or dataset version governance
For CD library workflows that publish curated datasets, Snowflake’s Time Travel is a concrete mechanism to support querying and restoring prior data versions rather than rebuilding pipelines under pressure.
Building a transformation library without code-defined tests and documentation generation
Teams creating reusable transformation libraries should model with dbt so automated tests and documentation generation stay tied to the transformation logic instead of living in separate spreadsheets.
Assuming dashboard access controls cover dataset access
Teams publishing governed analytics in Apache Superset should confirm row-level security maps to authenticated identities and that dataset permissions in the upstream system also enforce RBAC so the dashboard cannot bypass governance.
Choosing an interactive notebook surface for backend-heavy library distribution
Observable is browser-first and focuses on reactive cells and publishable notebooks, so it becomes a poor fit for CD libraries that require persistent backend services and heavyweight library packaging.
How We Selected and Ranked These Tools
We evaluated Google BigQuery, Azure Synapse Analytics, Databricks, Snowflake, dbt, Apache Airflow, Apache Superset, RStudio, Kaggle Datasets, and Observable using criteria aligned to feature coverage, ease of use, and value, with features weighted most heavily because CD library workflows depend on concrete orchestration, versioning, and governance mechanics.
The overall rating uses a weighted average where features carry the strongest influence at 40 percent, while ease of use and value each account for 30 percent of the final score.
Google BigQuery set itself apart for the top position by combining a serverless SQL engine that supports high-volume analytics with BigQuery ML for training and predicting using SQL inside BigQuery, which directly lifts both features coverage and practical suitability for analytics-backed CD library reporting.
Frequently Asked Questions About Cd Library Software
How do BigQuery, Databricks, and Snowflake differ as the CD library analytics backend?
Which tool is the best fit for Git-driven promotion of CD library artifacts?
What integration paths and APIs support automation across CD library pipelines?
How does SSO and RBAC enforcement work for governed CD library operations?
What migration strategy works best when moving an existing CD library data model into a target warehouse?
How should teams structure environment promotion for dev, staging, and production?
Which tool handles high-throughput refresh and orchestration for CD library reporting workloads?
What governance mechanism best fits versioned datasets and release rollback needs?
How do teams manage documentation and data contracts for CD library assets?
When CD library workflows require interactive analysis artifacts, how do Superset, Observable, and RStudio compare?
Tools reviewed
Primary sources checked during evaluation.
Referenced in the comparison table and product reviews above.
Keep exploring
Comparing two specific tools?
Software Alternatives
See head-to-head software comparisons with feature breakdowns, pricing, and our recommendation for each use case.
Explore software alternatives→In this category
Data Science Analytics alternatives
See side-by-side comparisons of data science analytics tools and pick the right one for your stack.
Compare data science analytics tools→FOR SOFTWARE VENDORS
Not on this list? Let’s fix that.
Our best-of pages are how many teams discover and compare tools in this space. If you think your product belongs in this lineup, we’d like to hear from you—we’ll walk you through fit and what an editorial entry looks like.
Apply for a ListingWHAT THIS INCLUDES
Where buyers compare
Readers come to these pages to shortlist software—your product shows up in that moment, not in a random sidebar.
Editorial write-up
We describe your product in our own words and check the facts before anything goes live.
On-page brand presence
You appear in the roundup the same way as other tools we cover: name, positioning, and a clear next step for readers who want to learn more.
Kept up to date
We refresh lists on a regular rhythm so the category page stays useful as products and pricing change.