Effect of Software on CSP Performance
In order to verify how (if at all) CSP's investments in software affect their business performance, the relationship between (i) CSP's spending on software and (ii) its key performance indicators (KPIs) is considered below. The investments in software are represented by internal, external, and total software spending a year or two years prior to the KPI values, while the KPIs include average revenue per user (ARPU), revenue, net income, EBIDTA, and change in revenue (as an indicator of company growth). Besides, Opex and Capex along with R&D expenses are used as control variables. We use the data collected for the years 2004–2007 (revenue, net income, EBITDA, market capitalization, R&D, CAPEX, OPEX), and 2004–2006 (ARPU). The data is used to construct regression models wherein the dependencies between KPIs and software spending are evaluated in terms of the variance explained by each of the independent (or control) variables.
The dependency of CSP's KPIs on software spending, and Opex, Capex, and R&D expenses are summarized in Fig. 1. The thickness of arrow lines indicates the magnitude of the variance explained (thick line indicates the coefficient of determination of R2>0.5, thin line − 0.1<R2 <0.5, and dotted line − 0.05<R2<0.1. The type of the arrows designates the sign of the regression coefficient: the double (single) line corresponds to the positive (negative) value of the coefficient.