USING MULTIPLE REGRESSIONS MODEL to ANALYZE REVENUE DECREMENT on PT.

ADXN

Andrew Richard Togatorop

1140003462

BINUS BUSINESS SCHOOL MASTER of MANAGEMENT in STRATEGIC MARKETING BINUS UNIVERSITY JAKARTA 2011

USING MULTIPLE REGRESSIONS MODEL to ANALYZE REVENUE DECREMENT on PT. ADXN

Andrew Richard Togatorop

1140003462

A Project submitted to the Graduated Faculty In Partial Fulfillment of the Requirements For Master of Management Degree Major: Strategic Marketing BINUS UNIVERSITY

USING MULTIPLE REGRESSIONS MODEL to ANALYZE REVENUE DECREMENT on PT. ADXN

Andrew Richard Togatorop

1140003462

Supervisor:

Ir. Muhril Ardiansyah, M.Sc, Ph.D 25 July 2011

TABLE OF CONTENTS

TABLE OF CONTENTS LIST of TABLES LIST of APPENDICES/LIST OF FIGURES CHAPTER 1: INTRODUCTION 1.1 1.2 1.3 1.4 BACKGROUND PROBLEM STATEMENT LEARNING OBJECTIVE SCOPE of THE STUDY

CHAPTER 2: LITERATURE REVIEW 2.1 2.2 THEORITICAL FOUNDATION RESEARCH PRACTICE

CHAPTER 3: RESEARCH METHODOLOGY CHAPTER 4: RESULTS and DISCUSSIONS CHAPTER 5: SUMMARY REFFERENCES APPENDIX

CHAPTER I. INTRODUCTION

1.1

BACKGROUND
PT. ADXN is one of largest display Ad Network Company in SEA and become

pioneer in Indonesia since 2007. Until now Admax has been partnership with hundreds of local publishers and running almost all the campaigns from major brands or advertisers from Indonesia market. ADXN business models basically rely on the advance technology that could connect all the publishers to become a large network that could represent most of the online activities in Indonesia. The network clustered based on the content and audience profile into channels which become ADXN products. Advertiser could buy the channels based on Cost per Mil (CPM) or Cost per Click (CPC). Through a single platform ADXNs advertisers could access multiple websites, target the audience that they desired and manage their campaign performance effectively. ADXN really concerned on clients campaigns performance since this the only key value that makes differences with other conventional Internet Advertising which often doesnt care about it.

In 2009 ADXN Indonesia succeed to grow the company in terms revenue and number of human resources as seen on chart 1. Variety of products also make huge impact on the uptrend progress, here is the product list: 1. ADXN Channels as the main product of the company.
2. Yahoo! (Official Reseller).

3. MSN (Official Reseller). 4. Friendster (Official Reseller). One of the fast growing social networking sites in the world before the Facebook era.

Chart 1.1 Organizational Structures

1.2

PROBLEM STATEMENT

The real turning point started when ADXN enter their fourth fiscal year. ADXN revenue started having a downtrend each month and management realizes there were several factors that cause the hassle: 1. Turnover number increase. Some of the ADXN employees were resigned and move to other online company. 2. Friendster has lost their value because Facebook started to grow even bigger than expected. 3. ADXN ended the contract as Yahoo official reseller because Yahoo will open their office in Jakarta. 4. Solely depend with ADXN channels for the product which most of the client still unfamiliar with the concept. 5. Competitors start to penetrate ID market, each with their unique selling point. This research will only evaluate the internal factors from ADXN therefore external factor such as competitor would not included.

1.3

LEARNING OBJECTIVE
Expected result from the research is to understand which factors that really

impact the campaign revenue performance. From there the management will have the knowledge that could help them to determined next step for the company. Since internet still showing never ending growth in terms of penetration and innovations, in the future the research result could help other online start up company to solve similar issue with ADXN case.

1.4

SCOPE of THE STUDY

from human resources, products, and sales performance. The collected data will only reflect from the last 30 months revenue report. The outcome of this research is focusing on how to capitalize each factor and should not be valid as reference for forecasting company performance.

CHAPTER II. LITERATURE REVIEW

2.1

THEORITICAL FOUNDATION
2.1.1 Internet Industry

Internet has become a very common term in public society. It already part of people lives, they could access it at school, working, home, mobile, caf, and many other places. Refer to www.internetworldstats.com table tells that internet penetration in 2011 has reached 30.20% of world population, see Table 1.

