India Peoplemeter Update

List of Contents

Section	Section No.	Page no.
Introduction		1
Sample Plan	I	4
Reporting	II	7
Town Selection	III	7
Technology Used	IV	9
Meter Migration	V	9
Time Plan	VI	10
Implication of Expansion Plans on Data Analysis	VII	11

List of Tables

No.	Table	Page No.
1	Top 6 Metro Sample Sizes	5
2	State level coverage and sample sizes	6
3	Sample sizes by reporting strata	7
4	Towns Selected per Reporting Stratum	8
5	Type of Market and Impact on analysis	14

Introduction

The original expansion plan presented by TAM to the Research Design Sub-Committee (RDSC) and the Technical Committee (TC) envisaged not only a broader coverage (more markets) as compared to the present panel but also a deeper coverage by covering the Less than 0.1 Mn. urban/semi-urban population stratum. After extensive meetings with the RDSC and the TC, the total sample size for this coverage - ~ 100% urban India was calculated at 5800 homes.

While the ultimate aim is to get to the above coverage, feedback received by TAM indicated that this is best done in stages.

Accordingly the expansion has been broken into two stages, each stage emphasising a different aspect of the expansion plan.

Stage I – Broaden coverage to ~ all states but within the existing reporting stratum i.e Greater than 0.1 Mn. population stratum

Stage II – Deepen coverage to the Less than 0.1 Mn population stratum for all markets covered in Stage I.

This document :

1. Describes the Stage I expansion plans as was arrived at after consultations with the

RDSC

2. Gives an update on the progress of the expansion &

3. Provides an end-user perspective on the expansion front by detailing the impact on data analysis

I. Sample Plan

The Stage I sample plan originally presented to the RDSC included reporting all Class I markets except for J&K, Assam and North-East, Bihar and Rajasthan.

The RDSC suggested making adjustments to the plan by including Bihar and Rajasthan so as to maximise coverage. Acting on this suggestion, the sample plan was reworked and is described below.

Representation and Coverage

a. An ~ 100% representation of urban TV owning homes in the Class I (greater than 0.1 Mn. population) strata.

b. The above coverage accounts for 31 mn TV homes – a 35% increase from the present coverage

c. Increase in sample homes from the present 3454 homes to 4555 homes – a 32% increase

d. The sample plan is designed to maximise the effective sample size. To achieve this, an extensive Sample Optimisation exercise was conducted so that the maximisation is effective whichever way the data in analysed, either from a market point-of-view (analysis across strata) or a stratum point-of-view (analysis across markets).

e. A comparison of the proposed and the present sample sizes in the metros is given below:

Table 1 : Top 6-Metro Sample Sizes

Metro	Existing Sample (homes)	Proposed sample (homes)
Mumbai	312	450
Delhi	355	425
Calcutta	217	265
Chennai	225	255
Bangalore	239	255
Hyderabad	225	255
Total	1573	1905
A 21% increase in Metro sample size

f. Similarly, a comparison at the State/Rest of State sample sizes is given overleaf :
Table 2 : State level coverage and sample sizes

Market	Existing Sample (homes)	Proposed sample (homes)
RoAndhraPradesh	220	305
Gujarat	195 (Ahmedabad)	270
PHCHP	165 (Ludhiana)	260
RoKarnataka	220	115
Kerala	208 (Cochin)	220
RoMaharashtra	221	295
RoTamilNadu	220	230
Uttar Pradesh	432 (Kanpur + RoUP)	330
RoWestBengal	0	110
Orissa	0	110
Bihar	0	90
Madhya Pradesh	0	215
Rajasthan	0	100
Total (excluding 6 metros)	1881	2650
Total (including 6 metros)	3454	4555
A 41% increase in the State/Rest of State sample size
A 32% increase in overall sample size – including metros

* The name Uttar Pradesh(UP) as will be used for reported purposes includes the newly formed state of Uttaranchal. Similarly Madhya Pradesh(MP) includes Chhatisgarh and Bihar includes Jharkhand

* PHCHP is Punjab + Haryana + Chandigarh +Himachal Pradesh

II. Reporting

While currently all non-metro markets were reported as Rest of States, these would now be reported by strata. Thus, Rest of Maharashtra can now be analysed as RoMah. 1 Mn.+ and RoMah. 0.1-1 Mn.

