Adoption of Digital Innovations: A Special Referenceto Routinised Incremental Innovations

Abstract

The study’s main purpose is to determine the adoption of digital innovations in a developing country that is continuously introduced on different service platforms. A quantitative approach is taken in the research where structured questionnaires based on the unified theory of acceptance and use of technology (UTAUT) model have been used to collect primary data. A total of 821 responses were selected for analysis using structural equation modelling. The findings suggest that the UTAUT model indicates a good fit for determining the behavioural intention to adopt incremental innovations in digital platforms. The model as a whole explains 31.6% of the total variance in the behavioural intention to adopt. Out of the four variables in the UTAUT model, social influence (SI) is seen to have no significant relationship with that of behavioural intention. However, the other three factors, namely performance expectancy (PE), effort expectancy (EE) and facilitating conditions (FC), do have a significant impact on the behavioural intention to adopt incremental innovations. FC has the greatest influence on the behavioural intention, followed by PE and EE. But important factors like age, gender, internet experience and voluntariness to use that are considered moderators are not seen to impact this aspect. Research focussing on radical innovations is quite abundant in past literature; however, those focusing on incremental innovations are quite scarce. This research, focussing on the adoption of incremental innovation in digital platforms in a developing country with a digital divide situation, is indeed a novel attempt.

JEL Classification: O1, O2, O320

Keywords

Adoption Digital Incremental Innovations Platforms

Introduction

Digital innovations are the modern-day take on implementing advertising and marketing strategies into a traditional brick and mortar store. With the advent of novel technologies, the spectrum of digital innovations has conquered new heights. From digital prescriptions by doctors to digital tours organised for a particular place, the list of possibilities has only seen a rise in recent times. Innovation has always acted as a powerful business booster for marketers in both the manufacturing and services sectors. Agostini et al. (2020) defined the concept of digital innovation as the creation of or changes in the offerings in the market, business processes, or models that occur due to the use of digital technologies. Digital innovation is defined as the combination of digital and physical components to produce new products or services to provide diverse services that dissolve product and industry boundaries (Yoo et al., 2010). Digital innovation does not only imply a shift in technology but also in the dynamics of the relationship between business and markets (Akesson & Akram, 2011).

The adoption of technology among the respondents has been thoroughly researched in the past. There are various technology acceptance models (TAM) that have been formulated in the past in order to determine the factors that influence the adoption of various technologies. There have been factors such as perceived ease of use and perceived usefulness put forward in the TAM , the factors of subjective norm and attitude being part of theories such as the theory of planned behaviour and the factors of performance expectancy, social influence, effort expectancy and facilitating conditions in the unified theory of acceptance and use of technology (UTAUT) model. These models have been established as prominent ones in this aspect and have been used by researchers for the adoption of various technological advancements . In this study, one remarkable transformation in the digital innovation process that has been made with the aid of updated technology is the rate at which these are being introduced into the market is being investigated. Thus, the scope of incremental innovations on digital platforms has increased. In a developing country like India, it becomes quite difficult for these incremental innovations to be adopted simultaneously with the pace at which they are being created. The internet penetration in the country has just started to rise in recent times, and due to the pre-existing situation of a digital divide in the nation, the adoption has become a bit restricted. Moreover, research focusing on these digital innovations of an incremental nature has not received much importance. This study is an attempt to study the adoption of such incremental innovations in digital platforms in a developing nation with the condition of a digital divide in play.

Literature Review

Incremental Innovations

Incremental innovations, as mentioned, are unlike radical innovations; they are simpler and take place in the form of improvements or modifications to the existing structure. Sundbo (1997) defined incremental innovations very well as innovations which are new to the market. Chan et al. (1998) defined incremental innovations as improvements in the process that do not require the aid of new technologies. Hurmelinna-Laukkanenet al. (2008) mentioned, citing Garcia and Roger (2002) that radical innovations are responsible for creating both marketing and technological discontinuities at a macro and micro level. But incremental innovations are relatively smaller levels of innovation where certain changes or modifications are made to the already existing service at a lower rate of risk. Kirzner (1997), in his theory of entrepreneurship, mentioned that incremental innovations have the chance of allowing the entrepreneur to achieve higher profits by using the same base product in a new arrangement. According to Pisano (2014), the profits generated by a firm are usually the most from the routine set of innovations. Pappenheim (2016) mentions how companies nowadays are using incremental innovations, but researchers are still focussed on working on radical innovations. Simple random sampling is used to select individuals for an interview in central Bangkok who were approached both in person and on internet platforms. The results showed that customers do tend to have a positive response towards the idea of incremental innovations. However, the success rate of an incremental innovation increases whenever it is perceived as a radical innovation by its users. Moreover, it is seen that consumers do not always expect radical innovations; they welcome incremental innovations in the form of product development in a similar manner. Thus, the literature involving incremental innovations is still scanty, and this is an attempt to dwell further to determine the factors influencing the adoption of incremental innovations in digital platforms.

