• Cert++
  • Practice
  • Certle
  • Review
  • Tracks
  • Checklist
  • Guides
  • Upgrade
Cert++
  1. Home
  3. CRM Analytics and Einstein Discovery Consultant

CRM Analytics and Einstein Discovery Consultant

Checklist progress

0/196Learned

CRM Analytics and Einstein Discovery Consultant

Study Checklist

  • Platform Administrator
  • Platform App Builder
  • Platform Foundations
  • Platform Developer
  • Platform Administrator II
  • Agentforce Sales Consultant
  • Agentforce Service Consultant
  • Platform Data Architect
  • Platform Development Lifecycle and Deployment Architect
  • Platform Identity and Access Management Architect
  • Platform Integration Architect
  • Platform Sharing and Visibility Architect
  • Heroku Architect
  • B2C Solution Architect
  • Experience Cloud Consultant
  • Agentforce Field Service and Operations Consultant
  • Agentforce Nonprofit Consultant
  • Data 360 Consultant
  • Omnistudio Consultant
  • CRM Analytics and Einstein Discovery Consultant
  • Platform User Experience Designer
  • Platform Strategy Designer
  • B2C Commerce Developer
  • JavaScript Developer
  • Omnistudio Developer
  • Platform Developer II
  • Marketing Cloud Engagement Administrator
  • Marketing Cloud Engagement Specialist
  • Marketing Cloud Engagement Consultant
  • Agentforce Sales Foundations
  • Business Analyst
  • Marketing Cloud Engagement Developer
  • Marketing Cloud Engagement Foundations
  • Agentforce Specialist
  • Agentforce Life Sciences Consultant
  • B2B Commerce Administrator AP
  • B2B Commerce Developer AP
  • Agentforce Consumer Goods AP
  • Agentforce Financial Services AP
  • Agentforce Health AP
  • Agentforce Manufacturing AP
  • MuleSoft Integration Foundations
  • MuleSoft Developer
  • MuleSoft Developer II
  • MuleSoft Platform Integration Architect
  • MuleSoft Platform Architect
  • Tableau Desktop Foundations
  • Tableau Data Analyst
  • Tableau Consultant
  • Tableau Server Administrator
  • Tableau Architect

