About the company
Anywayanyday is the largest online service for booking flights and hotels. Using Yandex Cloud to perform data analytics, the online travel service built their Business Intelligence (BI) platform on the Yandex Managed Service for ClickHouse and DataLens. This allowed the company to grant all employees access to monitoring and analysis of the indicators important to the business, which in turn accelerated decision-making in response to any deviations from target values. Moreover, the partnership with Yandex Cloud led to a reduction of 20% in the IT department’s labor costs associated with maintaining the reporting system.
Recognizing the need for a new approach
A trademark of the travel industry is its hyperdynamism: offers and prices change every minute. For companies to be able to adequately respond to this fast-changing market, modern data processing and visualization tools for key business indicators are must-haves.
Before transitioning to a single platform, Anywayanyday used a hodgepodge of unrelated tools:
- Web traffic data came from Google Analytics™ and Yandex.Metrica, was processed using BigQuery™, and visualized in Chartio.
- Sales data was gathered in a local Microsoft SQL Server™ database with pre-processing for SQL queries. For end users, data output was displayed in DB views, SharePoint forms, or exported to Excel.
- Metabase was used for product reporting. It was deployed on a local server in the office, but administrative and development costs meant that maintenance proved to be both complicated and expensive.
Anywayanydays' initial setup failed to provide employees the direct and quick access to data they needed in order to make decisions in real time. Moreover, the system was incapable of scaling quickly in response to increasing loads. The IT department’s workload was steadily increasing, leading other teams to have to spend time covering their tasks.
Anywayanyday decided to implement a single BI system, and came up with the following requirements:
- Server infrastructure should not require the purchase of equipment or support.
- The launch should take place as quickly as possible, without third-party consultants.
- The analytics team should take care of the setup using their own resources.
- Data from all sources should be available for download.
- The learning curve should not be an obstacle for employees.
- Users must be able to create and customize reports.
Taking all these requirements into account, AnyWayAnyDay came to the conclusion that a cloud-based BI was the best solution, and set out to formulate their plan.
To ensure the effective storage and analytical processing of data, while also reducing the IT department’s workload, Anywayanyday would use the Managed Service for ClickHouse in Yandex Cloud. As a managed solution, Yandex Cloud handles the deployment, maintenance, and integration with other services. To visualize data and build analytical dashboards, they would use DataLens, which is closely integrated with ClickHouse. The Anywayanyday management quickly accepted the proposal, and the implementation process was started.
Transition and training, all handled by one analyst
The switch to the new reporting system was complete in less than six months. A single analyst carried out the testing and transition, deployed the necessary services in Yandex Cloud, set up data uploads and the initial processing, and prepared a starter set of dashboards and reports. After that, he granted access to all relevant employees and trained them.
Commercial reporting and web traffic analysis were almost completely transferred to DataLens, which the sales and marketing departments now use.
Here is how data collection and processing is structured in DataLens:
- Data from Yandex.Metrica and AppMetrica are exported directly to ClickHouse.
- Sales data is collected in Microsoft SQL, then pre-processed by Python scripts and uploaded to ClickHouse.
- Afterwards, the data becomes available in DataLens for quick visual analysis and building dashboards.
DataLens: Usage scenarios and the most popular tools
Using DataLens for regularly updated live data helps the company quickly track changes, compare data from different periods, and promptly react to any anomalies.
For the marketing and sales departments, which work extensively with DataLens, top priority is given to dashboards in several areas:
- Gross sales figures by service type, sales channel, average check, etc.
- Penetration levels of additional services on sale.
- Dynamics for new services (such as installment payments).
- Indicator dynamics by channel: B2B, B2C, domestic and international flights, airlines, etc.
A variety of cross-sections offers additional detail for sales dynamics reports and dashboards. These dashboards regularly monitor changes in dynamics, and ad hoc queries in the report builder can identify their causes.
The customer base’s parameters are also regularly monitored. This cohort analysis of retention and outflow parameters, arrival dynamics of new customers, and a customer portrait profile is then used as a basis for customer retention measures. In DataLens, the data found can be easily summarized and visualized, giving Anywayanyday the option, for example, to create a list of customers who plan on leaving and come up with an attractive offer for them.
ClickHouse is effective for collecting, storing, and analyzing large volumes of unstructured data from various sources visualized using DataLens. This is important for seeing the impact of a promotion, the penetration of sales from a particular partner, or to track changes in dynamics in a particular segment.
For example, a joint campaign with a partner bank was launched to attract new customers at an affordable price. Customers received special conditions for purchases on anywayanyday.com. Using DataLens, Anywayanyday could track the entire course of the campaign: record the influx of new customers, promptly understand when a campaign loses its potential, quickly assess the cost of attracting customers, and compare communication channels by efficiency.
In DataLens, deviations from average indicators are clearly visible, making it easy to respond. Oncem when a decrease was observed in the sale of insurance as as additional service, DataLens made it possible to see the deviation, identify the cause, and bring additional insurance sales back to target levels.
Impact on the business
The transition to Yandex Cloud analytics allowed Anywayanyday to reduce their infrastructure and support costs, while also freeing up employees' time.
As a result of the 20% reduction of labor costs for the maintenance of the database and reporting system, the IT department was able to redirect these resources to other important tasks.
Overall dependence on the IT department decreased, along with waiting times to receive important information. It’s no longer necessary to ask analysts to upload data and then wait for execution. Now, marketing specialists can quickly monitor changes, make ad hoc reports and cross-sections, and prepare mailing lists in just a few minutes. They can also make more targeted offers by region.
Anywayanyday employees report that DataLens is easier to learn, faster to work with, and setup is more intuitive in comparison to other corporate BI systems. They plan to expand the areas of analytics available in DataLens, prepare public materials (dashboards and charts), and make this content available for integration on their website.
Opinion
By transitioning to Yandex Cloud, we were able to use DataLens to aggregate and visualize data from different sources, reduce our maintenance costs, and provide all departments with access to reporting. This made our employees less dependent on the IT department. Reports in DataLens are available for independent analysis and customization, so the marketing department can continuously monitor the market, make ad hoc reports and cross-sections, and quickly respond to changes based on the most recent data.