It has been a long time since I updated my blog. the reasons were many ranging form not getting free time to not knowing what to write. but few days back when i received a message from my friend notifying the change in his contact number from Airtel to TATA’s Docomo, a strange thought process developed in my mind regarding the Docomo plan “pay per second“.
At first even I was fascinated by the idea. But i felt sorry for TATA coz they seem to lose huge revenue by providing services at such cheap rates. but one day when I was traveling to my place in BMTC, again this question started to trouble in my mind. so i took a piece of paper and started to write some random minutes and seconds and tried to figure out what individual plans going to cost. To my surprise I got Docomo rates higher than normal 50 paise/min plan. So I thought to make it more structured and planned and make it free from human emotions in generating those random numbers.
Today i am really free with virtually nothing to do . so i thought let me write simple C program and make it more comprehensive. below is the simple C code. (for people who don’t know anything about programming write down some random numbers and figure yourselves out)
// rates.c :
#include <time.h>
#include <stdlib.h>
#include<conio.h>
#define NO_OF_TEST_CASES 900
#define MAX_MINUTES 10
struct callrates
{
int minutes[NO_OF_TEST_CASES];
int seconds[NO_OF_TEST_CASES];
float totalrate;
} Docomo,normal_50;
int main()
{
int count = NO_OF_TEST_CASES,min,sec,i;
srand((unsigned int)time((time_t *)NULL));
for(i=0;i<count;i++)
{
min = rand()%MAX_MINUTES;
sec = rand()%60;
printf(“%d\t\t%d\n”,min,sec);
Docomo.minutes[i] = min;
Docomo.seconds[i] = sec;
normal_50.minutes[i] = min;
normal_50.seconds[i] =sec;
Docomo.totalrate += (float)(Docomo.minutes[i]*60 + Docomo.seconds[i])/100; // convert to paisa and then to rupees
normal_50.totalrate += (float)((normal_50.minutes[i]*50)+50)/100; // convert to paisa and then to rupees
}
printf(“Docomo\t%f\nnormal_50\t%f”,Docomo.totalrate,normal_50.totalrate);
getch();
return 0;
}
( For Programming Geeks : Before commenting on the style of programming or code refinement or making it more optimal please know that I am not a great programmer so please don’t see the above code from that perspective . but if you find any logical errors in calculations or better algo for generating random nos etc commenting on that is highly appreciated.)
How to Use :
1. you just need to change NO_OF_TEST_CASES and MAX_MINUTES
How I chose NO_OF_TEST_CASES
Lets say i call 5 times a day.for one month its 5*30 = 150 and i keep this plan for around 6 months .so totally it will be 150*6 = 900 calls.
Choosing MAX_MINUTES
Lets say i speak for a maximum of 10 minutes . All my calls last within 10 minutes.
My Test Results (corrected to nearest integer…included only 5 cases for demonstration purpose)
DOCOMO: Rs. 2647
NORMAL: Rs.2431
DOCOMO: Rs.2746
NORMAL: Rs.2522
DOCOMO: Rs.2646
NORMAL: Rs.2434
DOCOMO: Rs.2727
NORMAL: Rs.2498
DOCOMO: Rs.2758
NORMAL: Rs.2525
GOT THE POINT?????????…
How we have been (at least me) fooled by the “pay per second” plan. I think this plan suits to only those who normally dont talk much over the phone. but the marketing was done so nicely that they made us to feel that we are under “gadi ka atyachaar”. When someone talks for around 10 mins then the next few seconds does not count at all. but we fail to recognize this and simply believe that this plan is the best. For those induviduals who speak a minimum of 10 minutes over phone like businessmen or the one having girlfriend/boyfriend this plan is suicidal.
You can try giving different values to the two macros according to your needs and find out which one suits you the best.
By looking at this I realized TATA is not making any loss by allowing us to pay per second and now why every operator seems to be liking this idea. After all nobody sells anything for a loss. Hats off to the guy who created this plan!!!!!!!!!
Disclaimer : I am not in favour of any operator or any plan. I was just letting a machine generate some random numbers and calculate the rates according to the paln. I assume the reader’s mind is evolved to the extent of making his own judgements without blind acceptance. If any reader claims he lost money coz of this article anything related to this, the author doesnot take any responsibility in such cases.
Any constructive comments which may lead to a healthy discussion is highly appreciated !!!!!!!!!!!
Dude ,
There are not many business magnets
and
Gf/BF – Love is nothing but a suicide. Only few ppl here also.
So gen’ly , ppl go for short calls(except parents calls)
where
Docomo Rocks !!!
But , I shd say, N/w is a serious problem if u stay in outskirts.
It’s a different type which is perfectly suited for some category.
Thanks
Shiva
@Shiva — thank you for the inputs but i am giving below stats of making MAX_MINUTE as 5…that means all calls lasts for below 5 minutes…..
