Revenue forecasting based on business opportunities / von Fabian Hauser
VerfasserHauser, Fabian
Begutachter / BegutachterinKopacek, Peter
ErschienenWien 2016
Umfangv, 74 Blätter : Diagramme
HochschulschriftTechnische Universität Wien, Univ., Master Thesis, 2016
Schlagwörter (EN)Sales forecast / Cognitive bias / Data analysis / Regression / Resampling / Value at Risk
URNurn:nbn:at:at-ubtuw:1-2956 Persistent Identifier (URN)
 Das Werk ist frei verfügbar
Revenue forecasting based on business opportunities [0.76 mb]
Zusammenfassung (Englisch)

CRM Opportunity data from a corporation are analysed for patterns. These data is created from salesmen and provide business opportunity information with win probabilities and volumes. The first step of the analysis is to create crosstables. In the table, observed probabilities and subjective probabilities vary a lot. We also check for probability independencies of opportunities at MDF compared to EFY volumes. As a result, we do not find dependences. The sum of all open weighed MFY opportunities from FY 2012 - 2015 calculated as a forecast for the EFY total volume is â 168.712.641. The realized volume is however â 53.178.551, that is only 33.8% of all weighted opportunities. Even if we assume that we have only the prior knowledge, the 30.6% of the FY are successful, the subjective probabilities are actually pointing into the wrong direction and we obtain a much better EFY estimation than using the subjective probabilities. There is no relationship between the subjective probability judgment and the won opportunities. Finally, a way is found to smooth the crosstable with a logarithm, where again similar probability categories are aggregated to make relationships clearer. In the end, it is clear to see that small volume opportunities tend to have higher success rates than large volume opportunities. Further a model with a deviation of less than 1% is found and a 10% VaR is calculated. It turns out that the model with three sized categories is the favourable, because 10 size categories seem too many and likely create an overfitting effect within the data. The main conclusion is that the company has an issue of very overoptimistic salesmen. We propose not to rely on the -experience- of the salesmen but consider the size of an OI as relevant indicator. Small and medium OIs have higher success probabilities than the average, while large OIs have much smaller ones. Big opportunities are won rather rarely (18%). Maybe the corporation should invest more in their acquirement efforts for big projects. However, a goal needs to be for this company to train their sales staff to make them more sensitive for their business estimations to get better data input and concluding to more data that are reliable to process