Strikingly Case Study - Deliver a Data Product Easier and Faster

A Website Builder for Everyone

Strikingly is hatched by Y Combinator, a well-known incubator in the United States. It is committed to helping users with no experience to build a beautiful website or e-commerce platform quickly and easily. Business owners, designers, photographers, and students can easily run brands and showcase themselves through Strikingly.

It helps customers to create personal, small business, or small e-commerce websites within a few minutes and has served millions of users in 200 + countries and regions around the world. On some major social media platforms, there has been a trend of evaluating, trying, and writing reviews about Strikingly.

Besides, it’s recommended by marketing guru and best-selling author Seth Godin.

Challenges of SaaS Customer Operations in the Cloud

Born in Silicon Valley, Strikingly stands out from many SaaS service providers and is a typical practice of SaaS vendor on the cloud.

The scenarios and pain points behind Strikingly's business are very common. After users build their personal websites and e-commerce platforms in Strikingly, they can see the classic website traffic, user behavior and retention data of their personal websites, such as pageview (PV) and user visits (UV), through the Analytics Platform that comes with them in the background. Such a website traffic analysis scenario has relatively stable data analysis requests, but there are many challenges at the technical level.

First, as Strikingly has a large number of users aroung the globe, the data services it provides must be able to support high concurrency.

At the same time, the users of Strikingly are usually very sensitive about the query response time. Users simply cannot understand why a dashboard does NOT come out within one to two seconds. Therefore, the data team needs to ensure that all queries are answered within two seconds. This is a big challenge faced by many to-C SaaS companies when providing data services to customers.

In addition, after the users complete building the website, the Analytics Platform becomes a key touchpoint of Strikingly to improve user retention. Besides, as Strikingly's users are mainly non-technical people, they need simple and easy-to-use analysis tools that are highly integrated with their websites. However, third-party analysis tools are often very complex and difficult to learn. So users rely on the built-in Analytics Platform very much. In order to ensure user satisfaction and retention, Strikingly must ensure high stability and availability of data services .

After consideration and comparison, Strikingly chose Apache Kylin as its technical base in 2017 and has been using it for years. The open-source Apache Kylin has helped Strikingly to achieve optimal performance and concurrency, but it has not solved 100% of the problems that Strikingly had. After several years' practice of Hadoop + Apache Kylin architecture on the cloud, Strikingly still has the following pain points when providing the Analytics Platform data service.

1. Difficult Operation and Maintenance

• The traditional deployment method, Hadoop + Apache Kylin on the cloud architecture, is very complex, and the operation and maintenance is difficult
• Need to continue to provide services to users for 7 × 24 hours without interruption, causing greater pressure on the IT team

2. High Total Cost of Ownership (TCO)

• High resource costs for Hadoop components on the cloud
• The recruitment of big data engineer is difficult and costly, and the stability of employees is difficult to guarantee
• Every time when there is an outage, IT team needs to invest a large amount of time (sometimes two or three days) to solve the problems, which is costly in labor, and affects user experience of the Analytics Platform, and has a negative impact on business

3. In addition to meeting the two conditions of TCO and operation and maintenance, it is also necessary to ensure sub-second response capability and high concurrency capability

Stability, Simplicity, and Security – All From One Solution

After in-depth understanding of Strikingly's business scenarios and customized requirements, Kyligence provided Strikingly with Managed Service and Kyligence Cloud products to help Strikingly solve its challenges at the lowest TCO. The overall services provided by Kyligence to Strikingly can be summarized as follows:

• Google Analytics-like website traffic analytics
• Migrated from EMR + Kylin to Kyligence Cloud
• Reduced 35% TCO with better performance
• Managed Services offered by Kyligence to take over IT operation
• Same SLA as AWS Redshift

At the overall architecture level, with the cloud-native features of Kyligence Cloud, Strikingly transformed its original data service architecture from heavy and difficult to lightweight and efficient.

Through the comparison of before and after architecture, it is clear that Kyligence products brought the following changes to Strikingly's original data architecture:

• Leveraging the new generation cloud-native technologies, the architecture becomes simpler and lighter, and the operation and maintenance costs are greatly reduced
• Shorten the entire data process, timeliness and the value of data are better reflected
• Support high availability deployment of all components, ensuring the stability of data services
• Provide a unified service management platform and has 24-7-365 automated operation and maintenance monitoring capabilities

The Values of Kyligence Managed Services and Kyligence Cloud

• The stability of services has been greatly improved
  • The automated operation and maintenance monitoring capabilities provided by Kyligence Managed Service is able to better ensure the stability of the service. Since the customer went live by using Kyligence Cloud, there have been no problems with the service and no complaints from users.
• Lighter architecture and lower hardware cost
  • Using the cloud-native architecture of Kyligence Cloud, it is no longer necessary to rely on Hadoop, which not only makes the entire architecture more lightweight, reduces a lot of hardware costs, but also reduces a lot of operation and maintenance costs.
• Lower TCO and higher personnel efficiency
  • The TCO after using Kyligence Cloud is 35% lower than the traditional deployment (Hadoop on Cloud + Kylin).
  • The IT personnel who originally need to be engaged in complicated operation, maintenance and troubleshooting are freed from repetitive work and put their energy into more valuable work
• Query performance is nearly doubled
  • With the increasing amount of data, the query speed after using Kyligence Cloud is nearly double the query speed of the previous traditional deployment method (E.g.: Hadoop + Kylin on Cloud), providing the better user experience and indirectly improving the user retention and payment rate.