In the 1990s the telecom industry has started conquering the market in DRC. Early in the fourth quarter in 2002, the Presidency of the Republic has issued the law 014-2002 establishing the Postal and Telecommunications Regulatory Authority (ARPTC in French). ARPTC mainly oversees compliance with telecommunications laws, regulations and conventions; in addition, it helps to define and adapt, in accordance with the orientations of government policy, the general legal framework in which the activities of the sector are carried out.

Operators have struggled to successfully expand nationwide their network coverage despite a slow internet connection in some outskirts and countryside. They strove to link the country from end-to-end, it’s one of their great contributions: Information and Communications Technology (ICT). This effort has paid off since 54% of the total population use their services. Vodacom leads the market in term of subscribers with 30%, 28% for Airtel, 23% for Orange and 19% for Africell. As a reminder, Oasis, Celtel, and Vodacom have almost come at the same period in DRC, after a while CCT came. -Zain acquired Celtel and Zain has been acquired by Airtel. -Tigo acquired Oasis ,and Tigo and CCT have been acquired by Orange.

From the 1990s to 2007, less than 50% of the customer base of each operator was not identified. It means that some subscribers have afforded a SIM card without being registered in the system. In one hand, this has shown in public how much they were less adamant at this point. In another hand, ARPTC issued a decision on the registration of all subscribers and promised an embargo whether a specific threshold is not reached. In fact, this measure shook them because the major part of unidentified subscribers consisted of prepaid, and a significant amount of revenue comes from them.

As for the postpaid base, it includes some diplomats, businessmen, members of the government, organizations, companies, and so on. These customers benefit from particular care based on the unlimited use of services offered by the operator. And the bill payment depends on the control protocols, it can be weekly, monthly, quarterly, or annually. This kind of contracts brings a lot of profits to the company and allows the company to guess with certainty the amount they may earn at the end of a month for all postpaid.

This decision arose from multiple complaints and abuse cases that have been reported to the police. Some people received calls and SMS from unknowns, sometimes these unknowns threatened them to death. As long as these unknowns are not identified, this makes it harder for detectives to unmask suspects for elucidating these cases for which they might have been collected logs. Without the substantial information related to the phone number as name and address, it’s time-consuming to assemble the pieces of the puzzle.

In my previous job as MIS (Management Information System) engineer, I was in charge of generating reports that highlight revenue, identify new business opportunities, predict future trends, customer behavior and calculate commissions of salesforce (super dealer, super vendor, etc.). I mainly worked with following departments: sales, customer service (CS), Revenue Assurance, and marketing.

As a new recruit, I started to learn the structure and role of every single table. Afterward, I developed some shortcuts to make easier my life while generating reports. There, I found an existing PL/SQL script for subscribers’ identification report. I spent some time reviewing this report and updating its chunks of code.

CS can only count identified subscribed to its own system but a bit limited to get the overview: how it works. Partners meet subscribers and ask them for the required substantial information, subsequently, they fill out an empty form. After collecting the data, at the end of the day, they bring these forms to the headquarters for encoding them. Apart from this system, IT has another data that CS does not access.

The revised version of my script approximately captured 68% of identified subscribers. It was good news for CS in particular and business in general.

This percent has smoothly increased day-by-day before our database crashed. I automated half of my tasks on the crontab interface for avoiding pressure. This allowed me to peacefully whistle in the morning and comfortably bite off pieces of the bread. When I dove into my job, against all odds the report mysteriously captured 40% of identified subscribers. I ran the script several times, but I repeatedly got the same result. This issue disappointed me, and I was that motivated to sort it out as soon as possible. After an investigation, I discovered a big black hole in our database: many data partitions disappeared, including the one of the previous day. It was a nightmare.

I initiated the highest severity incident ticket and I shared with the team in charge of OS and database administration. Meanwhile, this team was recovering this big data from the backup side: the duration to get entire data files reloaded was unpredictable. The boss came to give a gentle nudge to my manager saying: “I want everything to be done before the end of the day”.

This increased pressure amongst teams, my colleagues and I spent a sleepless night in the office. The time elapsed was tremendously annoying, at the end of recovery, I was excited to run my scripts and to close this issue. For identification report, I slightly got a rate less than 68%