OMIS 6230 V – Case Study 1

Yahoo’s Acquisition of Tumblr

数据分析mid代考 Yahoo is a digital media business that primarily focuses on informing consumers, connecting them with other people, and providing them…

Executive Summary

Yahoo is a digital media business that primarily focuses on informing consumers, connecting them with other people, and providing them with entertainment via research, communication, and the creation of digital content goods. Europe, the Middle East, Africa (EMEA), and the Asia Pacific are the four geographical regions in which the company operates. Jerry Chih-Yuan Yang and David Filo established the business in January 1994, and it is based in Sunnyvale, California. In 2013, Yahoo wanted to extend its service and decided to buy Tumblr for $1.1 billion with the goal to create an integrated social network that would make the digital world more open and stable.

Assuming the role of an analyst, we will investigate the details in the initiatives of Yahoo’s acquisition of Tumblr. Next, based on the available data, we would discuss how the web traffic relates to revenue and utilize a Lagged Regression model to generate the traffic prediction. Lastly, an explanation is provided to describe the rationale about whether it is worthwhile for Yahoo to spend $1.1 billion to acquire Tumblr based on company business models and synergy estimates.

  1. Why is web traffic an important metric to understand in Yahoo’s acquisition of Tumblr? How does it relate to revenue? 数据分析mid代考

Yahoo, a successful web service provider, focuses on providing popular web services such as their web portal, webmail, and search engine to consumers. Yahoo is continuously seeking opportunities in other related areas, such as blog acquisition rights for Tumblr. Marissa Mayer, the CEO of Yahoo, considered this acquisition a welcoming opportunity that will positively affect Yahoo’s revenue. The main reason for this decision was the massive web traffic that Tumblr would bring to Yahoo’s domain. Web traffic becomes an essential measurement for various reasons such as growth, target market, and revenue.

Traffic is crucial for a web service company, and Tumblr, a site with heavy traffic flow, is a great candidate for this reason. Tumblr has an established platform with over 300 million visitors, 120,000 new signups per day and blogs spreading in multiple areas. In addition, more than half of Tumblr’s users use their mobile app, with an average visit of 7 sessions per day.

The acquisition of Tumbler will allow the growth of Yahoo’s audience by 50% and traffic close to 20%. This is the baseline for Yahoo to reach their monetization goal by acquiring Tumblr.

It is also foreseeable for Yahoo to enhance and increase awareness among the younger generation by leveraging the fact that Tumblr has a higher attraction level towards teenage groups than other social networks. Tumblr’s consumer base is primarily composed of teens and young adults, thus creating a potential for Yahoo to consolidate the existing customer base in this age group. This effectively allows Yahoo to utilize this age group to open the market for young adults who are non-Yahoo users. Peer conformity refers to the matching behaviors and attitudes that adolescents tend to have due to peer influence. It is generally easier for teens to catch up with their friends because of this influence. This allows Yahoo toquickly become a popular trend amongst younger teens and lock web traffic of this age group to Yahoo’s web services.

Lastly, Tumblr’s advertising method similarly resembles Yahoo’s revenue strategies. 数据分析mid代考

Tumbler’s brand advertisers deliver this message by establishing emotional bonds with consumers, and it is a different method than Google or Facebook, which use direct clicking ad methods. Fortunately, this lies in Yahoo’s revenue strategy and is one of the primary sources of earnings. Most of Yahoo’s earnings are generated from Yahoo Search Revenue and Display Revenue. Yahoo Search Revenue generates revenues when an end-user clicks on a sponsored link. In contrast, Display Revenue recognizes revenue from the number of impressions during specific periods to a particular audience. Therefore, a similar advertising method will improve Tumblr’s ability in revenue strategy after being purchased by Yahoo.

The statistical explanation below also shows the relevance between web traffic and revenue. According to the regression statistics table (Figure 1), it could be known that the adjusted R Square is 0.991209. Here, this number is close to 1, which means the regression model conducted below is worthy of further study. From the above correlation matrix, 0.996085062 could indicate that there is a strong positive relationship between traffic and revenue (in millions). This evidence supports our idea that web traffic positively influences Yahoo’s revenue. Hence, web traffic is crucial to be taken into Yahoo’s consideration of acquiring Tumblr.

                                                (Figure 1)                                                                                              (Figure 2)

The line chart above (Figure 2) infers that there exists an increasing trend of the fitted regression line. It implies that there is a positive relationship between the independent variable web traffic and the dependent variable revenue.

