SevenTech is a systems integrator that develops and connects mobile services to deliver text and voice messages to customer phones by email, SMS, messengers, social media, or mobile apps.

By building an analytical solution based on Yandex Managed Service for ClickHouse and Yandex DataLens, the company saves up to 450 hours of effort for managers and technical experts per month. This minimizes the risks of gaps in accounts receivable or cash, and increases the efficiency of account managers. Here’s how they did it.

Collect, process, visualize

Within a month of the company’s launch in August 2018, SevenTech had connected 20 partners who sent 20 million messages to their customers.

Each message is far more than a text: there’s a message time, the addressee name and destination number, the type of traffic, message status, and much more. We needed a solution to quickly collect, process, analyze, and visualize data to generate partner reports and drive our business decisions.

Our legacy DBMS, MS SQL Server, was inefficient in Big Data analytics: we faced multiple issues. Data processing, for exmaple, took a long time. That’s why we decided to try ClickHouse. First, we tried it in our provider’s cloud infrastructure and got excited. ClickHouse showed 10x better performance: in a couple of minutes it processed 4 billion rows, which previously would have taken several hours. That’s why we abandoned MS SQL Server and launched ClickHouse in our domestic service provider facilities.

When we revised our data collection and processing workflow, the next thing was to quickly deliver information to our customers and employees. Ideally, we wanted to exclude programmers from this process.

As long as we had small amounts of data, we had a simple process. The manager told the programmer what data they needed, e.g. statistics on a certain operator or message type. The programmer quickly wrote an SQL query and sent the results to the manager in a couple of minutes.

But in four months, as we reach 65 partner companies, and the monthly message count exceeded 70 million, this model was no longer sufficient. Retrieving information consumed a lot of time and programmer resources, so we needed to find another solution to visualize data and build reports.

Based on our SQL queries, we wrote scripts to export data to Excel files. As a result, our managers could easily retrieve data samples and analyze them. The visualized metrics were presented in Grafana.

As of March 2019, SevenTech had 89 partner connections, 90 million messages per month, dozens of Excel files, and countless visualizations in Grafana. For partners to track operational statistics, we built several duct-taped web apps in Django. The process wasn’t easy to manage.

By the end of 2019, our services processed more than 190M messages per month. In two years, our revenue grew to billions of rubles. So, we needed a service to visualize data for our employees internally and partners externally, coping with the constantly increasing data volumes.

Yandex Cloud and Yandex DataLens for visualization

Yandex DataLens — the service that best suited the company’s needs in data visualization and analysis. It didn’t require complex setup: you could start working right away. Moreover, it allowed us to:

  • Provide access for every company’s employee and even external partners.
  • Operate in full gear, even on a free plan.
  • Build all the reports we needed for our business users.
    We learned about the product at Yandex Scale 2019 and immediately decided to try it. The following day, account managers could generate reports on their own and create the dashboards they needed. We didn’t consider other BI solutions: DataLens fully met our needs.
    But although we moved to DataLens in-the-cloud, we still ran a non-cloud ClickHouse DB internally at SevenTech.

Switching to Yandex Managed Service for ClickHouse

Properly maintaining and expanding a ClickHouse deployment takes time and money. We estimated the consumption cost for a cloud-based managed ClickHouse and decided to move to Yandex Managed Service for ClickHouse.
Thanks to the Yandex Cloud support, we managed to complete all migration-related tasks quickly. In September 2020, we finalized our migration to Managed Service for ClickHouse.


Now, managers use selectors to build their samples and solve analytical tasks, e.g. they can select a period, customer, messages sent, and messages billable.

That's what a tabular report looks like in Yandex DataLens. It helps us reconcile with operators and partners.

You can also build more sophisticated analytical reports to present business data to the executive team, or key metrics to your customers, visualizing the amount of traffic by customer groups, or planned KPIs. A sales director, for example, can assign KPIs for managers and track performance in every area, comparing indicators for previous periods if needed.

We connected our employees and partners to Yandex DataLens: everyone can view and analyze statistics in a convenient format now: this is another advantage of the service.

This is what the report in Yandex DataLens looks like. On the left: KPIs for each area by month. On the right: Normalized total indicators for all the areas. Below, there’s a table containing absolute figures for key metrics and KPIs by month and area.


Yuri Kazankin,
SevenTech CTO
Yuri Kazankin,
SevenTech CTO

We are constantly trying something new, with both mature and emerging products. Currently, we are looking to expanding the platform to Managed PostgreSQL based on Yandex Managed Service for Apache Kafka®. Managed PostgreSQL fits naturally into our development environment: both our developer and QA teams are satisfied with it. Some of our internal applications are already ready to migrate from local DB installations to the managed environment, and we’ll focus on this in the near future.