How Workhuman constructed multi-tenant self-service reporting utilizing Amazon Fast Sight embedded dashboards

This publish is cowritten with Ilija Subanovic and Michael Rice from Workhuman.

Workhuman’s customer support and analytics crew have been drowning in one-time reporting requests from seven million customers worldwide—a standard problem with legacy reporting instruments at scale. Enterprise intelligence (BI) admins confronted mounting stress as their groups grew to become overwhelmed with these requests. By rebuilding their analytics supply with Amazon Fast Sight dashboards, they eradicated the guide report era bottleneck for customer-specific necessities. With this transformation clients achieve {custom} reporting capabilities. Workhuman is a world chief in human capital administration (HCM) software program and focuses on worker recognition and engagement options. By utilizing Workhuman options, staff can acknowledge and reward one another, fostering genuine human connections within the office.

This publish explores how Workhuman reworked their analytics supply mannequin and the important thing classes realized from their implementation. We undergo their structure strategy, implementation technique, and the enterprise outcomes they achieved—offering you with a sensible blueprint for including embedded analytics to your individual software program as a service (SaaS) functions.

Workhuman delivers SaaS capabilities for social recognition, steady efficiency administration, and worker expertise analytics to enterprise purchasers. With twin headquarters in Dublin, Eire, and Framingham, Massachusetts, Workhuman serves over seven million customers throughout 180 nations, supporting a couple of million recognition moments month-to-month.

The enterprise problem

As Workhuman scaled to serve customers worldwide with legacy reporting instruments, their customer support and analytics groups grew to become overwhelmed by an unsustainable quantity of guide, one-time reporting requests. This reactive mannequin created a number of essential points:

• Useful resource constraints: Handbook report era consumed vital crew time, resulting in delays in knowledge supply and elevated operational prices. Every request for a {custom} report required developer involvement, making a bottleneck that slowed down Workhuman’s means to serve clients successfully.
• Restricted flexibility: Stories delivered to clients weren’t customizable to their particular wants. Modifications required further improvement assets, restarting the cycle.
• Lack of self-service: Clients couldn’t independently discover and visualize their very own knowledge. The dependency on inside groups created friction, diminished agility, and impacted buyer satisfaction.
• Entry management gaps: With out a sturdy mechanism for administering reporting entry or managing knowledge privileges securely, Workhuman confronted each safety dangers and operational complexity as their buyer base grew.

Workhuman wanted to construct an answer to unravel their distinctive must ship reporting at scale whereas empowering clients to handle it themselves.

Answer overview

Workhuman wanted to offer intuitive reporting experiences to program managers, HR professionals, and folks leaders so they might create {custom} visualizations as wanted instantly within the reporting product—all whereas respecting HR directors’ must implement granular reporting authorization privileges and preserve personalised dashboard entry primarily based on every consumer’s authorization degree.

Workhuman developed a complete self-service analytics platform that addresses the distinctive challenges of multi-tenant SaaS environments with structure patterns that preserve strict knowledge isolation throughout clients whereas maximizing useful resource effectivity. The answer was designed to empower each Workhuman’s inside and buyer customers to independently discover, analyze and visualize their curated recognition datasets for numerous insights. A key power of Workhuman’s implementation is the strategic use of Fast Sight embedded dashboards in current functions, adopted by automation approaches that scale analytics deployment throughout a complete buyer base with out guide intervention.

The answer covers the row-level safety methods used to undertake fine-grained entry management inside every tenant’s surroundings, complemented by steady integration and steady supply (CI/CD) practices for managing analytics belongings throughout improvement, staging, and manufacturing environments. Moreover, the answer showcases the real-world enterprise outcomes Workhuman achieved, together with diminished operational overhead and improved buyer satisfaction metrics that validate the funding in self-service analytics.

Workhuman chosen Fast Sight for its multi-tenancy and asset isolation options that instantly addressed their challenges:

• Multi-tenant structure: The namespace performance in Fast Sight creates logical isolation for every buyer group, offering strict separation of analytics belongings on the namespace degree and consumer administration whereas sustaining strict knowledge boundaries—a foundational requirement for any multi-tenant SaaS resolution.
• Embedding capabilities: The embedding SDK and API-first strategy let Workhuman management the consumer expertise whereas tapping into Fast Sight analytics. Growth groups can customise the appear and feel to match their software’s branding and consumer interface sample.
• Row-level safety: Entry controls assist customers solely see knowledge they’re approved to view primarily based on their position within the group.
• API-driven automation: Fast Sight APIs allow programmatic administration of analytics deployment points, supporting environment friendly administration of analytics belongings throughout a complete buyer base with out guide intervention for every buyer.

