The Single Household View For Mobile Operators

Updated: Jan 11
By: Tanul Mehta



With the advent of Big data, mobile operators are moving more and more towards a segment-of-one. This makes it necessary for operators to know who the real customer is and what is his product portfolio. It is even more important to understand not just the customer but the household that the customer belongs to. Mobile operators currently look at basic attributes like number of B numbers the SIM calls to with a certain defined frequency and applies demographic information to predict households.


The challenges with this approach are multi-fold. It is a simplistic approach with very low accuracy of less than 30%. This is because it applies too generic criteria to identify households and relies on demographic data that, in many cases, is incomplete or unreliable.


Besides, as the market characteristics change, the applied thresholds also need to change which makes it difficult to have an automated solution. Due to low accuracy of the prediction, the campaign take-up focusing on households tend to be much lower thereby impacting product penetration and top-line revenue.

With our SHV solution, we significantly improved the accuracy of identifying households from 30% to ~70%.

Single Household View is a pre-built solution by Lynx Analytics that addresses this problem and identifies the households within your customer base using your own CDR, location and device data. The approach relies on calling circle, the device attached to the SIM(s) and colocation of multiple SIM(s) and has an accuracy of ~70%.


How We Identify Households

The first step in identifying households is mapping all the SIMs that have been registered under the same ID. The ID could be either government registered IC, Billing Address or Email. For example, in the registration graph below, each vertex represents a SIM MSISDN and an edge connects two vertices if and only if the 2 vertices are registered under the same ID.




The above calling graph shows 2 segments: Segment 1 has 7 SIMs that are registered under one ID. Therefore, every MSISDN is connected to every other MSISDN. Similarly, Segment 2 has 4 SIMs that are registered under another ID.


We then identify MSISDNs that belong to the same customer using the Single Customer View (SCV) algorithm. SCV tags MSISDNs belonging to same unique customer if:

The two MSISDNs are co-located (colocation is based on cell tower location from Voice, SMS or data usage), ANDHave no calls between them, ANDThere are multiple instances where the 2 MSISDNs were connected to the same device


Below is a filtered calling graph, where the colour of the vertex represents the single customer. Segment 1 is a 4 member household with 7 SIMs, where


  • Member 1 is X97203761 (1 SIM customer)

  • Member 2 is X97203762 (1 SIM customer)

  • Member 3 is X97203763 (1 SIM customer)

  • Member 4 is X97203760 (4 SIM customer)


Similarly, Segment 2 is a 4 member household with 4 SIMs, where:


  • Member 1 is 4418047 (1 SIM customer)

  • Member 2 is 9114694 (1 SIM customer)

  • Member 3 is 2360680 (1 SIM customer)

  • Member 4 is 353154 (1 SIM customer)


Using the SHV, Lynx was able to identify 19% households with at least 2 on-net members.

We further analysed households with 2 or more on-net members and their connections with off-net customers to identify households with competition SIM. The chart below shows the distribution of on-net households with or without competition SIM.




How SHV Can Be Used

The SHV solution has been deployed in multiple markets including one of the leading Telecom operators in Malaysia which has a 40% post-paid and 60% pre-paid customer base. The operator uses the SHV for:


1. New customer acquisition above-the-line

a. Identified the households with competition SIM and target the household member with the operator’s SIM to refer their fellow household members for migrating their SIM and enjoy free subscription for 1st three months.


b. This approach generated a lift of 2x vs Business-as-usual take-up rates for the same number of target base.


2. Selling bundling SIMs to a family plans below-the-line.

a. Identified the households with multiple lines with the operator and target the primary household member with a family bundle plan. Churn rates among households with family plans tend to be 5x lower than the average churn rate among your base.


b. This approach generated a lift of 3.5x vs Business-as-usual take-up rates for the same number of target base.


The usages of SHV are numerous and through it, CSPs are able to attain a higher sense of understanding about their customers. Find out more in detail on how we successfully executed the Single Household View with this major Malaysian telco – get in touch with us today. 


Note: Above-the-line campaign is a targeted campaign that offers products that are available for the masses. The objective of such a campaign is to create targeted product awareness among your customer base to have a faster product penetration.


Below-the-line campaign is a targeted campaign that offers products that are only available to the customers receiving the campaign and it is usually not transferable.