Forecasting success via early adoptions analysis: A data-driven study

RESEARCH ARTICLE Forecasting success via early adoptions analysis: A data-driven study Giulio Rossetti1*, Letizia Milli1,2, Fosca Giannotti1, Dino Pedreschi2 1 Knowledge Discovery and Data Mining Laboratory, ISTI-CNR, Pisa, Italy, 2 Computer Science Department, University of Pisa, Pisa, Italy * a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Rossetti G, Milli L, Giannotti F, Pedreschi D (2017) Forecasting success via early adoptions analysis: A data-driven study. PLoS ONE 12(12): e0189096. https://doi.org/10.1371/journal. pone.0189096 Editor: Ming Tang, East China Normal University, CHINA Abstract Innovations are continuously launched over markets, such as new products over the retail market or new artists over the music scene. Some innovations become a success; others don’t. Forecasting which innovations will succeed at the beginning of their lifecycle is hard. In this paper, we provide a data-driven, large-scale account of the existence of a special niche among early adopters, individuals that consistently tend to adopt successful innovations before they reach success: we will call them Hit-Savvy. Hit-Savvy can be discovered in very different markets and retain over time their ability to anticipate the success of innovations. As our second contribution, we devise a predictive analytical process, exploiting HitSavvy as signals, which achieves high accuracy in the early-stage prediction of successful innovations, far beyond the reach of state-of-the-art time series forecasting models. Indeed, our findings and predictive model can be fruitfully used to support marketing strategies and product placement. Received: August 17, 2017 Accepted: November 18, 2017 Published: December 7, 2017 Copyright: © 2017 Rossetti et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Both code and data are made available on the GitHub repository https:// github.com/GiulioRossetti/hit-savvy. Funding: This work was funded by the European Community’s H2020 Program under the funding scheme “FETPROACT-1-2014: Global Systems Science (GSS)", grant agreement #641191 “CIMPLEX Bringing CItizens, Models and Data together in Participatory, Interactive SociaL EXploratories", https://www.cimplex-project.eu and under the founding scheme "INFRAIA-1-20142015: Research Infrastructures" grant agreement 1 Introduction Every day, new artists appear on the music scene, new products are launched onto retail markets, new restaurants and businesses open up. Every day, people make choices: which artists to listen to, which items to buy at the supermarket, which restaurants to visit. Consumers’ choices determine which innovations (artists, products, businesses) will reach success and achieve large diffusion, and which ones will not. To reach success, innovations need to reach/ target the right adopters. Several classical studies [1–3] analyzed the different phases of product’s lifecycle, from the first appearance on the market to vanishing. Rogers [4] described a peculiar family of adopters: the innovators, e.g., the ones that adopt an innovation before it becomes mainstream, the ones that do not need peer pressure to make their choice. From a different, psychological perspective, Tetlock et al. [5, 6] recently identified another exciting niche of individuals, called super-forecasters, that continuously make correct predictions of future events (in controlled Q&A experiments). Tetlock’s study, which ran over a number of years, aimed to understand whether people could predict an explicit yes/no time-limited question. To make predictions, the forecasters were explicitly allowed to prepare themselves researching the particular topics they were asked about. Moreover, forecasters were allowed to PLOS ONE | https://doi.org/10.1371/journal.pone.0189096 December 7, 2017 1 / 21 Forecasting success via early adoptions analysis: A data-driven study #654024 “SoBigData: Social Mining & Big Data Ecosystem", http://www.sobigdata.eu. There was no additional external funding received for this study. Competing interests: The authors have declared that no competing interests exist. change their predictions as time goes by to match their evolving feelings about the outcome as the deadline for the question grew closer. Citing [6], super-forecasters are “[..] people whose analytic ability is considerably better than random (or who, in financial analyst terms, “beat the market”)” namely, all those individuals able to provide the correct answers on a regular basis, thus making predictions having a precision far above the average. In this work we address a question at the intersection of the two lines above: are there innovators with passive super-forecaster abilities? Or equivalently, are there users that consistently adopt, before others, innovations that will later reach success? To answer such questions, we adopt a data-driven approach evaluated on two real datasets of supermarket transactions and musical listenings. Differently, from Tetlock’s approach, we do not ask users to express their forecast whether an innovation will be a success; we observe what and when they adopt (buy or listen) in the recorded transactions. Indeed, the niche of users we target—we call them Hit-Savvy—do not train themselves to produce a correct guess. Conversely, they regularly chose to adopt innovations (we will refer to them as “items”) that are likely to reach success in the future. Such difference is profound: adoption choices of Hit-Savvy are not driven by the desired outcome as for Tetlock’s super-forecasters (i.e., early-adopters who listen to novel music do not necessarily do so because they think the artist will be successful) but by personal taste. The first contribution of our study, described in Section 2, is the discovery that Hit-Savvy do exist and that their peculiar behaviour perdure through time. We empirically observe in our data a niche of innovators that exhibit a surprising propensity to adopt future successful artists and products prevalently: moreover, such Hit-Savvy tend to last in time, retaining their ability for months or even years. Moving from such results we addressed a question whose answer can deeply impact marketing strategies: can Hit-Savvy be used to predict whether an innovation in the early stage of its lifecycle will reach success in the future? Our second contribution, illustrated in Section 3, answer such question by describing a predictive analytical process that, using Hit-Savvy as signals, achieves high accuracy in the early-stage prediction of successful innovations. In Section 4, we validate our method on the previously analyzed datasets observing high precision and recall of successful items prediction, far beyond the state-of-the-art models based on ti (...truncated)