When Bad Data costs Millions: How an Indian Data Engineer helps Companies Streamline their Analytics

Companies face a lot of challenges due to bad and duplicate data. Indian data engineer Abhishek Anand has an architectural solution that has already cut Grubhub's routine analytics by 50% and he is focused on changing the whole industry of data analytics.

According to analysis from Dataversity, bad data costs companies an estimated $12.9 million each year. Approximately 70% of organizations experience bad data, duplications and inconsistencies in metrics, typically in the field of business analytics. Recently, HR-tech company PayFit discovered that up to 30% of records in their CRM department were duplicates. This resulted in wasted time and duplicate customer calls, along with decreasing their marketing effectiveness.

A KPI mistake on an old metric, a report that does not align with another report because a different SQL could be used, or two departments that define an "active user" differently: all of these examples therefore slow down analysts work, confuse strategic decision-making, and ultimately make the company less competitive.

While large companies spend millions of dollars trying to solve these problems, there are new solutions coming to market that allow companies to automate and seamlessly manage their metrics. An example of this is the work of Abhishek Anand, data engineer from India, who has reduced routine analytics for Grubhub, an American foodtech company, by 50%.

How to get rid of half of the routine operations

Abhishek Anand, who is a Data Engineer II of a large American food delivery company, also faced the issue of data duplication. Grubhub is a digital food delivery platform in the USA, whose primary project is the Metrics Layer Framework. Abhishek Anand has deployed an internal framework here that standardises access to metrics. The Metrics Layer Framework is a system within the company that allows all departments to use the same definitions for metrics, such as "active user", "successful delivery", or "cancelled order".

"Previously, engineers would have to manually compare formulas and make edits to reports. This would take hours or days and could result in data duplication. With the Metrics Layer Framework, everything is stored in a single database where each metric is defined once and can be accessed by anyone," Abhishek explains how this system works.

This is an internal tool that is meant to standardise metrics between all the company's teams, and eventually reduce the redundancy by 40-50%, from corrections in the code to approvals between departments. Also, Anand is undertaking considerable development efforts in building the AI-Powered Data Lineage Assistant, which is an intelligent assistant leveraging a graph database and LLMs to help track where data originates and its dependencies. Ultimately, the assistant will help engineers decide whether changing a metric will affect anything, utterly independent of searching through code or asking their team for multiple hours.

"This system has already become a reference for other departments within the company, not only for engineers, and is used as a basis for building new Grubhub's analytical solutions," Anand shares.

At this time, the Metrics Layer platform is becoming the only source of reliability for establishing data in the Grubhub architecture. The implementation of this technology allows this large company to avoid losing millions of dollars and customers, as well as maximise the amount of usability for the entire team. Anand has developed a centralized and reusable way to define and maintain business metrics that could be used by many companies outside of Grubhub. The hope is to run a lot of it at the same time, and reuse most of it in other reporting tools, and save very large dollars of development time, to make it so much easier.

Shaping Global Digital Data

Abhishek Anand's projects address a universal issue: disparate metrics, repetitive tasks, and the absence of a unified system. Grubhub has saved resources, accelerated processes, and, most importantly, implemented a unified analytics architecture. These are big for the data analytics market because any large company wants to get rid of at least the bad or duplicate data.

Prior to Grubhub, Anand had a career in the data industry and aided a company that he had previously helped improve its performance. At Meta, a major American technology company that owns popular services such as Facebook, Instagram, WhatsApp and Oculus, Abhishek was a Data Engineer III, built a system that automatically tracked where and how user data was being used and reported any violations in a timely manner. This allowed for quick problem resolution and reassurance to users that their information was secure. The increased trust from the audience was one of the reasons why the product quality improved and, consequently, advertising revenue increased by 25%.

Abhishek Anand's work has already been recognised internationally. He is a jury member for the annual international award that celebrates the best achievements and innovations in the field of technology, 2025 Globee® Awards for Technology and is an honorary member of the International Association of IT Professionals. This certainly recognises Anand as a developer, but even more importantly, it recognises him as a thought leader who is shaping the discussion on trends in data architecture.