1 Introduction
2 Related Work
Research goal | Methodology | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Literature | Satisfaction | Intention | Churn | Adoption | Usage Penetration | Data type | Statistical model | Observed vs Self-reported | Observed client characteristics | Observed activation measures | Vendor perspective |
(Ratten 2016) | x | Cross sectional | PLS | self-reported | |||||||
(Walther et al. 2015) | x | Cross sectional | PLS | Self-reported | |||||||
(Benlian et al. 2010) | x | x | Cross sectional | PLS | self-reported | ||||||
(Benlian et al. 2011) | x | Cross sectional | PLS | self-reported | |||||||
(Yang and Chou 2015) | x | x | Cross sectional | PLS | Self-reported | ||||||
(Karahanna et al. 1999) | x | x | Cross sectional | PLS | self-reported | ||||||
(Wu 2011a) | x | Cross sectional | PLS | self-reported | |||||||
(Wu 2011b) | x | Cross sectional | Rough Set | self-reported | |||||||
(Sharma et al. 2016) | x | Cross sectional | MLR, Neural Network | self-reported | |||||||
(El-Gazzar et al. 2016) | x | Cross sectional | - | self-reported | |||||||
(Wang et al. 2016) | x | Cross sectional | ANOVA, PCA | self-reported | |||||||
(Coussement and van den Poel 2008) | x | Cross sectional | SVM, Random Forest | observed | |||||||
(Frank and Pittges 2009) | x | longitudinal | decision trees, k-means clustering | observed | |||||||
(Goode et al. 2015) | x | Cross sectional | PLS | self-reported | x | ||||||
(Sukow and Grant 2013) | x | longitudinal | Array model | observed | |||||||
This Paper | x | longitudinal | LMM | observed + self-reported | x | x | x |
2.1 Continuance Use Intention in Software-as-a-Service
2.2 Identified Research Gaps
3 Hypothesis Development
3.1 Research Hypotheses
Category | Hypothesis | Expected direction | Explanation | |
---|---|---|---|---|
Software | 1 | Amount of Content | ↗ | More content leads to increase in users as the software value increases |
2 | Quality of Content | ↗ | This study hypothesizes that content quality leads to higher incentives to use software | |
3 | Single Sign-On | ↗ | Single Sign-On simplifies the login process and reduces any barriers created by having to remember passwords. Therefore usage increases. | |
4 | Design | ↗ | Contemporary and improved software design results in higher number of users | |
5 | Presence of Process-oriented Module | ↗ | Implementing a process-oriented module as opposed to only knowledge management modules requires regular user interaction resulting in a higher overall usage | |
Client characteristics | 6 | Age | ↘ | Users have difficulties to include software into their daily routine and will use it less over time |
7 | Involvement | ↗ | Clients motivating and incentivizing the use of the software have increased overall usage | |
8 | Management Structure | ↗ | A centralized strategic management on the client side leads to better decisions and activation measure which again result in higher user numbers | |
9 | Number of Contacts | ↘ | Since coordinating multiple contacts can be difficult, a high number of contact persons decreases the communication efficiency and SaaS usage | |
10 | Counseling Demand | ↗ | High counseling demand shows a strong connection between Saas vendor and client and therefore a higher motivation to use the software | |
Activation measures | 11 | Banner | ↗ | Banners redirecting users to the SaaS solution increases awareness and therefore user numbers |
12 | Newsletter | ↗ | Regular vendor newsletters increase usage by linking to the SaaS solution | |
13 | Training | ↗ | By conducting trainings for potential and current SaaS users, awareness and perceived usefulness is increased resulting in an increase in users |
3.1.1 Software Characteristics
3.1.2 Client Characteristics
3.1.3 Activation Measures
4 Methodology
4.1 Longitudinal Study Design