WORLD INTERNET USAGE AND POPULATION STATISTICS 31-Mar-11 Population World Regions ( 2011 Est.) Africa Asia Europe Middle East North America Latin America / Carib. Oceania / Australia WORLD TOTAL 1,037,524,058 3,879,740,877 816,426,346 216,258,843 347,394,870 597,283,165 35,426,995 6,930,055,154 Internet Users Dec. 31, 2000 4,514,400 114,304,000 105,096,093 3,284,800 108,096,800 18,068,919 7,620,480 360,985,492 Internet Users Latest Data 118,609,620 922,329,554 476,213,935 68,553,666 272,066,000 215,939,400 21,293,830 2,095,006,005 Penetration (% Population) 11.40% 23.80% 58.30% 31.70% 78.30% 36.20% 60.10% 30.20% Growth 20002011 2527.40% 706.90% 353.10% 1987.00% 151.70% 1037.40% 179.40% 480.40% Users % of Table 5.70% 44.00% 22.70% 3.30% 13.00% 10.30% 1.00% 100.00%

Table 2.1. Internet World Usage and Population

The most simple definition of Internet was mentioned on Vurro, Marcella essay that It is a virtual global place, were people globally come together to interact and exchange information (Vurro, 2009). Internet has released the distance boundaries for people to connect each other. Since the introduction of internet and lately web 2.0, there are millions active and non active websites. 2.1.2 Internet Advertising

The essence of Internet Advertising is an advertising medium that could provide a flow of information available to consumers on demand enabled through technological devices in interactive and personalized environment (Vurro, 2009). Online Ad Networks is a part of Internet Advertising tools that could aggregate traffic that was previously too difficult buy or undesirable because of too much website inventory wastage or reluctant to have an advertisement (DeSilva. Phillip. 2008). Ad networks Company has the technology that could manage inventory from many websites or publishers to be used by advertiser to perform their campaign. The technology will cluster publishers invetories based on targeting segmentation mostly based on content and user demographic. 2.1.3 Online Advertising Business Model

Ad network often positioned themselves as Media partner for Advertiser while in Publisher perspective they become the advertisers because part of the revenue contribution. Chart 2 could explain furthermore about the work flow. The essential of Ad network business model is the inventory acquisition strategy; representation, direct acquisition with profit share or direct acquisition via arbitrage. There were three ways for Advertisers to purchase Ad networks services (DeSilva. Phillip. 2008): CPM (Cost per thousand impressions) CPC (Cost Per Click) CPA (Cost Per Action)

Chart 2.1 Ad Network Work Flow.

2.1.4

Ad Network Business

Most of Ad network were considered as start up company with minimum resources and infrastructure. Maintain revenue level above average has always become every business goal, but each business on different factors to achieve their goal. Ad Network Company also relies on some fundamental factors to maintain revenue level as mentioned below. Sales order

Sales person Average sales value Product

2.2

RESEARCH PRACTICE
2.2.1 Past Research

The research model will use Multiple Regressions. Noor, Gusti (2008) stated that this model was suitable for identifying correlation between dependent variable and more than one independent variable. The formula is: Yi = 0 + 1X1i + 2X2i + i Yi = Dependent Variable 0= Y intercept 1= slope of Y with variable Xi 2= slope of Y with variable X2 i= random error in Y for observation i

CHAPTER III. RESEARCH METHODOLOGY

3.1

RESEARCH DESCRIPTION
This research used secondary data that has been extracted from monthly revenue report of PT. ADXN for the last 30 months. The last 6 weeks raw data has been through of variable selection, defined the research hypothesis, and learnt more about theoretical foundation. Next phase the data will go through quantitative analysis process with Multiple Regressions model to examine the variable correlation and test the hypothesis which leads to conclusions and recommendations.

3.2

FRAMEWORK
Research processes were breakdown into some phase as mentioned on the figure with below explanation: 1. Data extraction. In this case the researcher used secondary data that has been extracted from PT. ADXN internal system. 2. Problem Finding. The data was taken from year 2009 up to 2011 and it shown there is the declining trend in terms of revenue. 3. Identifying Variables. Number of sales person, number of proposal, and number of sales value were the chosen variables.

4. Defined the hypothesis. This is the key points because it will lead the research to the finding the conclusions. Fault hypothesis will mislead the entire process of research. 5. Data examination using Multiple Regressions. The analysis using multiple regressions will form in two ways, using Microsoft Excel and manual calculation. 6. Hypothesis Testing. This phase will involve two type of test which are F test and T test. 7. Conclusions and Recommendation. Finally the final phase where finally reach the conclusions of the research and share recommendation to solve the problem.

