Digital Transformation – Ready for the Future

-- As the No.1 quad-play service provider in Sri Lanka, how can Dialogue improve user experience, improve efficiency, and create new business models through digital transformation to facilitate the overall business development?

Dialog Axiata PLC started its journey as a connectivity service provider, with services starting from mobile then moving to fixed and to TV through acquisitions, eventually emerging as a quad-play service provider. Today, Dialog is the No.1 quad-play service provider in Sri Lanka and has acquired a strong market position across all business segments. Dialog currently holds a 51% market share in terms of mobile subscribers and approximately 66% in terms of mobile revenue. Dialog also holds 53% of the home broadband market and 72% of the Pay TV market. If we look at the overall performance of the leading operators in Sri Lanka, we can see that the top-performing operators are quad-play providers whereas others are limited to fixed or mobile services.

Addressing the CSP challenges through digital transformation

As in every other sector, telecommunications service providers in Sri Lanka face a multitude of challenges. Users always demand higher levels of experience and keeping pace with the market dynamics is difficult. We come up with the new technologies, but the applications grow at a much faster rate, further raising the demand for throughput. OPEX continues to grow, challenging the profitability of operators. Tranditional businesses like voice and SMS no longer generate revenue as they used to, so we need to invent new revenue streams. Digital transformation is the way forward to address all these challenges. It can impact the overall business by enhancing the user experience, improving efficiency and help create new business models.

Dialog's OWA Model for digitization

At Dialog, we have adopted an OWA model to drive digital transformation. It involves the three pillars: O for organization, W for work and A for analytics. We believe that any change has to start with the organization. In order to do so, we have to relook at the organization and plan how we can radically change it to face the new demands and address the new challenges. The second pillar, work, involves digitization of the processes. However, to achieve positive results, we need to critically assess all the processes and decide how we can simplify them for the best results before moving with automation and digitization. The third pillar, analytics, is all about data-driven decision making. Each and every decision to be made should be supported by data.

As a first step to achieve this, we reduced the organizational hierarchy from 7 or 8 levels to just 4. To help us inculcate the culture of agility and develop a better way of collaborating we moved to virtual teams. Individuals are not permanently attached to any one team or division; it's an agile way of working together. Next, we reinvented the processes and implemented tools to meet the new requirements such as simplicity, flexibility and agility. Also, we empowered the organization with automation and visualization, by giving everyone the right access to the necessary tools for this. Lastly, analytics, which forms the fundamental layer, helps stakeholders extract the data that is required for decision making.

The data-driven cloud transformation journey

Dialog's cloud journey was the outcome of its OWA strategy. It began when we realized that the conventional siloed organization will not support our digitization process. Thus, we came up with the concept of a 'data lake,' which is also our 'single source of truth.' Data generated or captured from different nodes (RDS, etc) are sent to the data lake in the cloud. This data is used in tandem with query engines and with APIs to create dashboards and reports. Further, end-user requirements are facilitated via various self-developed analytic models (ML/statistical/rule based). We then use several systems backed by the analytics engine to gain insights for our decision making.
Let's look at some interesting examples on how we leveraged the data lake for analytics.

Over propagation analysis: For this, we mainly use two key data sources, mobile data terminal (MDT) data and timing advance (TA) data to verify the reduction in over propagation following a cell optimization exercise. The radio network optimization team can look at a plot of this data to optimize the network, eliminating the impacts due to over propagation issues.

Coverage comparison: To compare Dialog's service with our competitors, we used to carry out physical drive tests on different routes with competitor SIMs to analyse network coverage and throughput. Now we use MDT and physical resource block (PRB) data bundled with data from apps which helps to identify poor coverage and high congestion clusters on major highways. With this, we can assess the performance of our network with greater accuracy, and avoid physical verification tests like drive tests. MDT data comes with great GPS accuracy, so we can detect where the poor coverage patches are and where we get better throughput, faster network, and the like.

Value-based CAPEX planning: All organizations face CAPEX constraints. This is the same reason we  need to focus our CAPEX investments on sites that provide the best ROI. By analyzing different parameters like performance data, coverage data, consumer complaint data and even social media, we are able to prioritize our site expansions to optimize our CAPEX investment.