Checklist progress

0/196Learned

  • What steps are required to enable CRM Analytics in a Salesforce org, including org-level settings and prerequisite licenses?
  • What is the difference between the CRM Analytics Plus license and the Einstein Predictions license in terms of what each grants access to?
  • Which permission set licenses are required to grant a user full CRM Analytics creator access versus explorer (viewer) access?
  • A company needs some users to build dashboards and datasets while others only need to view them. How should the administrator assign CRM Analytics permission set licenses to satisfy least-privilege access?
  • A company wants to enable Einstein Discovery for their sales team but only a subset of users need to build models. Which licenses and permission sets must be assigned to model builders versus prediction consumers?
  • Which org-level permission must be enabled in Setup to allow Einstein Discovery model deployment and prediction writeback to Salesforce fields?
  • Which CRM Analytics feature must be explicitly enabled in Setup before users can embed analytics dashboards in Lightning record pages?
  • What is the maximum number of rows that a single CRM Analytics dataset can contain?
  • What are the key governor limits that apply to CRM Analytics data syncs, dataflows, and recipes?
  • How does the CRM Analytics dataset row limit constrain dataflow or recipe design?
  • How do CRM Analytics governor limits affect solution design for data syncs, dataflows, and recipes?
  • A business requirement calls for near-real-time data in CRM Analytics dashboards for a large Salesforce object with millions of rows. Which data ingestion strategy best accommodates row limits and refresh windows?
  • A company runs 12 dataflows daily and is approaching their org's dataflow limit. What approach should the consultant recommend to stay within platform limits while preserving refresh frequency?
  • What dataset management operations can be performed using the CRM Analytics REST API?
  • What SAQL query and dashboard interaction operations can be performed using the CRM Analytics REST API?
  • A developer needs to programmatically query CRM Analytics datasets and retrieve results for an external application. Which CRM Analytics API capability enables this use case?
  • A consultant wants to automate the deployment of CRM Analytics assets between environments using the API. Which API approach supports this, and what are its limitations compared to the Analytics app?
  • What is the recommended approach for migrating CRM Analytics assets (dashboards, datasets, dataflows) from a sandbox to production?
  • When should a consultant use Deployment Manager versus manual app export/import for CRM Analytics migration?
  • A company uses change sets to deploy Salesforce metadata. Which CRM Analytics assets can be included in a change set, and which require an alternative migration method?
  • When deploying a CRM Analytics app from sandbox to production using the Analytics app export/import feature, what considerations must the consultant address regarding data source connections and credentials?
  • What is the role of the Data Manager in CRM Analytics?
  • What types of data connections does Data Manager support for loading data into datasets?
  • What external data connectors does CRM Analytics support natively (e.g., Amazon S3, Snowflake, Google BigQuery)?
  • What are the key prerequisites for configuring each type of CRM Analytics external data connector?
  • What is a data sync in CRM Analytics, and how does it differ from a direct dataset load via a dataflow or recipe?
  • How does a local file upload differ from a connected data sync when creating a dataset in CRM Analytics, including limitations on row count and refresh capability?
  • What is Direct Data Input (DDI) in CRM Analytics, and what limitations apply to datasets created via this method compared to those created by a recipe or dataflow?
  • A consultant needs to load data from an external relational database into CRM Analytics. Which connection type in Data Manager supports this, and what are the prerequisites?
  • A company stores historical sales data in Amazon S3. How should the consultant configure a Data Manager connection to ingest this data into CRM Analytics?
  • A business requires combining Salesforce object data with CSV files in a single CRM Analytics dataset. What is the most appropriate approach using Data Manager?
  • What scheduling limitations apply when using Data Manager to ingest data from external sources such as Amazon S3?
  • What scheduling options are available for data sync refreshes, and what is the minimum supported refresh interval?
  • How do incremental data syncs reduce load on the platform compared to full syncs?
  • What configuration is required to enable incremental sync for a Salesforce object in CRM Analytics?
  • What happens when a CRM Analytics dataflow or recipe exceeds the platform's maximum allowed runtime?
  • How should a consultant prevent a dataflow or recipe from exceeding the platform's maximum runtime?
  • A recipe is configured to run immediately after a data sync completes. The sync runs at unpredictable times. How should the consultant chain the recipe execution to the sync completion to avoid stale data?
  • A consultant observes that a recipe frequently fails due to upstream sync delays. What design pattern should be applied to make the recipe more resilient to sync timing dependencies?
  • A company's dataflow takes 4 hours to run and needs to complete before the business day begins. The org also has 3 other dataflows running nightly. How should the consultant configure scheduling to stay within limits and meet the timing requirement?
  • How are deleted records in a Salesforce object handled during incremental data sync in CRM Analytics, and what steps must a consultant take to ensure deleted records are removed from the dataset?
  • What is the difference between a dataflow and a recipe in CRM Analytics?
  • What criteria should guide the choice between a dataflow and a recipe for a new data transformation project?
  • What is the sfdcDigest node in a CRM Analytics dataflow, and how does it differ from using a Data Manager sync as input to a recipe?
  • What are the common data type transformation operations available in CRM Analytics recipes, such as date formatting, type casting, and formula columns?
  • What recipe transformation nodes are available for aggregating data (e.g., grouping and computing sums, averages, counts)?
  • How do CRM Analytics aggregation nodes affect the output dataset's row count?
  • How does the Augment (lookup join) node in a dataflow differ from the standard Join node, and when should each be used?
  • How does the Flatten transformation in a dataflow handle hierarchical data (such as a role hierarchy), and when is it the appropriate choice?
  • A recipe needs to join two Salesforce objects on a lookup field and add a calculated column. Which transformation nodes accomplish this, and in what order should they be applied?
  • A recipe needs to filter out cancelled opportunities before writing to the output dataset. Which node should be used, and how is the filter condition specified?