Rautela and Mavale (2016), talking about technological interventions in the innovation process, said that it has become impossible for any firm in any industry to ignore its online presence, especially for restaurant businesses. The firms that have employed technological incremental innovations have flourished in terms of handling food as well as attracting and retaining consumers. This has enabled the firms to be aggressive in marketing, which provides much convenience to them and also to the consumers. Incremental innovations with the aid of technology have enabled firms to perform better and, hence, can be predicted to last a long time. Therefore, identifying the factors that lead to its adoption amongst the masses becomes an important lookout. To continue with the study, the TAM of the UTAUT has been used as a conceptual framework. The next section explains in detail the origin of the model and why it is considered the best possible choice for conducting the study.

Unified Theory of Acceptance and Use of Technology

TAM have been widely used in the literature to determine the adoption of technologies by both individuals and organisations. Venkatesh et al. (2003) formulated a model by integrating eight previous TAMs to incorporate a wholesome aspect for the same and named it as UTAUT. The eight models combined to form the UTAUT model include theory of reasoned action (TRA), TAM, motivational model (MM), TPB, combined TAM and TPB, model of PC utilisation (MPCU), innovation diffusion theory (IDT) and social cognitive theory (SCT).

For validating the UTAUT, a longitudinal field study with 215 individuals at four organisations who were introduced to new technology at their workplace responded to a questionnaire at three different points in time—post their training of the technology, one month after the implementation of it, and three months after the implementation. The partial least squares method has been used to measure the reliability and validity of the scale. A total of 48 tests, including two studies (study 1 involuntary setting and study 2 in mandatory settings), were run in eight models at three time periods to test the convergent and discriminant validity. The reliability scores were found to be more than 0.70. The eight models were found to explain the variance of 17%–42%. Table 1 shows all the constructs that have been considered while formulating the UTAUT model. All these factors have been analysed to finally create the variables that sum up the entire eight models and help researchers find out the adoption behaviour using a single effective model. The factors from the eight models have been reduced into the following four: performance expectancy (PE), effort expectancy (EE), social influence (SI) and facilitating conditions (FC). These are the direct determinants of usage and intention, and let us look in detail at the characteristics of each of them.

Table 1.

List of Variables in the UTAUT Model.

Variable	Definition	Variables Used from Previous Models
Performance expectancy (PE)	The degree to which an individual believes that he/she would gain in job performance by using the system	PU from TAM/C-TAM-TPB, extrinsic motivation from MM, job-fit construct from MPCU, RELATIVE ADVANTAGE FROM IDT and outcome expectations from SCT
Effort expectancy (EE)	Degree of ease linked with the use of the system	PEOU from TAM, Complexity from MPCU and ease of use from IDT
Social influence (SI)	The degree to which an individual perceives that the closest one to him/her in the social system approves of him/her using the system	SI is a subjective norm from TRA, TPB and C-TAM-TPB, social factors in MPCU and image in IDT
Facilitating conditions (FC)	The degree to which an individual believes that there are organisational and technical support structures that exist to use the System	Perceived behavioural control from TPB, C-TAM-TPB, facilitating conditions from MPCU and compatibility from IDT

Source: The authors.

Apart from the four factors that determine the behavioural intention to adopt an innovation, the relationships in the model are also seen to be affected by the four moderators—age, gender, experience and voluntariness to use.

The model thus formulated can be depicted diagrammatically in Figure 1.

Figure 1.

Source: Venkatesh et al. (2003).

Over time, numerous researchers have focussed specifically on the UTAUT model to determine the adoption of innovations. Below are some of the attempts made in various fields.

Gupta et al. (2017) used the UTAUT model with an added factor of perceived credibility (PC) to find out the behavioural intention to adopt the payment bank services in India, especially among the underbanked population of the country. Data collection has been done in New Delhi from 660 respondents through convenience sampling, who are either small businessmen or migrant labourers and are underbanked. Structural equation modeling (SEM) is used to analyse the data collected along with measuring the moderating and mediating effects of a few of these constructs, and the effect of the variable FC is being directly measured on the BI of the customers who adopt these payment banks. The model reported PC as the strongest influencing construct on behavioural intention. The influence of PE and SI on that of PC does exist; with the effect of SI on PC being greater than that of PE. It is interpreted from this result that when one uses the payment bank services, he/she tends to attain a positive image in his/her social circle. Moreover, PC mediates the relationship of the UTAUT model variables’ PE and SI with that of BI. Further, the relationship between PE and BI is moderated by the other two variables in the UTAUT model—FC and EE.