Structure overview

A multi-tenant analytics resolution requires a number of key parts working collectively to offer a safe, segregated analytics expertise. Workhuman’s structure orchestrates these parts by a fastidiously designed workflow that balances isolation with operational effectivity. The structure makes use of devoted namespaces for every buyer whereas utilizing shared infrastructure and templates to cut back complexity.

Workhuman structured their strategy round these parts:

• Admin Hub and Reporting: The central software gives embedded Fast Sight dashboards and manages consumer entry
• Fast Sight namespace administration: Devoted namespaces for every tenant preserve isolation
• Evaluation creation and administration: Grasp templates and customer-specific analyses
• Dashboard publication: Analyzes revealed as dashboards for buyer consumer consumption and customization by authoring expertise.
• Dataset administration: Datasets with applicable customer-specific filtering
• Dashboard embedding: Safe embedding URLs built-in into the appliance
• Dashboard authoring: Customers can create personalized copies of current analyses

Workhuman’s AWS account serves because the central hub, housing the default namespace for inside operations and templated belongings with predefined datasets and evaluation instruments. Throughout onboarding, every buyer receives devoted segments that handle their particular belongings, together with filtered datasets tailor-made to their distinctive knowledge necessities.

The Admin Hub and Reporting software handles consumer administration, authentication, and authorization, interacting with templated belongings to publish dashboards and analyses utilizing an API. An Amazon Aurora PostgreSQL database helps backend operations, storing and managing buyer knowledge securely.

The structure confirms that every buyer operates inside their very own remoted surroundings, with devoted assets and knowledge entry controls, whereas utilizing shared infrastructure and instruments for effectivity and cost-effectiveness.

Workhuman developed this structured workflow, proven within the previous determine:

1. Handle consumer identities in Fast Sight: Create every buyer authoring consumer inside a Fast Sight group in a {custom} namespace
2. Create grasp analyses: Analysts construct grasp analyses for related product characteristic areas primarily based on core mannequin datasets
3. Duplicate for purchasers: Copy analyses for every buyer and share them with related buyer Fast Sight teams
4. Create filtered datasets: Construct customer-specific datasets with customer-specific filters and share them with Fast Sight buyer authoring teams
5. Replace analyses: Modify customer-specific analyses to make use of datasets with customer-specific filters
6. Publish dashboards: Generate buyer dashboards from the analyses with up to date datasets
7. Handle authorization: The applying layer manages consumer configuration and performs authorization checks, provisioning customers by way of API when approved
8. Generate embedded URLs: Fast Sight generates authoring embedded URLs for customer-specific analyses
9. Render in UI: The embedded URL renders within the consumer interface

Technical implementation

Workhuman’s implementation makes use of three core Fast Sight options: namespace isolation for tenant separation, template-based customization to keep up consistency, and row-level safety entry management. Every element builds on the architectural basis described earlier, working collectively to create an analytics platform that scales effectively whereas sustaining strict safety boundaries.

Namespace isolation

Every buyer group receives a devoted namespace in Fast Sight Enterprise Version. Every namespace incorporates one tenant’s assets, stopping clients from accessing one another’s knowledge or analytics. Namespaces present the foundational layer of consumer isolation required for multi-tenant SaaS functions, offering logical separation of every buyer group’s customers, belongings, and knowledge stay logically separated with boundaries that Fast Sight enforces mechanically

Grasp templates and customization

Grasp evaluation templates embrace customary KPIs, visualizations, branding, and placeholder filters. Throughout buyer onboarding, automation generates customer-specific variations from these templates, and the deployment automation framework then mechanically generates Fast Sight belongings, creating consistency whereas lowering guide effort.

Row-level safety

Row-level safety (RLS) restricts knowledge entry inside every buyer’s namespace primarily based on consumer roles. RLS guidelines filter knowledge utilizing column values that match consumer attributes. Row-level safety enhances namespace isolation by limiting knowledge visibility inside every buyer’s surroundings primarily based on consumer roles and permissions.

