How Visual Research reduced costs by 35 percent with the help of Google Cloud

How Visual Research reduced costs by 35 percent with the help of Google Cloud

Editor’s note: Visual Research, a Japanese company specializing in real estate solutions, offers systems that help real estate businesses and their customers execute rental and management processes. The company’s ability to modernize was being held back by increasing licensing costs, inefficient resource allocation, and performing manual maintenance tasks. By adopting Google Compute Engine, Cloud SQL for SQL Server, and Memorystore for Redis, Visual Research reduced its license costs by over 35%, eliminated manual tasks, improved performance, and gained the ability to scale to handle peak seasons. This ultimately enhanced the digital experience for its users and allowed Visual Research to focus its efforts on application development.


Visual Research was founded in 1995 with a goal to revitalize the real estate industry through platforms like rental management, rental agency, and buy-side and sell-side brokerage systems. These systems support everything real estate businesses need — from rental applications and vacancy searches to secure document storage and payments.

The Japanese real estate industry faces significant challenges due to its heavy reliance on paper-based processes and manual tasks, which have a profound impact on workflow efficiency and overall productivity. One noteworthy issue is the lack of standardization, as applicants and managers must navigate various portals and submission formats, each with their own unique processes. This results in unnecessary data duplication.

Adding to the complexity, the industry often prioritizes in-person or oral communication for sensitive and critical matters. As a result, historical documentation of conversations and exchanges in information is uncommon. In this context, real estate agents find themselves responsible for creating contracts, handling contract fees, and addressing inquiries across different real estate portals, further complicating their tasks.

From paper to enhanced digital experiences

Our goal was to enhance the efficiency of real estate management processes using IT solutions, shifting how users engage with digital experiences (DX). What we really wanted was to streamline account management and empower our users to prioritize customer engagement and service. We understood, however, that our existing hybrid infrastructure was curtailing our progress toward our goal.

In 2017, Visual Research adopted Google Cloud’s Compute Engine as part of our infrastructure, with the long-term goal of replacing our data center and Windows Server. Our previous setup also involved using SQL Server licenses for all our systems across data centers, which presented two primary challenges: First, the cost of licenses increased as our systems scaled up. Second, we had to estimate and size servers to accommodate peak loads, resulting in excess resources during off-peak periods. Given that the industry had two distinct peak periods — one monthly and one annually — resizing demanded a considerable amount of time and management effort, particularly since we lacked a dedicated infrastructure team. Ultimately, we were spending way too much time on tasks like server procurement, operational management, sizing assessments, and monitoring the condition and lifespan of server hardware and components.

Reducing costs by 35% with scalable, flexible, and high-speed databases

We chose Google Cloud as our preferred cloud provider for several reasons:

  1. Maintenance using live migration: Support for Live Migration in Compute Engine offered a seamless and efficient approach to system updates and maintenance tasks.
  2. Flexible resource changes: One of the primary factors that influenced our choice was the ability to make flexible resource changes, allowing us to adjust our computing resources as needed, with granularity down to 1 CPU and 1 GB of memory units.
  3. Enhanced storage performance: The platform’s high-speed disk capabilities met our requirements for storage performance.

In addition, the advantages we’ve experienced using Cloud SQL for SQL Server along with Memorystore for Redis are substantial:

  • Reduced license costs: By transitioning to Cloud SQL utilizing the SQL Server Web edition with high availability, we saw a cost reduction exceeding 35%. This stems from the fact that while the Web edition does not inherently support advanced features like mirroring or always on, Google Cloud compensates by offering high availability through its regional disks.
  • Reduced maintenance costs: The transition to Cloud SQL eliminated the need for manual tasks such as applying Windows updates and managing SQL Server patches, as well as storage management. This not only saved us approximately 4-8 hours per month but also brought peace of mind by avoiding potential blue screen issues following Windows updates. Additionally, the fixed cost structure for major database version updates, without the initial license purchase cost, has been another substantial benefit.
  • Enhanced visibility: We conducted performance measurements during the initial implementation that yielded remarkable results. Some queries demonstrated performance improvements of up to 20%, even under conditions initially perceived as performance disadvantages, and with the same or fewer resources compared to our previous Compute Engine setup. We addressed queries that initially exhibited performance degradation by performing a thorough review of statistics and indexes. We further improved efficiency with Memorystore for Redis as a caching layer, providing fast data access. Furthermore, the ability to scale up resources in anticipation of concentrated user access during specific dates has proven to be a valuable advantage.

These benefits are particularly significant for our team, as we lack platform engineers. This transition allows us to place a more substantial focus on our core objective — application development.

Reshaping the real estate industry with cutting-edge technology

At the heart of it, our goal is to simplify the entire process for real estate businesses and their customers. To push towards our mission, we sought flexibility and portability from Google Cloud databases, specifically Cloud SQL for SQL Server and Memorystore for Redis for rental management, and Cloud SQL for PostgreSQL and Memorystore for Redis for rental brokerage and sales. Combined, these database services help real estate agencies in Japan manage data more effectively during our peak seasons — from January to April, and then again in the summer. This ensures that agents can provide seamless services to their customers, handle increased demand, and make data-driven decisions to improve user experience.

Our recent experience with Cloud SQL for SQL Server has already yielded substantial cost reductions. Looking ahead, we eagerly anticipate faster backup capabilities to continue our journey of innovation and efficiency in the real estate industry.

Learn more: