Salesforce Einstein Analytics
Wave / CRM Analytics
Course in Pune

Getting data into CRM Analytics correctly is the foundation that determines everything else — if the data model is wrong, no amount of dashboard skill can produce accurate insights. This module covers the two primary data ingestion methods. Data Sync (previously called the Salesforce Local Connector) is how Salesforce native object data is pulled into CRM Analytics — we cover configuring which objects and fields to sync, full sync vs incremental sync (why incremental is faster and when it breaks), sync scheduling, and the Data Manager monitoring interface for diagnosing sync failures. Dataflows are the transformation engine — they take synced data and process it through a series of nodes to produce the datasets that dashboards consume. We cover every important dataflow node: sfdcDigest (extracting Salesforce data), Augment (the analytics equivalent of a JOIN — enriching one dataset with fields from another), Append (stacking multiple datasets vertically), ComputeExpression (adding calculated fields), Filter (excluding records), Flatten (denormalising hierarchies like Salesforce's role hierarchy for row-level security), and sfdcRegister (outputting the finished dataset). External data connectors (Snowflake, Amazon S3, CSV upload, remote Salesforce org connectors) are covered with practical examples. Understanding the full data ingestion layer is the most technically demanding part of CRM Analytics work and one of the most heavily weighted areas of the consultant certification exam.

Salesforce introduced Data Recipes as a more user-friendly, visual alternative to the JSON-based Dataflow editor — and for many data preparation tasks, Recipes are now the preferred approach. Understanding both methods and knowing when to use each is important for any CRM Analytics professional working on real projects. This module covers the Data Recipe editor in depth: adding input nodes (Salesforce objects, existing datasets, CSV files), applying transformations using the visual toolbar (Filter, Formula columns, Aggregate, Join — which is the Recipe equivalent of Augment, Flatten, Bucket, and Add Column), scheduling recipe runs, and monitoring recipe execution in the Data Manager. We pay particular attention to Join operations in Recipes — configuring left, inner, and lookup joins between datasets — because data joining is one of the most common requirements in analytics data preparation and also one of the most common sources of errors when not configured carefully. We also cover recipe limits (the maximum number of rows a recipe can process) and when complex transformations require a dataflow approach instead of a recipe.

A Lens is CRM Analytics's ad-hoc data exploration view — you pick a dataset and start slicing and dicing it visually. This is where users explore data before building formal dashboards. This module covers creating lenses from datasets, applying groupings and measures, switching between chart types (bar charts, line charts, donut charts, scatter plots, heat maps, compare tables, timeline charts), applying filters in the lens, clipping lenses to dashboards, and the lens explorer's interaction model. The bulk of this module is Dashboard development — building the polished, interactive analytics applications that business users work with daily. We cover the CRM Analytics dashboard editor: adding widgets (charts, tables, numbers, text, images), configuring steps (the data queries that power each widget), connecting widgets through faceting (where selecting a value in one chart automatically filters all other charts on the dashboard), building global filters for date ranges and dimensions, and creating binding expressions that make dashboards dynamic. Dashboard design principles are taught alongside technical configuration — we work through examples of poorly designed dashboards (too many charts, unclear hierarchy, no narrative) and how to redesign them for clarity and genuine usefulness. Students build three complete dashboards during this module: a Sales Pipeline Overview, a Service Team Performance dashboard, and an Executive Summary with dynamic date range controls.

SAQL (Salesforce Analytics Query Language) is what separates CRM Analytics developers who can build basic dashboards from those who can build genuinely sophisticated analytical applications. It is CRM Analytics's native query language — used to write custom queries that power dashboard steps when the visual configuration tools are not enough. SAQL is similar in concept to SQL but designed for CRM Analytics's denormalised dataset structure and optimised for aggregation and time-series analysis. This module starts from complete basics — what SAQL is, when you need it, and how it relates to the visual query builder — and builds through the key SAQL patterns that analytics developers use regularly. We cover the SAQL syntax: q = load "datasetName", q = filter q by condition, q = group q by dimension, q = foreach q generate expression, q = order q by field, q = limit q, and the SAQL functions that are most commonly needed — date functions (toDate, daysBetween, date), string functions (left, right, trim, length), comparison functions, and aggregate functions (sum, avg, count, min, max). We spend significant time on time-series SAQL — building queries that compare current period vs prior period, calculate rolling averages, and identify trends — because time comparison is one of the most common analytical requirements and one of the most complex to implement correctly. Students write SAQL queries for five real analytical scenarios: year-over-year sales comparison, running total calculation, top N performers with dynamic threshold, multi-dimensional funnel analysis, and period-over-period churn rate. The ability to write SAQL is one of the most asked-about skills in CRM Analytics interviews and consultant exam scenarios.

