Estimating the market potential pre-launch with search traffic

Transcription

1 Estimating the market potential pre-launch with search traffic Oliver Schaer Nikolaos Kourentzes Robert Fildes 38 th International Symposium on Forecasting 18 th of June 2018

2 Motivation for generating pre-launch forecasts Increased competition lead to shorter life-cycles and faster introduction of new products Forecast needed to adjust marketing planning and operations Difficult to estimate the market potential (Goodwin et al. 2014) Expert judgment bias when forecasting new products (Belvedere & Goodwin 2017; Tyebjee 1987) Consumer preferences change during pre-launch phase (Meeran et al. 2017)

3 Obtaining pre-launch market potential estimates By expert judgement Survey (Lee et al. 2014, Kim et al. 2013, Bass et al. 2001) By analogy Parameters from previous or similar products, i.e. product attributes (Kim et al. 2014, Goodwin et al. 2013, Lee et al. 2003, Lillien et al. 2000) By market research Conjoint (Orbach & Fruchter 2011, Lee et al. 2006) Pre-orders (Moe & Fader 2002)

4 Obtaining pre-launch market potential estimates By expert judgement Survey (Lee et al. 2014, Kim et al. 2013, Bass et al. 2001) By analogy Parameters from previous or similar products, i.e. product attributes (Kim et al. 2014, Goodwin et al. 2013, Lee et al. 2003, Lillien et al. 2000) By market research Conjoint (Orbach & Fruchter 2011, Lee et al. 2006) Pre-orders (Moe & Fader 2002)

5 Last year s ISF

6 Pre-release buzz (PRB) PRB is the aggregate anticipation of consumers towards a new product (Houston et al. 2018) includes (i) communication, (ii) search, and (iii) participation in experiential activities Why PRB supposed to work? Less information compared to post-launch forecasting available, i.e. univariate information PRB has smaller potential to have endogeneity than post-release phase Several studies which include pre-release buzz report increase in forecast accuracy.

7 Pre-release buzz forecasting literature

8 Investigating pre-launch buzz for sequential products Most products nowadays are not entirely new but released as sequential. What is the relation of pre-launch buzz to the market parameter for sequential product releases? Pre-release buzz also able to improve long-term forecasts and what lead time can we achieve? Using direct analogy of previous releases requires very few training sample.

9 Peak scaled search traffic Sales Generations of video games within a franchise Sales for Assassin's Creed j=1 j=2 j=3 j=4 j=5 j=6 j=7 Google Trends j=1 j=2 j=3 j=4 j=5 j=6 j=7

10 Peak scaled search traffic Sales Proposed methodology Sales Actuals Forecast j=1 j=2 j=3 j4 Search traffic t j=1 j=2 j=3 j=4

11 Predicting the market potential with pre-release buzz (PRB) In addition we define the same models also in an additive way

12 Empirical evaluation on 56 franchises (257 games) Google Trends with game title as keyword: PRB j w = 4 to 12 weeks Google Trends window l = 1 to 12 weeks lead time Set of parametric diffusion models Smoothed adoption shape of j-1 as non-parametric model Average Relative Absolute Error on the cumulative adoption with m j-1 (Davydenko & Fildes 2013)

13 Overall performance

14 Performance of GT percentage across lead times Forecasting accuracy gains available 10 weeks before release

15 The story so far for pre-release buzz Managerial implications for pre-launch buzz: Automated pre-launch forecasting process improving accuracy by up to 15% compared to established analogy based methods Search traffic information providing up to 10 weeks lead time Good accuracy with simplest model that is available for the 2 nd product generation Further research Can we improve the accuracy by including trend direction Compare to other marketing variables such as ad spending Can this information also be used to allocate success of products by just looking into the pre-launch buzz?

16 Thank github.com/mamut86

17 References Asur, S., Huberman, B., Aug Predicting the future with social media. In: Web Intelligence and Intelligent Agent Technology (WI-IAT), 2010 IEEE/WIC/ACM International Conference on. Vol. 1. pp Belvedere, V., Goodwin, P., The inuence of product involvement and emotion on short-term product demand forecasting. International Journal of Forecasting 33 (3), Craig, C. S., Greene, W. H., Versaci, A., E-word of mouth: Early predictor of audience engagement. Journal of Advertising Research 55 (1), Davydenko, A. and Fildes, R., Measuring forecasting accuracy: The case of judgmental adjustments to SKU-level demand forecasts. International Journal of Forecasting, 29, Dhar, V., Chang, E. A., Does chatter matter? The impact of user-generated content on music sales. Journal of Interactive Marketing 23 (4), Ding, C., Cheng, H. K., Duan, Y., Jin, Y., The power of the like button: The impact of social media on box office. Decision Support Systems 94, Divakaran, P. K. P., Palmer, A., Sndergaard, H. A., Matkovskyy, R., Prelaunch prediction of market performance for short lifecycle products using online community data. Journal of Interactive Marketing 38, Foutz, N.Z. & Jank, W Research Note Prerelease Demand Forecasting for Motion Pictures Using Functional Shape Analysis of Virtual Stock Markets. Marketing Science, 29(2),

