Decrypting the Success Path of High Potential Start-ups: The Application of Process Mining Algorithms on Key Developments

Numerous data-driven approaches made their way into the strategic decision-making process for identifying companies for strategic partnerships or M&A-activities. While the data availability such as of financials of global player is given sufficiently, it's not for start-ups. However, start-ups occupy a high potential in terms of growing yields. In contrast, only 10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> of European start-ups survive their first seven years. To overcome the lack of data to evaluate start-ups, we process key developments sourced by S&P. Key developments are daily-based events such as private placements or product announcements. We preprocess time-series of key developments for around 76,600 global start-ups acting in 5 different industries. Based on criteria derived from the reference literature, we divided start-ups into less or more successful companies to train several models by using the process mining algorithms alpha, heuristic and inductive miner in combination with classification algorithms like XGBoost, support vector machine and decision tree to classify a start-up’ performance. In doing so, we significantly identify processes of successful start-ups. Beside theoretical implications regarding the application of algorithms on key developments, our approach addresses managerial implications by supporting at finding high potential start-ups to enlarge business opportunities as well as for minimizing risks deriving from unprofitable investments.