Mansoori (2017) conducted a study involving the modification of the UTAUT model to determine the adoption of e-government services in Abu Dhabi. Two new constructs of Government trust and Internet trust are added to the original framework of UTAUT and a structured questionnaire involving 41 questions was used for collecting primary data through email. Responses from 638 respondents were finally considered for the study and were analysed using regression and SEM. The factors of PE, EE, FC, trust in the internet and trust in government showed a positive influence on BI. FC, however, did not show any influence on use. Gender moderated only the relationship between EE and BI; age moderated the relationship between PE and BI; and experience moderated the relationship between FC and BI. The study concludes with the thought that it is the satisfaction of the users in operating the e-government services that most influences them to realise the benefits arising from using these services.

Sorwar and Hoque (2014) used the UTAUT to develop a model determining the important factors in the intention to adopt and use mHealth services among elderly people. A structured questionnaire was used to collect data from 300 respondents who are over the age of 60 in Bangladesh. To analyse the data collected, the partial least squares method in SEM was used. The factors that were found to significantly impact the adoption and use are PE, EE, SI, technology anxiety and resistance to change. The study confirms the positive application of UTAUT in the field of mHealth services and presents valuable input to influence the potential users.

Alshehri (2013) used the UTAUT model as a conceptual framework to determine the adoption of e-government services in Saudi Arabia. Six hundred and eighty five primary responses were collected using a structured questionnaire for the first part of the triangulation method employed. The data collected was analysed using SEM using the software Analysis of Moment Structures (AMOS) version 19.0. The model is refined by adding the constructs of trust and website quality (WQ) along with PE, EE, SI and FC to measure BI and actual use in the presence of the moderators—age, gender, experience and voluntariness. The second phase of data collection involved qualitative focus group interviews from two groups, each having five participants. In this study, the factors of trust, PE, EE, FC and WQ showed a significant effect on BI and usage, while SI did not have significant results in either. BI influences the use of e-government services while the moderators’ age, gender and internet experience impact the actual usage.

Summarising the literature discussed above, it is seen that the UTAUT model emerged as the product of integrating eight different theories to measure the adoption behaviour of technological innovations. It records high reliability and validity scores and has been extensively used in the past for determining the influence of various factors on adopting technological innovations. However, the use of UTAUT in determining the adoption of incremental service innovations on digital platforms has not been addressed so far. As per Nambisan et al. (2017), going for interdisciplinary research in information technology is important to encourage theory development, therefore confirming the capabilities of UTAUT, which is used as a conceptual framework to understand this amalgamation of technology and innovation better.

Accordingly, this study lays down the following objective: ‘To determine the factors responsible for the adoption of incremental service innovations in digital platforms’.

The four constructs of the model need to be tested to see whether all of them influence the behavioural intention to adopt these frequent modifications happening all over the digital space. Although the original model measures the actual use of the innovation in question; here, as the concept is still quite novel among the people of India, we are only measuring the behavioural intention to adopt it and not its actual usage. The effect of facilitating conditions, which in the model is measured on actual use here, is hypothesised to have a relationship with behavioural intention as well. Thus, the hypotheses formulated for the study are as follows:

H_0a: PE doesn’t influence the behavioural intention to adopt incremental service innovations in digital platforms

H_0b: EE is not related to the behavioural intention to adopt incremental service innovations in digital platforms

H_0c: The SI on a user will not be related to the adoption of incremental service innovations in digital platforms

H_0d: FC of the user will not impact his/ her behavioural intention to adopt incremental service innovations in digital platforms

Research Method

Measuring Instrument

The research took a quantitative approach as empirical evidence is the basis for analysing the results. The UTAUT model has been used by researchers over time, however, for the study, the standard set of questions given by Venkatesh et al. (2003) have been adapted. The questions have been modified as per the requirements of the study. The questionnaire has been reviewed by experts in the field, and the modifications observed when conducting a pilot survey with 50 data sets have also been incorporated before the final sample collection procedure. A total of 16 items on a 5-point Likert scale have been incorporated to measure the variables in the study.