18 References Gelper, S., Peres, R., Eliashberg, J., Pre-release word-of-mouth dynamics: The role of spikes. Erasmus University Rotterdam working paper, 1-40, Working paper 14th Sept. Goodwin, P., Meeran, S., Dyussekeneva, K., The challenges of pre-launch forecasting of adoption time series for new durable products. International Journal of Forecasting 30 (4), Hann, I.-H., Oh, J., James, G., Forecasting the sales of music albums: A functional data analysis of demand and supply side p2p data. Working Paper Houston, M. B., Kupfer, A.-K., Hennig-Thurau, T. and Spann, M., Pre-release consumer buzz. Journal of the Academy of Marketing Science, 1-23 (Pre-print). Kim, H., Hanssens, D. M., Advertising and word-of-mouth effects on pre-launch consumer interest and initial sales of experience products. Journal of Interactive Marketing 37, Kim, T., Hong, J., Kang, P., Box office forecasting using machine learning algorithms based on SNS data. International Journal of Forecasting 31 (2), Kim, T., Hong, J., Koo, H., Forecasting diffusion of innovative technology at prelaunch: A survey based method. Industrial Management & Data Systems 113 (6), Kulkarni, G., Kannan, P., Moe, W., Using online search data to forecast new product sales. Decision Support Systems 52 (3), Lee, H., Kim, S. G., woo Park, H., Kang, P., Pre-launch new product demand forecasting using the bass model: A statistical and machine learning-based approach. Technological Forecasting and Social Change 86,

19 References Lee, H., Kim, S. G., Woo Park, H., Kang, P., Pre-launch new product demand forecasting using the bass model: A statistical and machine learning-based approach. Technological Forecasting and Social Change 86, Lee, J., Boatwright, P., Kamakura, W. A., A bayesian model for prelaunch sales forecasting of recorded music. Management Science 49 (2), Lee, J., Cho, Y., Lee, J.-D., Lee, C.-Y., Forecasting future demand for large screen television sets using conjoint analysis with diffusion model. Technological Forecasting and Social Change 73 (4), Lilien, G. Y., Rangaswamy, A., Van den Bulte, C., Diffusion models: Managerial applications and software. In: Mahajan, V., Muller, E., Wind, Y. (Eds.), New-Product Diffusion Models. Kluwer, Massachusetts, pp Liu, Y., Word of mouth for movies: Its dynamics and impact on box office revenue. Journal of Marketing 70 (3), Meeran, S., Jahanbin, S., Goodwin, P., Neto, J. Q. F., When do changes in consumer preferences make forecasts from choice-based conjoint models unreliable? European Journal of Operational Research 258 (2), Meeran, S., Jahanbin, S., Goodwin, P., Neto, J. Q. F., When do changes in consumer preferences make forecasts from choice-based conjoint models unreliable? European Journal of Operational Research 258 (2), Moe, W. W., Fader, P. S., Fast-track: Article using advance purchase orders to forecast new product sales. Marketing Science 21 (3),

20 References Mülbacher, H., Füller, J., Huber, L., Online forum discussion-based forecasting of new product market performance. Marketing ZFP 33 (3), Orbach, Y., Fruchter, G. E., Forecasting sales and product evolution: The case of the hybrid/electric car. Technological Forecasting and Social Change 78 (7), Tian, C. H., Wang, Y., Mo, W. T., Huang, F. C., Dong, S., Huang, Prerelease sales forecasting: A model-driven context feature extraction approach. IBM Journal of Research and Development 58 (5/6), 8:1-8:13. Tyebjee, T. T., Behavioral biases in new product forecasting. International Journal of Forecasting 3 (3), Wang, F., Zhang, Y., Li, X., Zhu, H., Why do moviegoers go to the theater? The role of prerelease media publicity and online word of mouth in driving movie going behavior. Journal of Interactive Advertising 11 (1), Xiong, G. and Bharadwaj, S., Prerelease Buzz Evolution Patterns and New Product Performance. Marketing Science, 33 (3),