TAM would now be providing a total of 29 reporting markets which is almost double the present reporting of 15 markets

Table 3 : Sample sizes by reporting strata

Proposed sample sizes at the state level (figures are number of sample homes)
Market	1 Mn.+	0.1-1 Mn.	Total Homes
RoAP	110	195	305
Gujarat	160	110	270
PHCHP	105	155	260
RoKar	0	115	115
Kerala	110	110	220
RoMah	115	180	295
RoTN	110	120	230
UP	150	180	330
RoWB	0	110	110
Orissa	0	110	110
Bihar	30	60	90
MP	105	110	215
Rajasthan	30	70	100
Total	1025	1625	2650

III. Town selection

¨ The town selection procedure makes use of the Simulated Annealing technique. The process ensures that the selected towns represent their respective reporting units on the parameters of C&S % Penetration, Terrestrial Availability of DD2, and Geographic Spread by SCR control.

¨ A highlight of the selection is that TAM has ensured the sampling of all 1 Mn.+ towns.

¨ With a fixed sample size per market one could either choose

1. More samples per town (with a lesser number of towns) or

2. Lesser samples per town (with more number of towns).

The present town selection has chosen the latter line of operation. This maximises spread and representation though being operationally and financially more demanding.

Table 4 : Towns Selected per Reporting Stratum
Market	1 Mn.+	0.1 - 1 Mn	Total
RoAP	1	6	7
Gujarat	3	3	6
PHCHP	1	4	5
RoKar	0	4	4
Kerala	2	3	5
RoMah	2	5	7
RoTN	2	4	6
UP	5	6	11
RoWB	0	3	3
Orissa	0	3	3
Bihar	1	2	3
MP	2	3	5
Rajasthan	1	2	3
Total	20	48	68
Total number of towns sampled including Top-6 Metros = 74

IV. Technology used

The currently used ‘Frequency Monitoring’ Technology will continue to be used by the service. The actual meters that will be used are a mixture of ACN and AGB frequency meters.

V. Meter migration

Due to the fact that a state will use either an ACN or a AGB meter alone, the markets have been split by type of meter to be used. Hence a market that is currently using ACN meters for reporting might need to use AGB meters for the same.

This means an entire process of migration so that each market would have the meter type it is allocated to. At the same time, TAM will ensure reporting of the current markets till such time that the markets graduate to the new market definitions.

Thus if market XYZ is to be metered with AGB meters and is currently ACN metered, it means de-installing the existing ACN meters and installing AGB meters – at the same time ensuring that any point in time, either the old or the new market definitions (e.g. RoXYZ or RoXYZ 1 Mn.+/0.1-1 Mn) are always present.

We endeavor to constantly inform you of the expected changes, during the migration process

VI. Time-Plan

The reporting of all existing markets will not change simultaneously and will be a gradual process. At the same time there are also new markets such as RoWB which will be introduced.

We expect meter installations to be completed by December 31^st, 2002 across all markets.

The exact timings of the changes in market reporting will be intimated to you in advance.

In the immediate future, we hope to replace the existing RoKarnataka with RoKarnataka 0.1-1 Mn. and the existing Ludhiana reporting with PHCHP 1Mn.+. in the first week of November.

VII. Implications of the expansion plan on Data Analysis

In this section the service has tried to anticipate user queries and provide solutions to the same. Since these are anticipated queries, they cannot be claimed as being exhaustive. However, as always, the service will be glad to assist users in case of any queries.

