Skip to main content
Join our Live Demos, Every Thursday. Register Now

What this blog covers:

  • Most data-driven organizations are migrating their FP&A functions from legacy systems to modern solutions for several reasons.
  • Cloud-scale data and cloud migration come with many challenges too.
  • Kyvos solves these challenges with a modernized approach based on its patented smart pre-aggregation technology.
Listen 09:04

Data in any domain is growing by the order of magnitude; Financial Planning and Analysis (FP&A) is no exception. The convergence of digital transformation and evolving business models bring in such volumes of data that most organizations are not yet equipped to handle.

With 61% of FP&A leaders reporting inadequate tools and systems as their biggest challenge, the problem is bigger and more complex than it seems. In addition, most financial institutions have not even accomplished predictive or prescriptive analytics yet. This goes on to emphasize that now is the time to unlock the full scope and potential of the FP&A function to deliver higher business value.

The first step is migrating data analytics from legacy systems to the cloud for better flexibility, cost optimization and faster query performance. But moving to the cloud also has many challenges, particularly due to increasing information overloads and query complexities.

Cloud-Scale Data and Analytical Challenges It Triggers

Modern data warehouses can handle huge datasets, even when answering analytical queries. However, with billions or trillions of transactional records inundating these warehouses, queries asked by FP&A users may sometimes encompass all these transactions at once.

Every time someone fires a query, the cloud data warehouse (CDW) will move massive amounts of data to a compute cluster which costs much more than its basic storage fees. Latency in query responses means lost opportunities in hyper-competitive markets. What if an organization fails to identify a looming liquidity risk due to delayed insights into its cash flows or customer accounts?

Let’s also agree that most queries across the enterprise still originate from information workers (not just data scientists or technical folks). For example, queries like – What is the expected profit margin for the next quarter by business unit or the expected return on investment for a new product in Region A? The results for such queries may involve pre-processing billions or trillions of rows of data, and some of these questions may be asked by many concurrent users over time using common dashboards.

CDWs like Snowflake or Azure use caching mechanisms to conserve the processing results for the same set of frequently used data and commonly asked queries. But the system is still rigged with challenges, such as:

  • Uncached data goes through a heavy processing stage to read the information and ends up delivering slow performance.
  • The cloud platform may not be built to handle the complexities, hierarchies, and modeling methods integrated with the legacy environment.
  • Ad hoc analytics are critical to FP&A functions, especially for predictive and prescriptive analytics. Answering questions – like what will be the expected revenue from a specific region or which loans might turn bad – need more than just cached information.
  • Rules for manual caching are cumbersome and automated management can’t rely on simplistically implemented models.

Enter Modern OLAP with Patented Smart Pre-Aggregation Technology

More than ever before, FP&A teams now need solutions that allow them to change financial plans or readjust budgets in real-time to ensure faster decision-making across business units. Modernized OLAP systems leverage a distributed architecture to accelerate query performance in an unintrusive manner while offering multidimensional insights to end users.

Here’s what the best solution for them brings to the table:

Built for the Cloud

TM1, Essbase, and SSAS – the legacy trio – work with an old scale-up architecture that needs to be installed on a costly server to manage growing capacity requirements. Unlike cloud-native platforms, they lack the ability to scale up and down with changing business needs and the number of users.

Kyvos harnesses a scale-out architecture for OLAP implementation on the cloud, meaning the analytics can scale limitlessly along with the CDWs. By supporting parallelism, the platform can manage data modeling and querying without any constraints of scale, complexity and costs.

Using a cloud data platform’s computing capacity, Kyvos can pre-process massive data volumes compared to legacy systems limited by their memory- one of the key TM1 limitations.

FP&A Model Consumption with a Universal Semantic Layer

A user-friendly semantic layer for decision-makers or information workers will be optimized in the backend to deliver consistent and high-performance query responses. This layer works as a bridge between users and underlying data platforms/sources, enabling an intuitive, secure and accessible interface for analytics.

At the same time, the ability to deliver sub-second query results and create user-friendly interfaces are connected to each other. A multidimensional data model designed by Kyvos can form the basic structure of a semantic layer to facilitate smart pre-aggregation while presenting data in a unified manner. The layer offers a single source of truth for every user and allows self-serve analytics for all, despite their technical expertise and skills.

In addition to these, a universal semantic layer simplifies complex data modeling with features like alternate hierarchies which offer multiple paths to roll up represented data while still preserving the integrity of the aggregations. With custom roll-ups, users can also define member values rolling up into parent values in a parent-child hierarchy.

ML-Based Aggregations and Intelligent Caching

Simple aggregation tables or a few SQL GROUP BY materialized views can’t tap into the full scope of OLAP. The complexity of aggregations brings several challenges due to countless attribute combinations (features, table columns) adding to the impracticality of processing and storing all these aggregations.

Even a relatively small data warehouse with not more than 10 customer attributes, 10 product attributes, 10 location attributes, 5 persona attributes, and 5 date levels can result in a staggering number of 25,000 combinations.

Kyvos mitigates this issue with a patented smart pre-aggregation that employs ML-based aggregation strategy. The system analyzes queries to select the most valuable aggregations that fit within the given constraints of finite time, computation, cost and storage. Leveraging this approach, the platform enables efficient computation for detailed financial planning and analytics while optimizing storage utilization and managing costs.

Kyvos’ modern pre-aggregation technology automatically pre-processes the data by anticipating user needs and adjusting pre-processed data models based on changing conditions. This intelligent approach calculates results proactively even before a user requests them. Various levels of caching and extensive calculations simplify the process even further.

Looking Ahead

From dozens to hundreds of users across organizations, Kyvos’ cloud-based analytics platform empowers FP&A teams to explore new possibilities with deeper analytics through a patented technology backed by a universal semantic layer and distributed architecture.

Kyvos Smart Pre-aggregation encapsulates the fundamental procedure for delivering data in an accelerated and user-friendly way.

Request Demo


What are the uses and benefits of a modern OLAP system?

A modern OLAP system enables users to perform advanced analytics on large-volume, multidimensional datasets. A platform like Kyvos accelerates query performance and supports parallelism for building data models, enabling businesses to deal with the scale and complexity of the modern data ecosystem while minimizing costs. These multidimensional data models based on AI-based smart aggregation technology also help create a smart semantic layer for presenting enterprise-scale data in a unified manner.

Why do we need OLAP on Cloud?

The scale-up architecture of legacy OLAP systems like TM1, SSAS or Essbase works on a costly server and is unable to scale up or down with changing needs. Modern OLAP on the cloud has a distributed architecture to manage data modeling and querying at any scale, even for complex datasets.

How can modern OLAP solutions improve FP&A performance?

Modern OLAP solutions like Kyvos work on a distributed architecture that accelerates query performance and offers multidimensional analytics on the cloud. A user-friendly semantic layer simplifies complex data modeling with custom roll-ups, alternate hierarchies, and many more features. 

What is OLAP in cloud computing?

OLAP in cloud computing is a technique to organize data into multidimensional models to get faster insights, slice-dice into any granularity and analyze large-scale data from multiple angles. The cloud-native approach leverages the flexible computing capacity of modern data platforms to build highly optimized data models across multiple dimensions.

Close Menu