DOCOMO : 1349
Normal : 1355
DOCOMO : 1384
Normal : 1387
DOCOMO :1358
Normal :1362
DOCOMO :1352
Normal :1362
DOCOMO : 1328
Normal : 1335
DOCOMO : 1323
Normal : 1337
see over a period of 6 months the advantage of Pay per call is negligible compared to normal plans …u can check for diff combinations and see
Here is a more precise analysis :
Consider this plot – http://img56.imageshack.us/img56/9704/plot.png
X axis is the number of seconds and Y axis the the amount paid in rupees.
Ignore the fact that the stepped function doesn’t appear stepped. It is because of the resolution of the plotting program… it doesn’t consider all values.
It can be clearly seen that up to about 300 seconds (5 minutes), there is a good chance that the 1ps/s plan will perform better than the 50ps/min. At exactly 300 seconds, the values are the same. So some one who speaks 5 minutes on an average per call will spend the same amount on both plans. After this point, the graph diverges and the 50ps/min plan performs MUCH better in the long run.
Even before the 300 second mark, it is not assured that the 1ps/s plan will do better. But by law of large numbers, as far as the company is concerned, they will be losing money compared to 50ps/min plan if on an average their customers speak for less than 300 seconds. But assuming most people speak randomly between 0 and 600 seconds (reasonable assumption), they will be making more money on the 1ps/s plan than the 50ps/min plan. This is again, by law of large numbers, at 300 seconds, the advantage is clear and even before that, there is positive probability that the 1ps/s plan will earn them more money, thus probability of making more money through the 1ps/s plan is more than even for a completely random time between 0 and 600 seconds.
Good tactics and good marketing. They will make a lot of money out of this if most people don’t realise this.
Nice article, Kote.. And nice follow-up too, Madhava.. An eye-opener for many, I must say..
Docomo is better in it’s own way !!!
It is Simply superb for short callers.
Go for “Pay per Call” in Tata Indicom option
where u pay 1/- for 10 min,
If u are a long time fone user.
Every thing is Best in it’s own way.
U may feel sad when u speak only 10-12 seconds when u r under ‘pay per call’ option or other fone.
Similarly , With Docomo u cant go for long time fones.
I conclude ,
Short term callers (1 – 3 min) – Docomo
Avg Time callers (4 – 8 min ) – Others
Long Time Callers ( 10+ min ) – Tata Indicom
So Select option Based on u.
Dnt calculate by Graphs and all…
Thanks
Shiva
Error: Unrecognized syntax “flaot” on line 40 in main()
Comment: You are a very bad programmer.
Solution: Please take a gun and shoot yourself.
yappa!!!!!! changed it
Too much programming has made shany into a compiler
Interesting topic.
@madhava: 1. My modification of your plot: http://img98.imageshack.us/img98/2686/phoneplans.gif [showing proper step function.]
2. Nice application of law of large numbers.
Another thing you might want to consider here is customer loyalty and the inertia to move onto something new.
The best example for this would be the Tata Indicom Pay-per-call plan.
At first sight it seems Tata Indicom would go into loss with such a plan. But we can also try and look at who the target for this plan would be.
1. Does not look like business people – they make far too many calls but may not speak for long.
2. Not for people who speak rarely. The validity charge may simply be unsuitable.
3. Then it’s probably for those who make a few calls a day and speak for a fairly decent amount of time. So, we could conclude that it’s for ‘normal’ people.
Looking at the population of mobile users, easily atleast 65-70% have to be the normal users. And generally, once you get used to something like pay-per-call, it might be hard to revert back to any of the other plans. So, it appears as if TataIndicom is – putting it the indian way – getting 2 mangoes at 1 shot. A large customer base as well as long-term loyalty.
(The above analysis was not done by me, I read it somewhere but am hard-pressed in trying to locate the original.)
Kote, just a small question.
How do you consider the average call duration?
Just to illustrate it, say if the average call duration is 400secs, for Docomo, there will be no change, but for other companies-the call duration will have to considered as 5 minutes. So an extra minute has to be accounted for non-Docomo networks.
Is the above factor accounted in your model? The problem is how the average is measured- it should be done in seconds rather than minutes for a fair comparison.
Nice analysis though.
I ll put in other comments later.
if the average call duration is 400 seconds then for non Docomo users it will be 7 minutes. For this the Docomo users will be paying Rs.4 where as others will pay 3.50 (considering 50p/min). The comparison is based on seconds only..
Nice analysis. As Madhava pointed out, the graphs for the two cases start diverging at the 300 secs. So ignoring other constraints here, if people talk for an average duration of less than 300 secs, the company may lose money.
The company makes money only if it is able to get in large numbers, or rather large volumes. Such large volumes can only be realised in urban markets, not in rural areas. This explains the kind of marketing blitz which is done by Docomo, in urban areas(cities)
Including other constraints in the model can actually make the analysis more complete-like the cost/sec for the company for using the network, bandwidth cost etc. A per second plan, also allows the company to cut costs substanially, which is further be aided by large volumes.
There is another cap which Docomo has set-STD calls will be at 1p/sec only for a call period of 7.5 hours. After that period, the charge will be 2p/sec!
Clever marketing. But I think the long term scenario may turn out to be different, if customers find out that they are actually losing more money.