Innodatatics rated among the "20 Most Promising Data Analytics Solution Providers - By CIO Review"
Analytics in Telecom and Technology

What are the challenges faced by Telecom & Technology Industry?

The Telecommunication Industry is an industry where there is no lack of information. Examining Telecom data permits exciting insights, including, estimating the client shake.

...Read more

Case Study 1:How to predict network outage to increase the availability of services
Business problem Statement

Customer frustration occurs when a service is unavailable due to maintenance work, despite the fact that a downtime alert is broadcast. We can assume the clients' situation if the unavailability were due to a network outage or a system malfunction. Telecom companies cannot guarantee quality service if they cannot identify the network outage, upfront. There are too many network devices connecting various geographies stretching thousands of miles and predicting the malfunction of any of these is a tedious task.

Business Challenges

  • Increase in maintenance costs due to frequent faults in the telecommunication network
  • Unable to proactively identify the vulnerable locations of network outage
  • Increase in operational expenditure because of replacement of the faulty equipment used in telecommunication
  • Limitations on analyzing the huge volumes of log files to extract meaningful insights to identify the problematic network components

Business Solutions

  • A network gear failure or malfunction was identified using a built-in prediction model, allowing for constructive monitoring.
  • Association rules & clustering techniques are used to identify those locations, which frequently break down leading to customer churn
  • Survival analysis performed on each cluster to determine the number of days where a network operates without any failure
  • Natural Language Processing & text mining performed to analyze the huge logs of data to identify anomalies

Business Impact

  • Reduction in the number of network outages / faults
  • Reduction in maintenance cost due to proactive strategies
  • Increase in customer satisfaction because of increased reliability in network
  • Improved network availability & up time
  • Reduction in customer churn & thereby increase in profits

Case Study 2: How to automate the call center service desk
Business Problem Statement

Service desks’ manual process of routing the incidents received from customers to appropriate resolver groups introduces questions on brain-drain (inappropriate usage of knowledge workers). This activity will bring down the throughput rate significantly and this manual process is prone to human errors. Research has shown that significant amount of time is spent in re-assigning incidents to correct resolver groups, which were earlier assigned to inappropriate resolver group. This rigmarole introduces risks of missing the service level agreements & paying huge penalties for the slippages.

Business Challenges

  • Enormous service desk staff needed for assigning incidents to appropriate departments
  • Large number of incorrect assignments of incidents to departments due to manual process
  • Large volume of emails going unaddressed leading to SLA (Service Level Agreement) slippages
  • Decrease in customer satisfaction due to high-resolution time, caused by delayed assignment of incidents to appropriate group
  • Lot of manual effort in identifying top issues pertaining to different applications supported

Business Solutions

  • Built a classification model, which scans all the incoming emails and sorts them to a distinct application, thereby automatically assigning them to the appropriate resolver groups
  • Unstructured data from the ticketing tool was run through text mining to generate word clouds, thereby identifying the top 5 issues pertaining to each & every application; giving an opportunity for problem management team to provide a permanent fix for the most occurred incident

Business Impact

  • Optimized utilization of service desk
  • Reduction in resolution time due to assignment of incidents to appropriate resolver group
  • Increased quality of resolution
  • Reduction in no. of SLA slippages
  • Increased customer satisfaction
  • Drastic reduction in manual effort spent on analysis

Case Study 3:How to predict the high/low utilisation of network bandwidth to strategize the marketing & pricing plans
Business Problem Statement

Of late, Internet has become an essential commodity in ones life. Due to the burgeoning demand and uncertainties, the prediction of network bandwidth peaks has become a challenge to the network providers. This is causing an imbalance in catering to the necessities of customers by the network providers. Often, these providers are left with improper strategies pertaining to Marketing and Pricing plans.

Business Challenges

  • Increase in maintenance costs due to frequent network outage
  • Unable to effectively determine the bandwidth load (peaks & troughs) across the daily 24 hour period
  • Complexity in determining the architecture for load balancing
  • Customer dissatisfaction due to reduced bandwidth availability to the customers at peak time
  • Uncertain capacity planning to meet the future demands
  • Haphazard marketing plan based on artificial increase in demand

Business Solutions

  • Statistical model built by considering detailed historical Net Flow data of network into account
  • Built prediction model to identify the utilization of network bandwidth
  • Association rules & clustering techniques are used to identify those locations where the utilization is high, helping in pricing analytics
  • Statistical model to predict the probability of customer purchasing if discount coupon is sent as part of marketing strategies

Business Impact

  • Improved accuracy in network capacity plan
  • Reduction in the number of customer complaints
  • Reduction in maintenance cost due to proactive strategies
  • Increase in customer satisfaction due to effective bandwidth plan for the future
  • Increased customer base due to effective marketing
  • Reduction in customer churn & thereby increase in profits

Case Study 4: How to Improve MS Lync Call Quality
Business Problem Statement

For an improved outcome of meeting, we need an effective clarity in the communication medium. There are lots of instances where many big deals have become void just because of low quality in the call. Connection issues result in dissatisfaction for employees. Frequent disconnection, voice degradation during the call adds to the frustration of an employee, resulting in company resorting to high cost equipment.

Business Challenges

  • Employees are facing issues in connecting the call using Lync
  • Frequent disconnect during the call
  • Voice degradation during the call
  • Potential cost increase on the VOIP based equipment

Business Solutions

  • Past 36 months of organization data which was stored in Hadoop platform were considered
  • Analyzed each session of the Lync call
  • Provided good insights on the Lync call quality such as if user calls to another user who is inside the company network Vs outside company network etc.
  • Tree based algorithm which predicts whether a call would be a good call or bad call
  • Issued best practices to enjoy the greater Lync call experience

Business Impact

  • Potential cost reduction – Employer contemplating to buy VOIP phone for every employee
  • Effective communication during the meeting
  • Less number of tickets related to Lync issues
  • Improved employee satisfaction

Contact Us