As a small part of further social network analysis of a mobile (cell-phone) customer base I have examined age differences between customers and with whom they communicate most frequently.
I was also looking at how reliable it might be to guess someones age (a customer or non-customer) by extrapolating from known individuals. There is customer age approx 97% of the time, and its accurate approx 92% of the time (unusally large numbers of people claim to be born on 1st Jan 2000 :)
I was surprised to see (but then maybe I'm naive :) how so many people have 'mobile calling relationships' mainly with people the same age...
The chart below shows the percentage of customers and the age average between the people they communicate with most frequently. Age differences over 4 years are comparatively rare...
The small spike at 30 years difference is probably parent-to-child communication.
I will be using this to support an estimation of age for prospects and customers where age is unknown, but age of social group members is known.
- Tim
7 comments:
How much of that spike at zero is caused by people calling their own voicemail?
Hi Adam,
No, this wouldn’t happen :)
Only customers are included in the chart. No age for voicemail.
The analysis examines *reciprocal* communication, and voicemail is removed (along with a few other services such as SurePage). The social network analysis I have conducted is focused on communication relationships between customers and any recipient (with voicemail numbers etc removed). As part of the analysis there were some specific telephone numbers that were identified as ringtones, music subscriptions etc that were consequently filtered out. These specific services become obvious because they form huge hubs that have many thousands of customers connected to them.
We match the telephone numbers to accounts (a billing construct) that has information such as address, age and gender. Therefore where we have two mobile telephone numbers that have a calling relationship we can link these to the customer account and obtain their age. We have accurate age for approx 95% of all customer mobile telephone numbers. Theer is a small issue of some account holders having two mobile services, in these cases one is probably for the spouse or child.
In situations where the recipient of the calling relationship is not a customer (or a second mobile service) we don’t know their age, but can use this analysis to make a good guess!
In the case of multiple services within a single account (let's say the account holder is 45 years old), we often see one mobile service has relationships with 40-50 years olds, whilst the other service has relationships with 15-20 year olds. We can fairly accurately state that the additional mobile service on the account is for the child.
- Tim
- just to clarify;
I retained only relationships of customer service numbers that link to a customer account where we have accurate age -> when creating this chart. That is a significant chunk of the data retained (because we have large mobile market share).
- I can then generalise to all relationships involving any service with an unknown age. Chances are that the age of an unknown mobile service will be within 3 years of the services they share the most frequent relationship with.
- hope that makes sense!
Makes sense - thanks for clarifying!
Hi Tim,
According to your graph, those have age differences of less than 4 years could be due to the married/ in relation couple. From here, this infer that the age difference between husband and wife in this country is about 4 years. Am i statement making sense?
If correct, maybe the marketing people may come out with a creative strategy to their loyal customers. It just just my opinion. Apologized for any inconvenience caused.
Regards,
Kelvid Pang (Malaysia)
kelvidpang@gmail.com
It is not necessarily a married couple, it could just be friends or work colleagues. Yes, the purpose of this was also to allow a creative marketing message targetted by age groups.
Since my original post I was able to improve my analysis and get down to an absolute mean error of just 0.85 years.
Hi Tim,
I found that this pattern is quite interesting. I'll explore further if I deal with telco data to check the similar pattern exist.
That's cool. By reducing your error, your forecasting/modeling become stronger.
I am looking forward to your blog.
Regards,
Kelvid Pang (Malaysia)
kelvidpang@gmail.com
Post a Comment