Sigma S7 — the solution for real-time airline sales control

S7 Airlines is the largest private airline in Russia, with the most modern fleet in the Russian air transit market. Their extensive network of routes allows passengers to travel to 181 cities in 26 countries across the world. S7 Airlines is among the top three airlines in Eastern Europe, according to the prestigious international Skytrax rating.

Key goals

1. Accelerate the processing of data on flights and services
2. Reduce labor costs for the accounting processes and settlements with contractors and partners
3. Provide management with a clear and accurate picture of their business as it is
4. Give S7 information systems with access to up-to-date and reliable sales data

Solution

Sigma by True Engineering is an accounting system that automatically processes data on sold tickets and rendered services, generates invoices for S7 interline transportation partners, and allows to create claims for agents.

The solution processes airline sales data as it comes from external sources. S7 employees are virtually spared from manual work on settlements with the company ticket sellers and partner airlines.

More than 100 controls created by the team search for errors in documents, reconcile data, correct simple discrepancies.

During peak periods, the solution processes over 2,500 transactions per minute.

The first step is to understand the processes

At the very beginning of the project, we took a deep dive into the industry standards, studied the under-the-hood operation of the previously used accounting system. The goal was to understand how all the key processes work: pricing, preparing documents and settlement letters, interacting with airport services.

We realized that the service was basically a very powerful Excel spreadsheet. Employees entered data along the chain, the system provided calculations. Everything depended on users who had to manually bring the data to the target state, and the actions had to be carried out sequentially — if there is no "A", "B" will not work either.

In fact, the share of automatic transactions was far from 100%, because the employees had to invest their resources throughout the process.

How the development went

We applied a hybrid model: instead of developing a large system for a long time, we implemented the product part by part.

Advantages of the hybrid model:

• Transition to a new solution is seamless and has minimal impact on the customer's processes.
• The company gets a faster tangible result allowing for evaluating the effectiveness of a new solution.
• Employees have time to get used to new tools, the gradual launch of functional blocks simplifies training.

First, accounting and settlements were decomposed into separate parts: sales from direct agents and aggregators (the so-called neutral environments), working with global reservation systems (GDS), cargo transportation, preparation of data for accounting and accounting.

The team then went on to take the functional modules one by one, create the corresponding blocks, transfer historical data from the old system, test and open the functions to users.

In each case, at first there was a transitional period, when the old and new modules worked side by side while we evaluated the effectiveness, corrected errors appearing in real working conditions. The team had to make sure that the new system provided the same result as the old one. So, module by module, we launched the new solution.

Architecture as an advantage

We created a pipeline-based architecture: the data moves through the process, the system enriches it, requests information from global schedule systems, tariff bases, cross-industry information systems.

Accounting systems use industry algorithms for working with data, proprietary S7 technologies, clearing rules, billing and reconciliation procedures.

Wherever possible, operations are carried out in parallel. Only if an error is found, the transaction is submitted for verification by an S7 specialist.

Result — 99% of transactions are processed automatically. The remaining 1% is the result of external factors such as changes in the tariff base.

How it works

1. Raw sales data comes into Sigma from exchange files or online sources.
2. The system generates transactions that are written to the database.
3. Transactions are checked for errors.
4. The airline employee corrects discrepancies, if any.
5. Closing of the day — separately for direct agents and neutral environments.
6. Settlement letters are created — the main document for mutual settlements between the airline and its agents.
7. Letters are received in personal accounts on the S7 agency portal.
8. Manual study of automatic processes

The faster the data is processed, the faster the airline can bill its counterparties and get money. To speed up these processes, we, together with the customer, divided the verification of transactions into several parts:

1. An invoice is generated when the system processes 80% of transactions
2. The remaining 20% ​​goes to the next bill
3. For each process, we worked out the timing — how to ensure profit faster
4. To speed up the result, in some cases we add computing power, in others — refine the algorithm logic

Success from the very start

Already at the MVP stage, the system proved its efficiency. Parallel data handling accelerated processes, automation relieved employees. Correcting 1-2 obvious errors is not at all the same as double-checking 30-40% of transactions.

S7 has its own knowledge base. The airline methodologists understand how the algorithms work, participate in product development along with the True Engineering analysts and developers. Together, they turn business needs into algorithms that translate into new product features.

The airline provides itself with full control over the product. The system develops along with the business. Flexible architecture helps to quickly respond to external needs. Our goal is to provide a minimal Time-To-Market of new features.

To further increase the efficiency of Sigma, we have integrated its internal processes into the existing S7 workflow:

The main calculations take place at night and early in the morning

• The beginning of the working day in Novosibirsk — the system provides for verification transactions that could not be processed automatically.
• By 10 a.m. Moscow time, the central office receives ready-to-work, verified sales data for the previous day

Results

• Control and management of S7's core business — all sales and sales of aviation services are connected in a single chain.
• Automation of complex work — airline employees get rid of many manual operations for processing flights and controlling sales.
• Increasing labor productivity — the airline receives a technological foundation for the development of high-tech expertise.
• Reducing labor costs — about 200 people were engaged in maintaining the old accounting system, only 30 employees work with Sigma.
• Growth and development — high level of automation allows to scale business without bloating the staff.