Security in CRM Analytics is more complex and more important than most beginners expect — and it is one of the most heavily tested areas in the consultant certification exam. The core challenge is this: CRM Analytics datasets are denormalised copies of Salesforce data, and Salesforce's standard sharing and visibility rules do not automatically apply to these copies. A sales rep who can only see their own opportunities in Salesforce might be able to see everyone's opportunities in a CRM Analytics dataset if security is not configured correctly. This module covers the four mechanisms for securing CRM Analytics: App Sharing (controlling who can see a CRM Analytics app), Dataset Sharing (controlling who can see individual datasets), Row-Level Security (RLS — the most important mechanism, controlling which rows of data each user can see based on their profile, role, or attributes), and the Security Predicate — a SAQL-like expression embedded in the dataset that filters rows at query time based on the running user's identity. We cover the common RLS patterns: User attribute predicates (the current user can only see their own records), Role hierarchy predicates (using the Flatten node to model Salesforce's role hierarchy in the dataset so managers can see their subordinates' data), and multi-object predicates (applying visibility rules based on relationships across multiple objects). We also cover Salesforce Sharing Inheritance — the mechanism that automatically applies Salesforce org-level sharing rules to CRM Analytics — and when to use it versus custom security predicates. Security configuration exercises in this module mirror the exact scenarios that appear in the consultant exam.

Einstein Discovery is the component of CRM Analytics that elevates the platform from business intelligence (showing what happened) to predictive intelligence (showing what is likely to happen and what to do about it). It uses machine learning algorithms to analyse your Salesforce data, identify statistically significant patterns and correlations, and present findings in natural language stories that business users can understand without data science expertise. This module covers the complete Einstein Discovery workflow. Story creation: selecting a dataset, choosing the outcome variable you want to predict (deal won/lost, customer churn yes/no, case resolution time), selecting the predictor variables, and running the automated model training. Story interpretation: reading the Insights panel (which factors most influence the outcome), the Prediction panel (which shows predicted outcome scores for new records), and the Improvement panel (which recommends specific actions that would improve the probability of a positive outcome). Model governance: understanding confidence scores, coverage rates, and when a model is reliable enough to act on. Write-back configuration: pushing Einstein Discovery prediction scores and recommended actions back into Salesforce records — so a sales rep can see their deal's "probability to close" score in the Opportunity record and the recommended actions to improve it directly in Salesforce. We also cover how to evaluate model quality, identify when a model needs retraining, and the data requirements for reliable Einstein Discovery models (sufficient historical data, balanced outcomes, relevant predictor variables). Einstein Discovery is tested in a dedicated section of the CRM Analytics Consultant exam and is increasingly the capability that enterprise clients prioritise when implementing the platform.

Extended Metadata (XMD) is CRM Analytics's system for controlling how datasets are presented to end users — field labels, field types, formatting, measure vs dimension classification, colours, and icons. Well-configured XMD makes dashboards significantly easier to use because field names display as business-friendly labels rather than technical API names, measures are correctly classified for aggregation, and colours align with brand guidelines. This module covers the XMD editor, bulk XMD updates, and the common XMD configurations that every analytics developer needs to understand. App design covers building CRM Analytics Apps — the packaged collections of dashboards and datasets that form the analytics products delivered to business users — including setting app icons, navigation, featured dashboards, and app sharing. Advanced dashboard features covered include embedded dashboards (embedding CRM Analytics dashboards directly in Salesforce record pages and Visualforce/LWC), custom actions that let dashboard users navigate to Salesforce records or trigger flows, Watchlist for user-defined alerts, Collections for organising related analytics assets, and the Template App framework that allows analytics solutions to be packaged and deployed across multiple orgs. We also cover the dashboard migration process using Workbench and the Analytics REST API — essential for moving dashboards from sandbox to production environments in enterprise deployment workflows.

The final module ties practical real-world experience with exam preparation. Industry use cases covers the most common CRM Analytics implementation scenarios across key verticals — Sales Cloud analytics (pipeline health, forecast accuracy, activity trends, territory performance), Service Cloud analytics (case volume trends, SLA compliance, agent productivity, CSAT correlation analysis), Financial Services Cloud analytics (relationship manager performance, portfolio health, lead conversion), and how to approach scoping and designing a CRM Analytics solution for a new client. Performance optimisation covers techniques for keeping dashboards fast — dataset partitioning (dividing large datasets by date for faster queries), compact form factors for dataset storage, query optimisation, and the Analytics Monitor for identifying slow queries. We also cover dataset version control and rollback, scheduled dataflow management, and the Analytics Notification framework for automated email alerts when dashboard thresholds are breached. The CRM Analytics and Einstein Discovery Consultant exam preparation covers all six exam domains: Data Layer (20%), Security (20%), Admin (18%), Analytics Dashboard Design (22%), Analytics Dashboard Implementation (20%), and Einstein Discovery Story Design (remaining %). Three full timed mock exams under exam conditions are included, with detailed review of every wrong answer and coaching on the exam's scenario-based question format. Exam registration guidance via Webassessor is included. The exam is considered moderately difficult — it requires both conceptual understanding and hands-on experience, which is exactly why the Aapvex programme prioritises live org practice throughout.