Complaint bucketing: By analyzing the customer experience management (CEM) data, we are able to identify the root causes of customer complaints. This data is then correlated, to identify issues that are causing multiple complaints which can then be addressed at the network level or customer level.

AI/ML data-driven virtual NOC journey: Our Network Operations Center (NOC) used to operate with 30 or 40 people on site. Now, with the virtualized network, there is no longer a requirement for staff in the NOC to receive and process alarms. Instead, the processes are automated with auto detection, auto ticketing, and auto escalation. This empowers the regional staff and the field teams who receive alarms directly and implement solutions to resolve the detected issues. Only if it does not get resolved at the last point, do we need to attend to it personally.

Reactive vs. Predictive Maintenance: By running historical data in the prediction engine, we are now able to calculate the probability of a fault occurrence at a site or a specific piece of equipment. Over time, we have achieved greater accuracy with our predictive engines, which has given us a 20% reduction in reactive tickets and site visits, 10% to 15% reduction in cost per site visit and 0.5% – 1% improvement in site availability.

Data-driven crowdsourcing: As a quad-play service provider, we require on-site engineers to perform the installation and service at customer premises. We are now leveraging AI/ML technologies to crowd-source engineers to complete the work orders. The system can identify the right skill set for the specific job and allot the work to the next available agent, all without any human intervention. There's also a review mechanism to assess the performance of the agent and to rate their work. This not only improves work efficiency but also enables great cost savings because we are not required to maintain a large workforce, and also because we pay according to the task accomplished by the agent.

Network Transformation to Cloud Native

For the cloud transformation, we started with Network Functions Virtualization ("NFV") cloud on the telco side and a hybrid cloud on the IT side. Later we launched on-premises clouds, one NFV-based cloud for telco and the other virtual cloud for IT applications. Now, we have adopted a cap and grow strategy where we can have a mix of in-house as well as public clouds with over 2000 ecosystem partners, offering 200 industry solutions and advanced services.

On the network side, Huawei is our key partner for core site and core applications. We run a number of different applications including CEM, CWR and OWS. In Phase 1, we have started the migration of all these applications to a single cloud. In the second phase, which is planned for 2022-23, we migrated the siloed applications such as BSS to a cloud environment. This will be used to implement an online charging system and to enhance the billing system. We are now targeting a full-fledged cloud implementation starting from on-prem, which is then expanded to hybrid and eventually to a public cloud.

Enterprise offerings

For enterprise, we started by providing infrastructure, which means we built data centers and power facilities for our enterprise customers. Now we are expanding to other services like storage, compute as well as platform, and then to application development. Dialog offers the end-to-end cloud lifecycle, starting from connectivity infrastructure, platform, application development, integration, billing, and support. 

We have Dialog's own on-prem cloud as well as the public cloud and we have also launched partner programs. We also work with multiple partners to offer different services like IaaS, PaaS, SaaS, and Security aaS. In addition, we offer system integration, support, and all the other related services as one package to our enterprise customer services. We have direct sales channels for large enterprises and for SMEs and we have the marketplace where they can go and subscribe to the service. We also have local and international channel partners. Under professional services, we have partners for consultation, architecture, migrations, integration/operation, security and modernization.

We offer multi-cloud platform subscriptions and support, which provide cross cloud connectivity, multi-cloud management tools, and professional services. For multi-cloud management, we offer an integration marketplace where the customers can choose their preferred cloud provider and migration partner. A lot of enterprises may not have the skill sets to migrate their applications to the cloud, so we can extend our support there too. Even if we don't have a particular skillset, we can offer them through our partners. We also provide 24-hour monitoring of cloud services, availability, disaster recovery and a range of related services including technical support, automation, architecture, design, implementation, best practices and the necessary documentation.

Overall, the journey that we have taken has brought us to a place where we are ready for cloud, in terms of organization, processes, and technical knowledge. We collaborate with multiple partners to offer the best-in-class services to internal customers as well as our valued enterprise customers.