  • A consultant needs to append rows from two datasets of the same schema into a single output dataset in a recipe. Which node type supports this, and what schema requirements must be met?
  • A company needs to bucket opportunity amounts into revenue tiers (small, medium, large) in their CRM Analytics dataset. Which transformation or recipe node supports this, and how is it configured?
  • A recipe joins two large datasets and the output dataset exceeds the row limit. What options does the consultant have to resolve this within the recipe?
  • What is extended metadata (XMD) in CRM Analytics, and what aspects of a dataset can it control?
  • What XMD configurations can be applied to a measure field, such as defining display format, currency symbol, or order in widgets?
  • How can a consultant use XMD to assign consistent colors to dimension values across all dashboards that use a dataset?
  • A consultant wants to display friendly labels for a field's values (e.g., 'Closed Won' instead of an internal code) without changing the underlying data. How should this be accomplished using XMD?
  • When is it more appropriate to edit XMD versus transforming the data in a recipe to improve label clarity for end users?
  • How does CRM Analytics support versioning for dataflows?
  • How can a consultant roll back to a previous dataflow version after a breaking change?
  • A company wants to convert an existing dataflow to a recipe. What steps and considerations are involved, and what functionality available in dataflows may not be directly replicated in recipes?
  • After converting a dataflow to a recipe, the output dataset field types differ from the original. What step in the conversion process causes this, and how should the consultant resolve the discrepancy?
  • How do CRM Analytics user roles and permission sets interact with Salesforce profiles to control access to analytics apps and assets?
  • What is the difference between Manager access and Viewer access when sharing a CRM Analytics app?
  • What actions does each app sharing permission level (Manager, Viewer, Edit) allow when sharing a CRM Analytics app?
  • A company requires that only members of a specific Salesforce public group can access a particular CRM Analytics app. How should the consultant configure asset-level security to enforce this?
  • A new employee should be able to build dashboards in a CRM Analytics app but must not be able to share the app with other users. Which permission level should be assigned?
  • How does Salesforce field-level security interact with CRM Analytics datasets?
  • If a field is hidden by FLS, what is the effect when that field is synced into a CRM Analytics dataset?
  • What is sharing inheritance in CRM Analytics, and when is it the preferred approach over a custom security predicate?
  • What is a security predicate in CRM Analytics?
  • How does a security predicate enforce row-level security on a CRM Analytics dataset?
  • What is the Row Level Security (RLS) dataset pattern used with security predicates in CRM Analytics?
  • A security predicate uses a user attribute (e.g., UserRoleId) to filter rows. What are the steps to define a user attribute and reference it in a predicate?
  • A company requires that sales reps only see opportunities owned by them or their direct reports. How should the consultant implement row-level security using a security predicate, and what dataset columns are required?
  • How do CRM Analytics predicate functions differ from a simple equality predicate?
  • What predicate functions (e.g., 'hasRole', 'inGroup') are available in CRM Analytics security predicates?
  • How does a consultant configure a security predicate to reference a separate RLS lookup dataset?
  • A business has complex territory-based access rules that do not map directly to Salesforce record ownership or role hierarchy. Why is sharing inheritance insufficient, and what approach should the consultant use instead?
  • How does the security predicate interact with the dataset's columns at query time, and what happens when a required column referenced by the predicate is missing from a row?
  • A dataset contains a column 'OwnerId' used in a security predicate. A new recipe run adds rows where 'OwnerId' is null. What is the security impact, and how should the predicate or recipe be modified to handle null values safely?
  • What are the three access levels available when sharing a CRM Analytics app?
  • Which operations does each app sharing access level (View, Edit, Manage) permit in CRM Analytics?
  • How does sharing an app affect access to individual assets (dashboards, datasets, lenses) contained within it?
  • A company wants a Salesforce public group of regional managers to have View access to a CRM Analytics app while a smaller group of analysts has Edit access. How should app sharing be configured to meet both requirements simultaneously?
  • A user has Viewer access to an app but the dataset powering its main dashboard has a security predicate. The user can open the dashboard but sees no data. What is the most likely cause, and how should the consultant investigate?
  • What is the role of a KPI inventory and data availability assessment in validating that a proposed CRM Analytics dashboard is feasible before development begins?
  • How should a consultant prioritize conflicting dashboard requirements from multiple business stakeholders when time and resources are constrained?
  • A stakeholder provides vague requirements: 'We need a sales dashboard.' What discovery questions and scoping techniques should the consultant use to define measurable, prioritized dashboard requirements?
  • A consultant is gathering requirements for a new dashboard. The business sponsor wants 20 KPIs on a single page, but user interviews reveal most users only care about 4. How should the consultant handle this conflict during the scoping phase?
  • What is a CRM Analytics lens?
  • How does a CRM Analytics lens differ from a dashboard as a deliverable to end users?
  • What CRM Analytics best practices govern layout, widget density, and color use to ensure dashboards are readable and actionable for business users?
  • What is the Flex layout in CRM Analytics dashboards, and how does it differ from the fixed-grid layout in terms of responsive design and widget positioning?
  • What is the purpose of the container widget in CRM Analytics dashboard design?
  • How does the container widget improve user experience in CRM Analytics dashboards?
  • A consultant is designing a dashboard for executive users who need a quick summary of pipeline health. Which widget types and layout principles should be prioritized, and why?
  • How should a dashboard be structured differently for a mobile audience versus a desktop audience in CRM Analytics?
  • A consultant is building a dashboard that must render correctly on both desktop and mobile. Which layout option and widget configuration practices should be used?
  • What is a CRM Analytics templated app, and how does it differ from a custom-built analytics app?
  • How is a CRM Analytics templated app configured during installation?
  • What data sources and field mappings must be provided when configuring a CRM Analytics templated app during installation?
  • What limitations does a templated CRM Analytics app impose on customization, and when should a consultant recommend a custom app instead?