Chart 3.1 Research Framework

3.3

POPULATION

The secondary data was taken form the internal record systems of PT. ADXN. It recorded monthly sales data from January 2009 up to June 2011, in total there 30 month (see appendix 1). The captured data was list as below: 1. Sales persons 2. Date 3. Clients 4. Sales order
5. Average sales amount

6. Revenue

3.4

SAMPLE
Multiple regressions model need at least 30 samples data to be consider valid. Therefore number of sample will reflect the number of population. The sample data only used number of sales person, sales order, and average sales value to be process on the research, see table. The graphic shows that revenue declining and this research will find the solutions to keep the trend downward in the future.

Month Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09 Jul-09 Aug-09 Sep-09 Oct-09 Nov-09 Dec-09 Jan-10 Feb-10 Mar-10 Apr-10 May-10 Jun-10 Jul-10 Aug-10 Sep-10 Oct-10 Nov-10 Dec-10 Jan-11 Feb-11 Mar-11 Apr-11 May-11 Jun-11 Jul-11

Sales person 3 5 6 6 6 6 6 6 6 6 6 6 4 3 3 2 3 3 3 4 3 3 3 3 3 3 2 2 1 3 2

Sales Order 1 2 3 3 1 4 2 2 2 1 4 7 6 4 7 12 8 11 13 12 16 8 4 4 6 6 5 5 9 9 4

Average Sales Value 228,690,039 129,111,875 148,528,642 311,506,671 45,472,000 143,749,657 176,523,839 118,869,483 158,634,283 93,251,352 112,139,703 119,946,647 141,103,324 199,497,113 118,984,650 151,575,112 159,125,993 344,704,463 184,086,314 133,033,157 162,084,341 97,523,852 151,597,010 57,604,967 153,685,255 63,333,339 87,500,005 96,884,240 61,666,689 97,280,512 113,685,188

Revenue 228,690,039 258,223,750 323,619,800 623,013,342 45,472,000 574,998,627 353,047,678 237,738,965 317,268,567 93,251,352 448,558,812 839,626,528 846,619,944 797,988,452 832,892,552 1,818,901,348 1,273,007,944 3,791,749,097 2,393,122,078 1,596,397,880 2,593,349,457 780,190,815 606,388,038 345,629,800 922,111,532 380,000,032 350,000,018 484,421,200 555,000,204 875,524,606 454,740,750

Table 3.1 Data Sample

Graphic 3.1 Data Sample Trend

3.5

RESEARCH MODEL
3.5.1 Variable

Based on the table 1, the sample data contained dependent variable which is revenue and independent variables where the correlation needs to be proven in this research. These are the independent variables:

1. Sales person. Ad network business really relies on their sales force.

In this research will see the correlation in terms of quantity rather than quality. 2. Sales Order. Number of sales order that contributes by sales person every month.
3. Average Sales Value. It shows average numbers of sales that

happened each month.

3.5.2

Hypothesis

This research was meant to find the significant correlation between dependent variable and independent variables. These are the hypothesis statements:
1. H0: There is no correlation between overall independent

variables (sales person, sales order, and average sales value) with revenue performance in PT. ADXN. H1: There is correlation between overall independent variables (sales person, sales order, and average sales value) with revenue performance in PT. ADXN.

2.

H0: There is no correlation from each independent variable

(sales person, sales order, and average sales value) with revenue performance in PT. ADXN. H1: There is correlation from each independent variable (sales person, sales order, and average sales value) with revenue performance in PT. ADXN.

3.5.3

Multiple Regressions

In order to understand the correlation between the dependent variable and independent variables, therefore the analysis will use the multiple regressions model as follows:

R = b0 + b1 SP + b2 SO+ b3 ASV R SP SO ASV = Revenue = number of sales person = number of sales order = average of sales value

3.5.4

Hypothesis Testing

1. F Test.

This test objective to determine whether there is a significant relationship between the dependent variables and set of independent variables. Below is the hypothesis model to be used: Ho: b1=b2=b3=0 Ha: b1 or b2 or b3 0
SSR k SSE n k 1

MSR F= = MSE

Figure 3.1 F-Test Graphic 2. T test

The objective for this test is to shows if there is linear relationship from each independent variable with dependent variables. Below is the hypothesis model to be used: Ho: bi = 0 (no linear relationship) Ha: bi 0 (linear relationship exist)

t=

bj 0 Sbj

(df = n k 1)

Figure 3.2 T-Test Graphic

3.6

PROGRAM for RESEARCH

CHAPTER IV. RESULT and DISCUSSION

4.1 4.2 4.3

DESCRIPTIVE STATISTICS INFERENCE STATISTICS DISCUSSION

CHAPTER V. CONCLUSIONS and RECOMMENDATION

5.1 5.2

CONCLUSIONS RECOMMENDATION