Sansan Corporation

FlyData Autoload contributed to a dramatic improvement in business speed.

Reduced time to collect data and issue KPI reports from 10.5 hours down to a minute. Contributed to quicker turnarounds in the service improvement cycle.

Today, Sansan Corporation: Eight Division gave us a discussion about how they started using FlyData Autoload.

Sansan Corporation

Mr. Shishikura, Production Development Vice President, Eight Division

Mr. Mase, Infrastructure Manager, Eight Division

Mr. Suzuki, Data Analysis Specialist, Eight Division

– Sansan Corporation Services

Sansan Corporation was founded in 2007, offering “Sansan”, an enterprise business card management service (originally called “Link Knowledge”). Since then, we have offered our services to more than 2000 companies in Japan.

In 2012, we added another service called “Eight”, a business card management service targeting individual customers. Eight, with the concept of “making the business card into a new business connection”, offers an app through which people can manage business cards easily. By taking a picture of a business card through your smartphone, the app will store contact information to your smartphone and make it accessible also through their website. Eight also provides a “Connect” feature, which allows users to connect to others they have exchanged business cards with in the past. Users can then message each other or get updates on each others’ statuses in real-time.

– Growing Pain-Points

Two years after having launched our service, our user base has now grown to more than 600,000 users with a total of more than 20 million business cards processed. Now, more than 100,000 business cards are added each day. To stably process the data, we redesigned our backend infrastructure to be scalable, using AWS.

At Eight, we are working to improve our service by extracting and analyzing our KPIs as daily reports. Before using AWS, we had been extracting and quantifying the number of new users and users’ information using RDB and creating reports based on data from it. But over time as the number of users grew, the amount of data we handled became larger and larger; to the point that we couldn’t query fast enough to issue our daily reports. It was a serious problem to our business since we could not access the information vital to our business decision-making.

– Research on Potential Solutions

Pushed by these circumstances, we started looking into possible solutions. We realized that we needed to introduce a solution quickly to avoid losing user trends and insight into user activity.

First we looked at Hadoop. Redshift was still a new service, so initially, we were just testing it out alongside our other tests to get an idea of how well it performs. That was when we got introduced to FlyData through AWS, and started working on a solution together.

The key factor in our decision to choose FlyData and Amazon Redshift was the benchmark we ran, which showed Redshift returning results much faster than Hadoop. Moreover, rather than modifying a query with Hadoop Hive, we could continue to use a lot of the standard SQL queries that we have built up when querying against Redshift because it is compatible with PostgreSQL. On top of that, by using FlyData, we found that we could make data transfer and loading to Redshift much smoother and easier.

With query performance better than Hadoop, as well as the flexibility provided by FlyData and Redshift, we decided to go ahead and test the entire implementation. As a result, we were able to implement our analytics infrastructure using FlyData and Redshift, without any technical difficulties.

– Noticeable Results

We have traditionally set KPIs such as conversion rates and duration of active usage, but after moving to Redshift and reports taking 1 minute to generate instead of 10.5 hours, we are now able to share KPI reports across the team every morning at a set time. Now that reports can be generated so quickly, there are people who request getting reports even faster!

So we started preparing an environment that allows us to capture our KPI in near real-time, utilizing the real-time continuous loading capabilities of FlyData. Today, we project our dashboard on a screen so that the whole company can share the same KPI at any time, and understand exactly how our service is performing.

Going forward, we plan to start implementing an even more seamless infrastructure by utilizing FlyData Sync.

FlyData handles real-time replication for Amazon RDS and Aurora, MySQL and PostgreSQL.

Get set up in minutes. Start uncovering data to make faster, better business decisions today.

By using this website you agree to accept our Privacy Policy and Terms & Conditions