  • A company wants to quickly deploy sales performance analytics for their Sales Cloud users without custom development. Which CRM Analytics templated app should be recommended, and what are its configuration prerequisites?
  • Which standard CRM Analytics templated apps are available out of the box for Service Cloud use cases?
  • What datasets do Service Cloud templated CRM Analytics apps require?
  • A company has installed the Sales Analytics templated app and the vendor releases an updated version. What is the upgrade process, and what customizations risk being overwritten?
  • What is SAQL, and how does it differ from SOQL for querying data in CRM Analytics?
  • What are the different query types available in CRM Analytics (SAQL, compact, grouped)?
  • When should each CRM Analytics query type (SAQL, compact, grouped) be used in dashboard widget configuration?
  • When is the compact form query sufficient for a CRM Analytics dashboard widget?
  • What are the limitations of the compact form query compared to a full SAQL query in CRM Analytics?
  • What are the four main SAQL query modes (aggregateflex, values, timeseries, cogroup)?
  • When is each SAQL query mode (aggregateflex, values, timeseries, cogroup) appropriate for a dashboard widget?
  • A chart widget needs to display the top 10 accounts by revenue with a custom sort. Which query type and widget parameters should the consultant configure?
  • A chart widget must show monthly trend data over the past 12 months. Which SAQL query type is most appropriate, and what date grouping syntax must be used?
  • A business requires a number widget displaying a percentage metric calculated from two measures. How should this be implemented in the dashboard query?
  • How does configuring the 'limit' and 'offset' parameters in a SAQL query affect what data is displayed in a dashboard widget?
  • What is the role of steps in a CRM Analytics dashboard, and how do they enable cross-widget interactions?
  • What is the difference between a selection interaction and a result interaction in a CRM Analytics dashboard, and when should each be used?
  • What is a static step in a CRM Analytics dashboard, and how does it differ from a query-backed step in terms of how it stores and provides values for widget interactions?
  • How does a faceting interaction differ from a manual binding in CRM Analytics, and what are the trade-offs of each approach?
  • A dashboard requires a filter that lets users switch between 'This Quarter', 'Last Quarter', and 'This Year' time ranges. A query to a dataset is not needed for the options themselves. Which step type should the consultant use, and how is it configured?
  • A consultant needs to configure a global filter that applies to all queries on a dashboard page simultaneously. What approach should be used to implement a single filter widget that drives all step queries?
  • A dashboard has a toggle widget that should filter multiple charts to a selected time period. How should the consultant configure the interaction binding to propagate the selection across queries?
  • A consultant wants a list widget selection to drive a SAQL query on a second widget. What JSON configuration is required to implement this result interaction?
  • How can a consultant configure cell-based bindings in the CRM Analytics dashboard JSON to pass a specific result cell value from one widget's query into another widget's SAQL query?
  • What is a compare table in CRM Analytics?
  • What types of calculations does a CRM Analytics compare table support beyond a standard chart widget?
  • How does the time series analysis feature in compare tables work, and what data requirements must be met for it to function correctly?
  • A company wants to display a rolling 3-month average of closed-won deals in a compare table. Which windowing function should be used, and how is it configured?
  • A business needs to show month-over-month revenue change percentage in a CRM Analytics dashboard. How should the consultant implement this using a compare table with windowing functions?
  • A consultant is asked to show each region's sales as a percentage of the company-wide total in a single compare table query. Which windowing function accomplishes this, and what partitioning syntax is required?
  • What conditional formatting options are available in a CRM Analytics compare table, and how are thresholds and color rules configured?
  • What are the methods for embedding a CRM Analytics dashboard in a Salesforce Lightning record page?
  • What configuration is needed for each method of embedding a CRM Analytics dashboard in a Lightning record page?
  • What are CRM Analytics actions, and how can they be configured to allow users to update Salesforce records directly from a dashboard?
  • How does a CRM Analytics URL navigation action work, and how can it be configured to open a related Salesforce record or external URL from within a dashboard widget?
  • A manager wants to see the CRM Analytics pipeline dashboard directly on the Account record page, filtered to that account's opportunities. How should the consultant configure the embedded dashboard to respect the record context?
  • A company wants users to be able to create a follow-up task directly from a CRM Analytics dashboard widget. What type of action should be configured, and what are the prerequisites?
  • A consultant needs to configure a dashboard widget so that clicking a row navigates the user to a different dashboard page filtered to that row's values. Which action type and configuration accomplish this?
  • What is the Dashboard Inspector in CRM Analytics, and what metrics does it provide to diagnose slow dashboard performance?
  • A dashboard takes over 30 seconds to load. The Dashboard Inspector shows several queries with high execution time. What optimization strategies should the consultant apply to reduce query runtime?
  • The Dashboard Inspector shows a query with a very high 'Result Count' but the widget only renders 10 rows. What does this indicate, and what change should the consultant make to the query?
  • A consultant identifies that a query is scanning a very large dataset on every dashboard load. What pre-aggregation or dataset design changes would most improve performance?
  • How can reducing the number of dimensions or measures in a query and applying dataset-level filters improve CRM Analytics dashboard performance?
  • How can reordering or parallelizing queries in the Dashboard Inspector improve perceived dashboard load time?
  • How does dashboard versioning work in CRM Analytics?
  • How can a consultant restore a previous dashboard version after an unintended change in CRM Analytics?
  • A development team needs to maintain a stable 'production' version of a dashboard while continuing to develop a new version. How should they use CRM Analytics versioning to support this workflow?
  • What is Dashboard Publisher in CRM Analytics, and how does it differ from standard dashboard sharing?
  • What is a published dashboard in CRM Analytics?
  • How does publishing a dashboard version affect what viewers see when the underlying data refreshes?
  • A company wants to distribute a static snapshot of a CRM Analytics dashboard to executives who do not have CRM Analytics licenses. What delivery mechanism should the consultant recommend?

Build a model by assessing data and selecting one of the three types of predictions

0/11

  • What are the three prediction types in Einstein Discovery (numeric, binary, multi-classification)?
  • What business question does each Einstein Discovery prediction type (numeric, binary, multi-classification) answer?
  • A company wants to predict whether a support case will be escalated (yes/no). Which Einstein Discovery prediction type should be used, and what data requirements must be met?
  • A company wants to predict which product category an opportunity will close in (one of six categories). Which Einstein Discovery prediction type is appropriate, and how does multi-class classification differ from binary in terms of model outputs?
  • What is the minimum recommended dataset size and row count for training a reliable Einstein Discovery model?
  • How should a consultant assess data quality and completeness before building an Einstein Discovery model, and which data issues most significantly degrade model accuracy?
  • What is the difference between an Einstein Discovery Story and a deployed Einstein Discovery Model?
  • How does the transition from an Einstein Discovery Story to a deployed model occur?
  • How does Einstein Discovery handle the training/test split?
  • What does the holdout test set in Einstein Discovery reveal about model generalization?
  • A company's dataset has a highly imbalanced target variable (e.g., 95% negative outcomes). How does this affect model training in Einstein Discovery, and what steps should the consultant take to address it?

Given business requirements, analyze the model results and propose data improvements to the customer.

0/8

  • What metrics (e.g., AUC, RMSE, accuracy) does Einstein Discovery provide to evaluate binary and numeric model performance?
  • What metric values indicate a well-performing Einstein Discovery model?
  • How does the Einstein Discovery Story view present model results, including key drivers, contributing factors, and predictive strength indicators?
  • What are 'prescriptions' (improvement suggestions) in an Einstein Discovery Story, and how should a consultant present them to a business stakeholder as actionable recommendations?
  • How does the confusion matrix displayed in an Einstein Discovery Story help the consultant evaluate false positive and false negative rates for a binary classification model?
  • A binary model has a high accuracy score but very low recall for the positive class. What does this indicate about the model, and what adjustment should the consultant propose?
  • After reviewing an Einstein Discovery model, a consultant finds that the top predictive factor is a data field that is unavailable at prediction time (a data leakage issue). What should they recommend to the customer?
  • A customer is concerned that their Einstein Discovery model is biased against a demographic group. What tools or features in Einstein Discovery allow the consultant to investigate and mitigate this concern?