1. What impact will the expansion have on the user interface ?

¨ The user interface remains as is – with, of course, more markets added!

¨ For the users to distinguish the newly defined markets from the existing ones at a glance, the newly defined markets have an asterisk (*) added before the market name, as seen in the screen shot below.

The existing market definitions are viewed on scrolling down – they are present after the new market definitions.

2. Can I group Existing Market definitions and New Market definitions ?

¨ Yes, you can group markets belonging to these sets provided that the complete data for all markets you have chosen in your group, during the selected analysis period, is available

¨ For instance, assume your group comprises of Mumbai, Rest of Maharashtra

& Gujarat 1 Mn.+ where Gujarat 1 Mn.+ is a new market, you will be able to evaluate this group only if the data for Gujarat 1 Mn.+ is available from the start date of the period you have chosen. Similarly, the existing definition of Rest of Maharashtra should also exist throughout the chosen analysis period.

3. How will the expansion process impact my existing Market Groups?

¨ Your existing market definitions will not change if you are evaluating a period prior to the new definitions coming in.

¨ But as new markets keep getting added, the old market groups may not be valid. As a simple example, your existing “India Surveyed” definition is based on the existing 15 markets. This will not be valid once data for new markets is released since the markets that make up the actual India Surveyed would be different.

¨ If you intend to look at the newly added markets as a part of your overall market group, then you will need to recreate this group in some other name.

¨ The definitions you use will be dependent on the constituent markets actually reporting data for the analysis period. If the data for any market in your market group is not completely available in your selected analysis period then the s/w will display an error “Invalid selection period, please re-select the period”.

4. I understand that you will be covering all 1 Mn.+ cities. Now assume that Cochin is reported separately till November 1^st and post-November 1^st ,Kerala 1 Mn.+ starts getting reported. Can I conduct an analysis selecting Cochin with the analysis period being October 20^th to November 20^th?

¨ Since Cochin’s identity as a separate market ceases on November 1^st and therefore not present for the entire selected analysis period, the analysis run will have to be broken into two runs : October 20^th – October 31^st and November 1^st – November 20^th.

¨ If the data for a selected market is not available for the entire selected time period then Media XPress will display an error “Invalid selection period, please re-select the period”.

5. How will the expansion impact plan building?

¨ Old market definitions for a state will not be operational once the new market definitions for the same state are in place.

¨ Thus there will be a period during the transition when the pre-plans are based on a certain market definition for a state while when the campaign is in progress the new market definitions become operational.

¨ This situation is partially comparable to the analysis period when TAM conducts the yearly universe updations ; users break the analyses runs into pre-universe updation and post-universe updation.

6. Do all the above directions relate to State and Metro markets ?

¨ The above directions apply mostly to state level markets where there is a market definition change.

¨ The top 6 metros can be run as at present.

The table overleaf summarises the market type and its impact on analysis:

Table 5 : Type of Market and Impact on analysis
No.	Market	Effect on Analyses
1	Top 6 metros	- Can be analysed at any point of time - Markets can be grouped unrestrictively within this set
2	The 4 metros reported separately now but to be amalgamated into a state reporting as part of the new market definition (Kanpur, Ahmedabad, Cochin, Ludhiana)	- Can be analysed as these markets only upto the point where they are reported separately - Once new market definitions are in place, cannot be analysed separately. - For e.g. one cannot click on Cochin when Kerala 1 Mn.+ is operational so get a valid output. - Similarly, one cannot click on Kerala 1Mn.+ and choose a data period that precedes that market definition with the aim of getting the Cochin component of Kerala 1 Mn+.
3	Existing Rest of states	- Same treatment as above (No. 2)
4	New Markets	- Can obviously be analysed only post-launch - Care to be taken that the analysis start-date is on or later than the first reporting day of the market. - E.g. Assume MP 1 Mn.+ is reported from the week beginning November 24^th. Then the start date for an analysis including MP has to be on or later than November 24^th.