Industry Trends/ Overview/ Challenges:
Declining ARPU (Average Revenue per User), shrinking profits are key challenges of Telcos across the world; in fact, it’s very difficult for smaller Telcos to survive. The revenue loss is the biggest challenge faced by Telcos. It’s inevitable to look at new revenue opportunities/ streams.
Consumer behavior is a key driver for all markets and industries with no exception to telecom.
Understanding the consumer behavior and consumer preferences to provide personalized, valuable smart offerings at the right time to grow customers and business. The key for success is by combining data analytics with real-time, contextual information to “detect” and “capture” more opportunities for monetization.
The only route to tap the subscribers when it comes to ARPU is by offering personalized promotions based on consumer behavior/ preferences. The right product at the right time for the right customer at the right price/ cost is the Mantra for CSPs/ Telcos.
The proposed approach can be broadly categorized into three steps –
- Insights/ Analytics – Capture insights from device and network and combine these insights with consumer profile, usage, preference and transform into valuable business intelligence
- Monetize – Personalize service plans based on business intelligence thereby increase the ARPU
- Retain – Retain the subscribers by improving customer satisfaction
Such systems are typically known as Personalized Ads/ promotions/ recommendations or Smart Ads/ Promotions/ Recommendations.
Sample Use Cases:
Few example use cases are shared here.
- Real-time location based promotion – Increase the uptake rate of promotions by triggering a unique promotion based on subscriber’s current location details combined with consumer’s past usage behavior, subscriber’s preference.
- Data usage based promotion – Engage subscribers when it is immediately relevant (while browsing YOUTUBE, heavy on data) and suggest offers that enhance their experience for a limited-time.
- QoS (Quality of Service) based promotion – Monitor the QoS status and trigger a promotion when QoS status goes below the desired level.
- Sentiment Analysis – Analyze the comments/ posts/ tweets related to the offered products/ campaigns through Facebook and Twitter and provides the summary of opinion whether it is positive/ negative/ neutral and the past trends. It needs to run NLP (Natural Language Processing) engine for analyzing the post.
CSP/ Telco Service Provider example:
“MyTelco” is an operator (fictitious operator) out of the modern telecommunication world. This operator is committed to provide a unique customer experience. MyTelco is in a highly competitive market, causing decline in ARPU and shrinking profits.
- Total number of active subscribers for MyTelco ~ 70M
- Average event size (Event data records includes voice call, SMS, data, recharge and other tickets) daily ~ 1.5TB
“Anna Mahony” is a young and modern lady, professional journalist, always on tour and represents the end-user or MyTelco’s subscriber.
Anna is traveling to Singapore, the moment she arrives to Changi airport, the smart phone triggers the backend engine. The engine augments and correlates Anna’s profile data, with past usage details – data, international calls; along with Anna’s action for previous recommendations to identify a contextual/ personalized, tailored offer and then recommends Anna with a promotion in real-time.
Given that MyTelco has high-volume, high-velocity, and high-variety information asset demands cost-effective, innovative forms of information processing for enhanced insight and decision making. A traditional RDBMS is neither cost effective nor scalable, that’s why Hadoop eco system is proposed. Various MLibs (Apache Spark’s scalable machine learning library) are proposed to recommend a contextual and personalized offer.
Data security/ privacy:
Subscriber tend to think – “Targeted Online Promotions: A threat to personal Identity and Security?”; several measures need to be taken/ proposed to address these concerns.
- Consent management – The proposed solution shows useful ads/ promotions by using data collected from subscriber’s devices, location, past usage, reaction towards past ads and profile details within CSP’s database. Subscriber is presented with ‘explicit consent’ dialog, without end user’s explicit consent, data collection/ trigger/ analysis for smart recommendation will not happen.
- Sensitive data – Sensitive data such as credit card details will never be stored in the proposed system.
- Preferences – Subscriber has the option to control his/ her preferences, for example, not more than 2 promotions a day to avoid spamming. Subscriber also has the provision to sign-off from smart promotions/ ads any point of time. The system should provide an ability to store the subscriber’s preferences.
Research indicates that consumers want to see relevant advertising/ promotions. Personalization in advertising is becoming ever-more sophisticated. The advantages of personalized advertising/ promotions for consumers is invaluable. Personalized Ads/ Promotions is the future of marketing.