Given derived results and insights from the model, adjust data parameters and add/remove data or columns to improve the model.

0/6

What process should a consultant follow to iteratively improve an Einstein Discovery model?

Learn this concept
Unseen

How does adding or removing predictor columns in Einstein Discovery affect model accuracy?

Learn this concept
Unseen

What is the role of the 'exclude' and 'include' controls in the Einstein Discovery model settings for managing predictor variables?

Learn this concept
Unseen

A consultant adds a date-derived feature (e.g., day of week) that improves model performance. How are custom feature engineering columns introduced into an Einstein Discovery dataset?

Learn this concept
Unseen

After retraining an Einstein Discovery model with additional rows, the AUC score does not improve. What alternative data quality or feature adjustments should the consultant investigate?

Learn this concept
Unseen

A model's accuracy improves significantly when a column with very high null rates is removed. Why does this happen, and how should the consultant decide which columns to exclude?

Learn this concept
Unseen

Enable prediction features on Lightning record pages across Salesforce and CRM Analytics.

0/7

  • What are the steps to deploy an Einstein Discovery prediction to a Salesforce Lightning record page, including model deployment and page layout configuration?
  • How does the Einstein Discovery prediction widget on a Lightning record page display score explanations to users, and what controls the level of detail shown?
  • What is writeback in Einstein Discovery, and how does configuring prediction writeback to a Salesforce field enable downstream automation or reporting?
  • What is the Einstein Prediction Service, and how does it differ from the Einstein Discovery prediction widget on a Lightning record page in terms of invocation mechanism and use case?
  • A company wants sales reps to see an Einstein Discovery win probability score directly on the Opportunity record page. What configuration steps are required, and what permissions must be granted?
  • A company wants to score all open opportunities in bulk overnight and write the scores back to a custom field for use in reports. Which Einstein Discovery prediction approach should be used, and what configuration is required?
  • A company wants to display Einstein Discovery predictions in a CRM Analytics dashboard rather than on a record page. What integration approach should the consultant use?

Monitor and interpret a Model Card to improve or maintain model performance.

0/9

  • What is a Model Card in Einstein Discovery?
  • What information does the Einstein Discovery Model Card surface about model performance, data distribution, and bias?
  • What is model drift in Einstein Discovery, and how is it detected using the Model Card?
  • How often should an Einstein Discovery model be retrained, and what signals in the Model Card indicate that retraining is overdue?
  • What remediation options does a consultant have when model drift is detected in Einstein Discovery?
  • How should a consultant use the Model Card's segment analysis to identify subpopulations where the model underperforms?
  • A consultant reviews a Model Card and notices that model accuracy has declined since the last training run. What actions should they take to investigate and remediate the performance degradation?
  • A model's AUC score on the Model Card is 0.61. What does this indicate about model performance, and what steps should the consultant recommend to improve it?
  • A Model Card shows that prediction accuracy is significantly lower for one demographic segment than for others. What does this indicate, and what steps should the consultant recommend to address the disparity?

Prepare for the Exam

Play Today's Certle
Back to track

Study Community

Ask questions and get the latest info from other CRM Analytics and Einstein Discovery Consultant studiers. 593 members and growing.

Go to Discord

A model's accuracy improves significantly when a column with very high null rates is removed. Why does this happen, and how should the consultant decide which columns to exclude?

Explainer

Learn More

Practice Question

Keep going

Next conceptWhat are the steps to deploy an Einstein Discovery prediction to a Salesforce Lightning record page, including model deployment and page layout configuration?

Checklist progress

0/196 (0%)

0 of 196 concepts learned

Tip: You can filter concepts by status.

Prepare for the Exam

Play Today's Certle
Back to track

Study Community

Ask questions and get the latest info from other CRM Analytics and Einstein Discovery Consultant studiers. 593 members and growing.

Go to Discord

Explainer

Einstein Discovery examines data for quality issues including frequently missing values, which can trigger data quality alerts during model training. Consultants improve models by selecting columns with strong associations to the outcome and removing extraneous columns that add complexity without analytical value.

Core information
  • Einstein Discovery analyzes data for quality issues such as duplicate impact (multicollinearity), potential bias, and frequently missing values, notifying consultants with data alerts.
More details and nuances
  • Data quality alerts appear in the model performance overview under Assess Deployment Readiness, with a View All Alerts button to examine all detected issues.