(Figure 3)

In addition to the regression line, the regression table (Figure 3) provides more details regarding the coefficients, standard error, t-Stat(t-statistic), P-value, and confidence interval. The Linear Regression line could be written as 数据分析mid代考

In order to further interpret the table, the significance test should be conducted to determine whether web traffic has significant impacts on revenue. It can be seen that the t-Stat of coefficient 1 of the web traffic is |31.87054227|, which is greater than the critical value |1.96|.

This means that it rejects the null hypothesis H0:1 =0 at a 5% significance level with a 95% confidence interval. Thus, this variable web traffic has significant effects on the revenue and is statistically different from zero. Furthermore, 1 (0.053880801) can be interpreted as a one-unit increase in the web traffic which would give rise to 0.053880801 unit growth in revenue. Thus, the regression model that we built proves the positive relationship between web traffic and Yahoo’s revenue.

Additionally, the network effect can be utilized to underline Yahoo’s initiative for acquiring Tumblr. The network effect is defined as a phenomenon where increased numbers of participants can improve the value of a good or service. Yahoo, viewed as a social and web platform, connects both advertising agencies and users. Understanding Yahoo’s major revenue streams, this effect assists Yahoo to attract more advertisers while having a more extensive user base. Since Tumblr has a large user base, with more users being brought to Tumblr, Yahoo can utilize this potential to attract more advertisers due to the different age groups that Tumblr targets. In return, more revenue and profitability will be generated as the network effect continuously develops.

  1. Create a demand prediction model to forecast the amount of web traffic Tumblr will observe over the 100 months from June 2013 to September 2021. What model do you believe best fits the data? Comment on why it’s important to use this model as compared to calculating the average monthly growth rate and extrapolating from this value over the next 100 months. 数据分析mid代考

To estimate the amount of web traffic that Tumblr experiences over 100 months from June 2013 to September 2021, the most effective and appropriate model is the Lagged Regression model.

The People column from Exhibit 1 provides data to plot an initial trend observation, providing the most direct and intuitive method for past traffic analysis. For visual inspection, Tumblr’s past 38 months’ traffic has a non-zero mean, and there is no significant seasonality observed from the diagram below. However, there is an overall upper trend of traffic as time increases, as shown in Figure 4.


(Figure 4)

Continuing with in-depth analysis, the ACF Chart (Figure 5) shows that the autocorrelations for different time lag intervals are very high. The red lines on the graph below indicate the 95 percent confidence intervals in probability. The data set cannot be considered Gaussian White Noise since any autocorrelation values outside the confidence bands could be viewed as a signal of relevance between past traffic and present traffic. Especially from Lag 1 to Lag 8, the autocorrelation is way more extensive than the UCI. In addition, the PACF Chart (Figure 6) shows that there is a significant partial autocorrelation value at Lag 1. This implies that the past data of Tumblr’s traffic could be used to predict future traffic.

                                                                 (Figure 5)                                                                           (Figure 6)

After implementing Partition under time series, the data can be split into 80% training set and 20% testing set, with the 38-month data divided into 30 months and 8 months. Multiple demand forecasting models were taken into consideration, namely Moving Average, Exponential Smoothing, Holt-Winters, SARIMA, and Lagged Regression. Each model has its own unique strengths and limitations.

Beginning with SARIMA, it can not be used to forecast the value of the next 100 months. Since there are only 38 observations, SARIMA requires at least 50 observations to be considered a moderate size.

The comparative analysis should be utilized to determine which model is optimal amongst the four options for model selection. The first step is to determine the eligible model by comparing the RMSE to the standard deviation.


As illustrated below, all four models successfully fulfill the requirement that the RMSE should be smaller than the standard deviation. By contrast, the Lagged Regression model outstands the other three models with the smallest RMSE by 7862395.75.

It is challenging to obtain reasonable time lags to establish the Lagged Regression model. 数据分析mid代考

The initial time lag was chosen from 1-month lag, 6-month lag, and 12-month lag. There is a diminishing trend presented in the traffic forecast, which is not the most accurate or valid data to be considered in this case. These estimators were abandoned because the data was not displayed on a yearly basis, and there were only nine months given in the year 2013. Thus, the final decision was made quarterly, resulting in a time lag of 1-month, 2-month, and 3-month.

The Lagged Regression model for the training set can be generated using the Data Mining feature, with the regression equation as written below.

The projected value of the test set is determined by using the formula above. The residual of the test set is then calculated by finding the difference between actual and forecasted traffic. Using Lag Analysis under ARIMA, the ACF and PACF Chart of the testing set residuals can be shown below. There is no significant value in both ACF and PACF charts, which indicates that the model adequately explains all the signals, and the residual is White Noise.


The spreadsheet “LaggedReg” primarily shows the predicted web traffic for 100 months using the Lagged Regression model. There is an upward tendency based on the diagram shown below.

