Data Sources
Native connectors to Zoho apps (CRM, Desk, Books, Inventory, Projects, People, Campaigns), plus 250+ external connectors covering databases (MySQL, PostgreSQL, MS SQL, Oracle, Snowflake, BigQuery, Redshift), cloud storage (S3, Azure Blob, Google Cloud Storage), spreadsheets (Excel, Google Sheets), files (CSV, JSON), and REST APIs. Pull data on schedule (15-min minimum) or via streaming for the database connectors that support CDC. The Zoho DataPrep module handles transformations before the data hits Analytics — schema normalization, deduplication, and type casting belong there, not in report-level formulas.
Reports
Six report types: Tabular (raw rows), Chart (column, bar, line, pie, area, scatter, heatmap, funnel), Pivot Table, Summary, KPI Widget, and Query (SQL-like via the Analytics Query Language). The drag-drop builder covers 80% of cases; AQL handles the remaining 20% where you need window functions, complex joins, or correlated subqueries. Custom formula columns support 200+ functions across math, string, date, lookup, and statistical categories.
Dashboards
Collections of reports with global filters, drill-downs, and parameter passing between widgets. Filters can be report-specific or dashboard-level (e.g. a Region filter applies to all 12 widgets). Embed in other apps via secure iframe with row-level filtering — pass the logged-in user’s identifier into the URL and Analytics returns only that user’s data. White-label the embed for customer-facing dashboards. Dashboard refresh respects the underlying data source schedule, so a 1-hour-stale Salesforce sync produces a 1-hour-stale dashboard regardless of when you load it.
Zia Insights
AI-driven insights: forecast (time series with seasonality detection), anomaly detection (flags outliers in time series), and natural-language queries through Ask Zia. Type “top 5 customers by revenue this quarter” and Zia builds the chart. The natural-language model understands metrics, dimensions, filters, and time ranges; it does not yet handle multi-step analytical questions (“show me customers whose spend dropped after their renewal date”). For those, fall back to AQL.
Scheduling
Reports can email on schedule (daily, weekly, monthly, custom cron), publish to a Slack/Teams/Cliq channel, or write back to a database table. Dashboards refresh on schedule independently of report email. Operational use requires both — a Monday-morning dashboard email is useless if the underlying data hasn’t refreshed since Friday.
Common Failure Modes
Report performance dies when you build a 2M-row tabular report with no filters; use sampling or aggregations instead. Workspace sprawl (50+ workspaces, half stale) makes governance impossible — quarterly cleanup with a deprecation workflow keeps it manageable. Permission inheritance is non-obvious: a user with View on a workspace inherits View on all underlying tables, which can leak sensitive data. Audit shared dashboards monthly.
Cost Considerations
Pricing is per-user with row-count caps per workspace. The Basic plan caps at 0.5M rows per workspace, Standard at 1M, Premium at 5M, Enterprise at 50M. Compare against Tableau, Power BI, and Looker on three axes: per-user cost (Zoho roughly 30-50% cheaper), Zoho ecosystem integration (Zoho wins decisively), and advanced analytics (Power BI and Tableau still lead on geospatial and statistical functions).
What to do this week
Audit your top three workspaces for permission scope, schedule one operational dashboard with refresh aligned to its email cadence, and try one Ask Zia query against a live workspace to test natural-language understanding on your data shape.