4.1.1 Types of Factors and their Associated Effects in a LMM
4.2 Study and Variable Description
Hyp. # | Variable Name | Type | Explanation | Source |
---|---|---|---|---|
– | id | numeric | Assigns every client to a number from 1 to 17 | Identifier |
– | t | date | measurement times | |
1 | r.statement | numeric | Amount of research findings a client’s software provides to its users | system exports |
1 | r.verbatim | numeric | Amount of consumer quotes available | |
1 | r.project | numeric | Amount of information on past projects | |
1 | r.library | numeric | Amount of research libraries | |
1 | r.searchdoc | numeric | Number of uploaded documents which are searchable by users | |
1 | r.report | numeric | Amount of company reports | |
3 | sso | binary | single sign-on enabled | interview (time varying) |
4 | design | binary | contemporary design used for landing page | |
5 | process | binary | Did a client implemented one of two process management modules | |
6 | age | numeric | Maturity of a software solution | |
11 | banner | binary | Is a banner redirecting to the SaaS solution | |
12 | newsletter | categorical | type of newsletter | |
13 | training | categorical | Information on whether trainings took place | |
2 | quality | numeric | Quality of the Softwares content | interview (static) |
7 | involve | numeric | clients involvement | |
8 | mgmt.central | binary | How was the SaaS solution managed | |
9 | contact | numeric | Amount of people taking care of SaaS solution on the client side | |
10 | phone | numeric | Intensity of phone based counseling demand | |
10 | email | numeric | Intensity of email based counseling demand | |
– | usage.penetration | numeric | percentage of active users | dependent variable |
5 Descriptive Analysis
5.1 Conditioning on Predictors
Hypothesis | Variable | Result of Exploratory Analysis | |
---|---|---|---|
H1 | Amount of Content | r.statement | |
r.verbatim | \(\leftrightarrow \) | ||
r.library | \(\leftrightarrow \) | ||
r.searchdoc | \(\leftrightarrow \) | ||
r.project | \(\leftrightarrow \) | ||
r.report | \(\leftrightarrow \) | ||
H2 | Quality of Content | quality | |
H3 | SSO | sso | ↗ |
H4 | Design | design | \(\leftrightarrow \) |
H5 | Presence of Process-Oriented Module | process | ↗ |
H6 | Age | age | ↘ |
H7 | Involvement | involve | |
H8 | Management Structure | mgmt.central | \(\leftrightarrow \) |
H9 | Number of Contacts | contact | \(\leftrightarrow \) |
H10 | Counseling Demand | email | \(\leftrightarrow \) |
phone | \(\leftrightarrow \) | ||
H11 | Banner | banner | ↗ |
H12 | Newsletter | newsletter | ↗ |
H13 | Training | training | ↗ |
5.2 Model Building
5.2.1 Model Selection Process
6 Results
6.1 Hypothesis Testing
Variable | Age | Banner | Newsletter | Training | |||
---|---|---|---|---|---|---|---|
P-value Likelihood ratio | 0.06193 | 0.00003556 | 0.0002527 | 0.0001642 | |||
P-value Satterthwaite | 0.0609803 | 0.00002883 | 0.0002673 | 0.0001737 | |||
P-value Kenward-Roger | 0.0839306 | 0.00006497 | 0.0002861 | 0.0001875 | |||
Variable | Intercept | Age | Banner | Newsletter Content | Newsletter Other | Training New | Training Existing |
P-value Kenward-Roger | 0.001993 | 0.083931 | 0.000065 | 0.000564 | 0.028102 | 0.000127 | 0.080815 |
6.2 Predictive Accuracy
7 Discussion
# | Hypothesis | Expected impact | Modeled impact |
---|---|---|---|
1 | Amount of Content | ↗ | Not sign. |
2 | Quality of Content | ↗ | Not sign. |
3 | Single Sign-On | ↗ | Not sign. |
4 | Design | ↗ | Not sign. |
5 | Presence of Process-oriented Module | ↗ | Not sign. |
6 | Age | ↘ | Not sign. |
7 | Involvement | ↗ | Not sign. |
8 | Management Structure | ↗ | Not sign. |
9 | Number of Contacts | ↘ | Not sign. |
10 | Counseling Demand | ↗ | Not sign. |
11 | Banner | ↗ | + 5.2% |
12 | Newsletter | ↗ | up to + 2.9% |
13 | Training | ↗ | up to + 2.7% |