SHERP for Yalantis Scaling:Cultivating a Data-Driven Decision-Making Culture with SHERP

Amid rapid growth in 2020, Yalantis encountered operational challenges due to using too many disparate tools for project management and manual HR processes. The lack of a single source of truth, data fragmentation, and significant manual labour costs slowed decision-making and impeded scaling.

The SHERP product ecosystem was developed to unify data on a single platform and automate processes. SHERP was designed to integrate with all internal systems, generating real-time data with necessary visualization. This all-in-one BI solution would become the new engine for Yalantis’s growth.

---

Pre-Implementation Context: A Cycle of Growth Constraints

Technology Zoo:
In 2020, Yalantis used up to five different tools for managing HR processes and time tracking. This included third-party solutions for resource utilization and timekeeping, along with Google Sheets with varying levels of accountability and inconsistent data.

Lack of a Single Source of Truth:
The data was disintegrated and inconsistent, with information about projects, resources, and workloads fragmented across different files. This made it impossible to operate with unified, consistent data for decision-making. For instance, a single project could have up to three different data sources, each with controversial data about specialist workloads.

High Costs Due to Manual Workflows:
Teams spent over 25% of their time weekly reconciling reports and handling approvals manually, increasing the risk of errors and lost productivity.

Impacts of Inaccurate Skill Assessments and Management Visibility:
Inaccurate assessment of employee skills increased the risk of project delays and led to time lost on additional training. When technical staff were not properly matched to their roles, it lowered the quality of work and heightened risks. Junior employees did not receive the necessary support for rapid development, and the absence of a competency benchmark matrix resulted in subjective evaluations of senior specialists’ abilities, potentially leading to demotivation among the staff.

Challenges with Performance Review Transparency:
The absence of a standardized approach to personnel management and performance reviews led to inconsistent interview practices across different managers. Without a unified framework, interviewers conducted reviews in varying ways, preventing the collection of consistent and actionable feedback for specialists. This hindered the ability to accurately assess employee performance and provide meaningful development insights.

---

Goals and Project Definition

• Establish a Unified Platform: Establish a single platform for managing projects, finances, human resources, and HR processes that would reduce manual work and provide real-time access to up-to-date data.
  
• Automate Processes: Implement automated solutions for resource planning, time tracking, leave management, and personnel utilization to improve workflow transparency and accuracy.
  
• Automate Competency Management: Develop a competency management system that enables performance assessments, creates individual development plans (PDP/IDP), and provides a transparent picture for salary reviews and career development.
  
• Reduce Dependency on External Tools: Many tools do not integrate seamlessly. Additionally, when data is distributed across various tools, there is an increased risk of security issues and potential data leaks.

Development and Integration Phases

Stage 1 – Data Pipeline Automation

Our data engineers created well-governed data pipelines that securely and efficiently move data to its destination to drive analytics and decision-making.

They did this by:

• orchestrating data workflows with Apache Airflow to schedule and automate extract, transform, and load (ETL) tasks to process data from internal systems and manual files
• cleansing, shaping, and validating data using Python packages (Pandas) for efficient data manipulation and transformation
• preparing quality data for analysis with Tableau Prep’s visualization tools
• extracting data from QuickBooks and Pipedrive (which Tableau cannot connect to) using APIs and database connectors
• loading processed data into the centralized data storage (Yalantis DWH) powered by PostgreSQL to provide scalable SQL-based data warehousing and analytics capabilities
• monitoring pipelines to quickly identify and resolve any data flow failures, ensuring reliable and accurate data delivery

Stage 2 – Data Consolidation and Optimization

Data engineers worked with business analysts to restructure the data foundation and optimize methods for entering and retrieving data across internal CEP, HRM, and other systems.

They did this by:

• creating a centralized data dictionary that documents all data elements, definitions, structures, owners, etc. in one unified metadata repository.
• reducing data duplication by designating the data dictionary as the single source of truth, instead of having separate definitions across systems.
• making internal systems independent to save employees’ time. For example, the HRM system no longer has to query the CEP or ERP system for employee data. It stores relevant fields locally as numerical IDs that link back to the dictionary, reducing data redundancy.
• optimizing data access for quicker system’s operation. Yalantis engineers consolidated all databases, creating a centralized database for internal systems. Now, changes are instantly reflected across systems. For instance, when an employee’s grade is updated in CEP, that information now automatically flows to ERP, HRM, and FinOps.

Stage 3 – Data-Driven Insights

To enable comprehensive data-driven insights, Yalantis management systematically gathers and processes data from across the organization.

This involves:

• leveraging ERP integrations with channels like Jira, Slack, and email to get a convenient consolidated view of operations and processes.
• uncovering correlations between parameters with multivariate analysis that would be impossible to see in isolated data sets: for example, analysing the relationship between assignment duration and task order processing to improve resource management and inform staffing decisions.
• discovering correlations, such as noticing that staffing without proper task order processing negatively impacts key metrics like project success rate and gross profit margin.
• identifying issues that may be missed from a single lens with holistic analysis. For instance, spotting that department X lacks individual development plans (IDPs) by combining perspectives from HR, mentors, and the department itself.

Stage 4 — Building Operational Interfaces for SHERP (ERP, HRM, CEP, Payments)

At this stage, we focused on developing interfaces for various operational modules of SHERP, such as ERP, HRM, CEP (Competence Evaluation Platform), and the Payments module. This process involved integrating with existing tools and configuring interfaces to ensure convenient access to information and the automation of key business processes.

---

Tools used:

Integrates with	Technologies:
Slack, Email, Tableau	Golang, React, AWS Hosting

---

Result after launched:	Business value:
The only one software — all-in-one system (ERP, CEP and HRM);	60% monthly saving without third-party applications;
No more using Google Sheets, Excel and manual analytics;	25% time saving without manual work;
Single Source of Truth (Data Lake);	+15-20% – Revenue increase;
Institutionalized and autonomy salary management;	+ 8-12% – Resource Utilization (~$40-50k);
Adaptation of software to the process demands;	+10-15% – Revenue Utilization (~$40-50k);

---