Safe dashboard embedding

Dashboard embedding generates safe, time-limited URLs for every consumer session. Workhuman personalized the interactivity choices and built-in with their current authentication system.

Dashboard authoring

Clients customers with authoring expertise can create personalized variations of embedded Fast Sight analyses by a custom-developed course of:

1. Create a duplicate of the unique evaluation utilizing the Fast Sight API
2. Outline RLS guidelines for the evaluation
3. Create a Fast Sight group related to the given RLS permissions utilizing the API
4. Add customers to the Fast Sight group

The method verifies that every {custom} evaluation has related permissions and teams, permitting customers to belong to completely different Fast Sight teams with completely different RLS permission units for various analyses.The next picture reveals the Reporting Admin house web page itemizing all obtainable analyses.

To create a {custom} evaluation, buyer customers enter the evaluation identify and choose the evaluation kind within the dialog field. They then select a pre-created evaluation to make use of as a template.

Static and dynamic asset administration

Workhuman’s implementation distinguishes between two forms of Fast Sight belongings, every managed in another way primarily based on their lifecycle and replace frequency.

• Static belongings: Namespaces, customary analyses, dashboards, folders, and creator teams keep comparatively secure throughout clients. Python scripts utilizing the Fast Sight API generate these belongings mechanically. The automation screens the standing of Fast Sight belongings for every consumer, validates evaluation definitions, detects supply definition updates, and mechanically provisions belongings upon detection of latest consumer IDs. Shopper report knowledge comes from a centralized Amazon Redshift desk. Throughout Fast Sight dataset creation, the system applies client-specific filters primarily based on consumer ID to keep up knowledge isolation.
• Dynamic asset administration handles assets related to Fast Sight evaluation authoring experiences—{custom} analyses created by customers and the related permissions. The frontend software makes use of the Fast Sight API, in order that customers can create new analyses and handle their lifecycle. A PostgreSQL database shops RLS permission units and evaluation metadata.

Throughout {custom} evaluation creation, the UI dialog lists obtainable analyses for personalization. Customers outline RLS permissions derived from obtainable columns and values within the dataset. Permissions are inserted into the RLS database and linked to the dataset. The evaluation is created, and a brand new Fast Sight group is created.

The separation between static and dynamic belongings signifies that Workhuman can preserve centralized management over customary templates whereas offering flexibility for customer-specific customizations.

CI/CD pipeline

The deployment pipeline automates Fast Sight asset updates throughout three phases. The three-stage strategy balances improvement agility with manufacturing stability:

1. Growth: Analysts create and take a look at new dashboard templates in an remoted surroundings
2. Staging: Templates bear testing with production-like knowledge to determine points earlier than buyer publicity
3. Manufacturing: Validated templates deploy to buyer namespaces

The pipeline contains approval gates between phases and rollback capabilities in case points are detected. Terraform deploys infrastructure, whereas AWS Lambda capabilities and AWS Batch processes execute Fast Sight asset creation automation.

The next picture is the CI/CD pipeline workflow. The structure demonstrates an automatic deployment workflow that integrates GitLab model management with AWS companies similar to AWS Batch, Amazon Fast Sight, Amazon Lambda, and Amazon Aurora to handle analytics dashboards and reporting.

Knowledge sources

Amazon Redshift serves as the first knowledge supply for experiences. Devoted extract, rework, and cargo (ETL) workflows create underlying Amazon Redshift tables. To take care of present knowledge, Fast Sight dataset refreshes set off mechanically following Redshift desk updates utilizing the refresh dataset API. Amazon CloudWatch metrics observe refresh timestamps, knowledge row counts, and processing length. Dashboards and alerting mechanisms monitor dataset freshness and assist confirm knowledge reliability.

Buyer-facing dashboards

The Workhuman dashboards powered by Fast Sight present clients with actionable insights from worker recognition knowledge. These dashboards show the forms of analytics Workhuman’s clients can entry and customise:

Award Distribution dashboard: The dashboard, proven within the following screenshot, tracks recognition attain throughout the group. The metric exhibiting 29.3% of awards given to people in numerous departments highlights recognition’s position in fostering cross-functional collaboration. Organizations can use these metrics to determine gaps in recognition protection and observe program effectiveness over time.