To further clarify the selection of the Lagged Regression model, it provides a multi-linear relationship while the average monthly growth rate is presented as a convexity shape. TheLagged Regression model takes various potential factors into consideration. One of the factors is an advantage that results in a more accurate and precise relationship between each component and predicted web traffic.

Beyond the intersection point, the average monthly growth rate continues rising while the Linear Regression stays within its gradually increasing pattern. If Yahoo solely focuses on the average monthly growth rate, Tumblr’s forecast will be biased, and the gap will be magnified as the prediction moves forward. In addition, every dependent variable of the time lag has its own corresponding coefficient, revealing whether there is a positive or negative relationship between the independent variable and each time lag.

  1. Should Yahoo have paid $1.1 billion for Tumblr? 数据分析mid代考

Yahoo’s acquisition of Tumblr for $1.1 billion had sparked controversy amongst business venturists. While some believe it is an improper decision for Yahoo to invest in Tumblr, others argue the amount of money invested is not ideal. Based on the data observed from question 2, Yahoo should not have paid $1.1 billion for acquisition based on a variety of perspectives listed below.

By applying the Lagged Regression(1/2/3) model, the estimated number of users using Tumblr’s website worldwide was 530.479 million in September 2021. The estimated firm value of Tumblr is $1.23 billion, which is roughly the amount that Yahoo paid for the acquisition. As shown in Exhibit 5, the average monthly growth rate is 1.52% each year. Applying this growth rate over a course of 100 months will result in an inappropriate prediction of 615.99 million users visiting Tumblr’s site across the globe. This prediction by Yahoo is 85.507 million more than the number derived from our Lagged Regression model. This infers that it is not sufficient to predict the firm value of Tumblr solely based on the average monthly growth rate, and Yahoo should have reassessed whether the firm value of Tumblr was actually worth $1.1 billion.

Yahoo highlighted in their case that Tumblr has more than 300 million unique visitors each month. 数据分析mid代考

However, since this figure also includes visitors who only use Tumblr for a short period of time, this information could prove deceptive towards many businesses. Unlike other social media platforms, Tumblr is an open network, which means that any user can visit the site without registering. The fact that marketers do not have access to each user’s personal information makes it even more difficult for them to earn profit from these short-term visitors. Furthermore, Tumblr refused to make the number of true users accessible to the public. When Bain & Company interviewed 352 global leaders, they discovered that overestimating synergies was the second most frequent cause for poor estimate of web traffic. It is critical to have an appropriate estimate of both synergies since even a tiny error from these estimates may have a significant impact on the true value of an acquisition.

As discussed in case, the acquisition of Tumblr has the potential to increase Yahoo’s viewership by 50% to more than a billion monthly visitors, and thus enhance traffic byroughly 20%. Acquiring Tumblr may assist Yahoo to transform its public image and gain users from the highly sought-after 18-to-24-year-old demographic group, which is a crucial target for advertisers.

However, despite the fact that the revenue potential of Tumblr has yet to be proved, Yahoo is too optimistic about the future return, and it could potentially lead to misjudgement of how long it would take for synergies to materialize. Most of the analysts from Disruptive Technologies at HFS Research remain skeptical about Yahoo’s potential to turn a profit on Tumblr. (Raman, S., 2013.) When integrating new operational processes to large existing workforces, it is normal to take a significant amount of time and resources to ensure smooth functionality. Acquisition and operational procedures are time-consuming processes that can take months or even years to complete. There would be a risk that Yahoo is overestimating the synergy, which will not be realized in the end.

Reference: 数据分析mid代考

Ankeny, J. (2013). Reports: Yahoo in talks to acquire tumblr for $1B. FierceMobileContent, Retrieved from

Engert, O., Kaetzler, B., Kordestani, K., & MacLean, A. (2019, March 27). Organizational culture in mergers: Addressing the unseen forces. McKinsey & Company. Retrieved October5, 2021, from

Forbes Magazine. (n.d.). Yahoo. Forbes. Retrieved October 4, 2021, from

Raman, S. (2013). Could yahoo’s tumblr acquisition be the next instagram?FierceMobileContent, Retrieved from

ValueWalk: 70% of M&A firms overestimate synergy and savings (2014). . Chatham:Newstex. Retrieved from 数据分析mid代考

What are Mergers & Acquisitions? 4 key risks: HBS online. Business Insights – Blog. (2019,July 25). Retrieved October 7, 2021, from

Understanding the Network Effect. Investopedia. (2021). Retrieved 8 October 2021, from

Cuofano, V. (2021). How Does Yahoo Make Money? Yahoo Search Business Model -FourWeekMBA. FourWeekMBA. Retrieved 7 October 2021, from