Photo credit: venturebeat.com
Stay informed with the latest insights and exclusive content on industry-leading AI developments.
Running multiple business units that utilize various applications can pose significant challenges in data management. Navigating through this intricate landscape with data analytics often adds another layer of complexity.
As Broadcom has expanded through a series of acquisitions involving software and hardware companies, it has encountered the difficulties of merging diverse data ecosystems.
Notably, after acquiring VMware for $61 billion in late 2023, Broadcom faced an extensive data integration challenge. VMware maintained a network of 1,800 heterogeneous applications along with an impressive 187,000 product SKUs. The strategy typically employed by many companies in similar circumstances would include establishing master data management systems, constructing large-scale data lakes, and executing complex integrations to link the various systems over multiple years.
However, Broadcom chose a significantly different path. Rather than layering additional complexities onto their existing systems, they opted for a clean slate approach. The company consolidated numerous data analytics tools, settling on a single platform from Incorta, which is now utilized across all divisions, including VMware.
“VMware had a notable mix of applications,” said Alan Davidson, Broadcom’s CIO, in an exclusive interview with VentureBeat. “We actually deprecated every single application to unify under our singular source management plan for the data model.”
Davidson clarified that this data model serves as Broadcom’s primary point of reference for its data, with the objective of simplifying and consolidating the data landscape to enhance reporting, analytics, and decision-making.
Broadcom’s strategy emphasizes data cleansing at the source rather than at a later stage in the process, a principle that has been central to their integration efforts.
The technical implementation encompassed several critical elements:
ERP consolidation: VMware had relied on seven different enterprise resource planning (ERP) systems. Davidson humorously noted that they had a system for each day of the week. Broadcom streamlined this to a single ERP solution.
SKU rationalization: The vast number of SKUs at VMware, which totaled to 187,000, has now been drastically reduced to just 500. This simplification has resolved numerous issues pertaining to compatibility, pricing discrepancies, and support scenarios that would have otherwise required extensive mapping.
Master data restructuring: Rather than creating elaborate translation layers between different systems, Broadcom instituted a single master data framework. This provided one reliable source for customer information, product details, procurement data, and supplier records.
Contract and provisioning integration: A vital aspect was integrating contracts with orders down to serial numbers and provisioning settings. This ensures that when customers purchase software, there is a clear visibility from contract stipulations to actual usage and entitlements.
The outcome was a significantly streamlined architecture, eliminating the need for cumbersome data transformations.
Standardizing on data analytics to enable enterprise insights at scale
A consolidated and reliable data source is the cornerstone of Broadcom’s data operations.
To enhance this further, Davidson noted that the company has unified its data analytics approach on a single platform. Following a thorough evaluation of various options, Broadcom selected Incorta to manage analytics across its extensive data repository. Presently, more than 17,000 internal users leverage Incorta to analyze over 200 terabytes of operational data.
“When you look at competitors like Microsoft Power BI, Snowflake, or Tableau, there’s a multitude of user interfaces, each with its own challenges,” Davidson remarked. “There hasn’t really been a comprehensive end-to-end solution that offers a flexible interface, self-service capabilities for thousands of users, and scalability with diverse data mappings.”
Broadcom’s array of business units, each with distinct market strategies and data requirements, posed a considerable challenge. Davidson emphasized Incorta’s capacity to manage scale and connect divergent data sources. Incorta’s adaptability in data access segmentation and facilitation of self-service analytics has proven transformative.
“I need to enable segmentation and connect the pieces while managing user scale,” Davidson elaborated.
Incorta’s AI and automation capabilities expand data analytics
Broadcom is poised to take advantage of new AI and automation enhancements accompanying Incorta’s latest updates.
Osama Elkady, CEO and co-founder of Incorta, shared with VentureBeat that businesses constantly seek enhanced data accessibility. To respond to these demands, Incorta has launched Nexus, which provides organizations like Broadcom with generative AI functionalities.
Nexus enables users to utilize advanced natural language processing (NLP) for various tasks, including data cleansing, model development, and bridging machine learning with generative AI. The aim is to empower Broadcom users to swiftly and effortlessly access the data they need, eliminating the previous burdens associated with manual data preparation and complex dashboard configurations.
Applying AI with a data-first mindset
While many enterprises rush to implement generative AI without resolving foundational data quality issues, Broadcom opts for a more cautious strategy that prioritizes data integrity.
“Utilizing AI with inaccurate data is one of the worst mistakes we could make, especially concerning financials and telemetry, where accuracy is crucial,” Davidson stressed. “If you start with bad data, you’ll quickly arrive at incorrect conclusions.”
With Incorta’s Nexus, Broadcom is strategically utilizing AI where it can deliver clear benefits—primarily in democratizing access to information.
“The essence of our AI initiatives with Incorta is to ensure data transparency and ease of access without necessitating the effort of creating personal dashboards,” Davidson clarified.
What this means for enterprise data leaders
For organizations grappling with data complexity—stemming from acquisitions, organic growth, or digital transformation efforts—Broadcom’s approach offers valuable insights:
Question the need for complex data lakes: Before accepting data complexity as a given, consider if drastic simplification at the source could yield better results.
Standardization beats flexibility in core systems: Though making tough decisions may be necessary, having a singular record system leads to fewer integration challenges than juggling multiple platforms.
Self-service analytics requires guardrails: Broadcom’s success is attributed to empowering users within a carefully orchestrated framework rather than providing unbridled access that could lead to maintenance issues.
AI implementation requires data discipline first: Prior to rolling out advanced AI solutions, ensure your core data is accurate, accessible, and systematically organized.
In Davidson’s succinct reflection: “It’s challenging to simplify processes, but it’s all too easy to complicate them. Our goal is to undertake the difficult tasks and streamline them.”
In a technology environment crowded with tools aimed at addressing integration challenges, Broadcom’s commitment to rigorous simplification offers a compelling alternative for organizations navigating extensive mergers and acquisitions. By prioritizing foundational data cleaning, they have created a robust base that enhances rather than hinders future growth.
Source
venturebeat.com