Government Insights dashboard: The view, proven within the following screenshot, focuses on inter-departmental recognition patterns. Within the instance proven, 89.29% of staff have acquired recognition, indicating sturdy program adoption and reveals departments like Operations, Buyer Excellence, and Expertise actively receiving awards from exterior their groups, indicating wholesome cross-functional appreciation. Executives use this dashboard to evaluate organizational tradition well being and determine departments that will want encouragement to take part extra actively in recognition packages.

Recognition (Recipient) dashboard: The evaluation identifies which awards and staff most affect firm tradition. The instance, proven within the following screenshot, highlights non-managerial staff as vital contributors to recognition tradition. The perception helps organizations perceive that culture-building isn’t restricted to management roles and might inform recognition program design.

Recognition per Worker dashboard: This dashboard analyzes recognition exercise by worker segments together with nation, division, and managerial standing. Organizations use this view to determine and tackle disparities in recognition distribution, confirming equitable program participation throughout completely different worker populations. Be aware: Foreign money is USD for quantities displayed within the dashboard examples.

Advantages and outcomes

Workhuman’s self-service platform reduce {custom} reporting requests dramatically:

• Decreased {custom} reporting requests: Clients now create and modify their very own experiences, releasing up the event crew for different work.
• Improved buyer satisfaction: Self-service analytics capabilities acquired constructive suggestions from Workhuman’s clients, who recognize the flexibleness and management. Empowering clients with self-service instruments improved their general expertise with the product.
• Quicker time-to-insight: Customers can entry and analyze knowledge instantly quite than ready for {custom} experiences, altering how clients work together with their knowledge and make choices.
• Scalable resolution: The multi-tenant structure helps a rising buyer base with out requiring proportional will increase in improvement assets. As new clients be a part of, automated provisioning handles the complexity.
• Decreased improvement time: The automated CI/CD pipeline deploys new analytics options extra shortly. What beforehand took weeks of {custom} improvement can now be rolled out throughout the shopper base in days.

The answer additionally freed up improvement assets that have been beforehand devoted to fulfilling {custom} reporting requests, permitting Workhuman to deal with core product innovation.

Seeking to the long run

Workhuman plans so as to add extra dashboard customization choices, new visualization sorts for his or her particular use circumstances, chat brokers, Pixel Good Stories, deriving insights from unstructured knowledge and expanded API capabilities.

Conclusion

Workhuman’s implementation of Amazon Fast Sight demonstrates how organizations can ship highly effective self-service reporting capabilities whereas sustaining strict multi-tenant knowledge isolation in SaaS functions. By utilizing Fast Sight enterprise options, groups can create scalable options that enhance buyer satisfaction and scale back improvement overhead.

Key takeaways

Apply these key classes from Workhuman’s expertise:

Use namespaces to separate tenants: That is foundational for multi-tenant analytics.

• Implement row-level safety for knowledge governance: Add row-level safety for extra knowledge governance inside namespaces, serving to to confirm that customers see solely their approved knowledge.
• Create grasp templates for consistency: Construct grasp templates to keep up consistency whereas permitting customization, balancing standardization with flexibility.
• Automate deployment by CI/CD: Implement a sturdy CI/CD pipeline to automate deployment and testing. Automation is essential for managing analytics belongings at scale throughout a number of clients.
• Use embedding capabilities for native expertise: Reap the benefits of the embedding capabilities offered by QuickSight to combine analytics naturally into functions, making a cohesive consumer expertise.

Study extra

To be taught extra about implementing embedded analytics with Fast Sight:

In regards to the authors

Kanniah Vagathupatti Jaikumar is a Senior Options Architect at AWS on the UKI ISV Answer Structure crew. He helps clients design and construct cloud workloads following AWS finest practices, with a deal with resilience, operational excellence, and cyber resilience. Kanniah collaborated with Workhuman to architect a scalable analytics service utilizing Amazon QuickSight, enabling embedded analytics for his or her multi-tenant SaaS platform. Outdoors of labor, he enjoys journey and experimenting with new cuisines.

Michael Rice is the Director of Knowledge Platform Engineering at Workhuman. He led the implementation of embedded analytics utilizing Amazon Fast Sight and continues to champion knowledge engineering rules and finest practices throughout Workhuman’s knowledge and analytics surroundings.

Ilija Subanovic is a Principal Engineer at Workhuman. He performed a key position in implementing and testing the embedded analytics resolution utilizing Amazon Fast Sight and different AWS companies, guaranteeing profitable supply of product options for Workhuman’s knowledge and analytics capabilities.