Participants and Procedure

One essential criterion for selecting the respondents includes their previous experience of availing of services online. Hence, a non-probabilistic method with judgement sampling and convenience sampling has been used to arrive at the best quality of responses. For the study, the geographic location of Assam has been specified. As the state corresponds to one of the highest digital users in the entire north-eastern region and has shown high growth in the past few years, it aptly represents the situation prevailing in India with respect to the use of digital platforms. For selecting the districts, the Employment and Livelihood Quality Index (ELQI) put forward by the OKD Institute of Social Change and Development (2014) has been used. The index includes parameters such as per capita income, number of individuals working in non-agriculture-based jobs, etc., which are found to have an impact on the internet usage pattern. The top four districts in the index have been used for collecting data. The districts are—Kamrup Metropolitan, Nalbari, Jorhat and Sibsagar. A proportional data collection system has been employed in each of the districts.

A total of 1,000 questionnaires have been distributed to the effective population, from which a total of 887 have been received back. However, careful examinations of responses for missing data and incompetent responses led to the discarding of 66 responses. Thus, with a net response rate of 82.1%, a total of 821 responses were finally analysed for the study.

Data Analysis and Interpretation

The data collected was analysed using IBM Statistical Packages for Social Sciences (SPSS) and IBM Analysis of Moment Structure (AMOS) 20.0.

Demographic Representation

The participants for the study have attempted to represent the actual internet users’ profiles in the country as shown in Table 2. Keelery (2020) stated that most of the internet users in India in 2019 were in the age group of 20–29 years. Users above the age of 40 years are reported to be the least in proportion. It is interesting to note that the number of users aged 20−29 years from rural areas is slightly higher than that of urban users. Varshney et al. (2014) mentioned age as an important demographic factor for internet usage, and their study stood in line with previous literature. The samples thus represent the population of internet users in terms of age and place of residence. As an implication of laying age group as a criterion for data collection, the numbers of respondents in certain demographic categories are seen to be greater in number. For example, due to the specifications in the age group when collecting responses, the income status has been divided into two broad categories: earning and non-earning users, to better identify the difference between the two.

Table 2.

Demographic Representation.

Demographic Variable	Frequency	Percentage
Gender
Male	410	49.9
Age (years)
18–24	466	56.7
25–34	292	35.6
Above 35	63	7.7
Earning status
Earning	385	46.8
Non-Earning	436	53.2
Level of education
Graduate	303	36.9
Postgraduate	232	28.3
Undergraduate	286	34.8
Place of residence
Urban	412	50.1
Non-urban	409	49.9

Source: The authors.

Data Processing

The study has been conducted in three stages. SEM Analysis is used to arrive at the required results for the study. However, the data structure has been identified and confirmed before the SEM process through exploratory and confirmatory factor analysis (CFA). Last, the demographics of the variables would be analysed for any possible significant associations.

Table 3 represents the items included in the questionnaire with their respective mean and standard deviation values.

Table 3.

Descriptive Statistics of the Items.

Variables	Number	Items	Mean	Std. Deviation
Performance expectancy	PE1	I find the modifications made in digital platforms useful in availing my service	3.83	2.15
	PE2	I can finish my tasks more quickly using the improvements made in these platforms from time to time	3.80	0.827
	PE3	These modifications improve the quality of services offered by the service providers	3.67	0.886
	PE4	My overall experience of using digital platforms is good because of the frequent improvements made by them	3.62	0.874
Effort expectancy	EE1	I find these modifications easy to use	3.56	0.894
	EE2	The instructions about the use of these adjustments are made clear and understandable	3.52	0.866
	EE3	I can easily become an expert at using the modifications in digital platforms	3.40	0.938
Social influence	SI1	People who are important to me think that I should use the improvements made by the digital platforms	3.34	0.965
	SI2	I use them because a section of my friends/ colleagues also uses the improvements in digital platforms	3.13	1.03
	SI3	People who influence my behaviour think that I should use these modifications	3.16	0.982
Facilitating conditions	FC1	The resources (hardware, software and networks) I use daily are sufficient to use these modifications	3.44	0.933
	FC2	It is easy to get the knowledge necessary to use the modifications made in the digital platforms frequently	3.71	1.32
	FC3	I can get help from a specific person (or group) in case I face difficulties using the various improvements by digital platforms	3.64	0.924
Behavioural intention	BI1	I intend to use such modifications in digital platforms in the coming future.	3.77	0.904
	BI2	I would use the modifications to improve my digital service experience.	3.72	0.899
	BI3	I plan to use such modifications in the next month	3.36	0.894

Source: The authors.

Exploratory Factor Analysis

The process of exploratory factor analysis (EFA) involves identifying the factors corresponding to the list of items provided to the respondents. Hair et al. (2006) stated that conducting a factor analysis helps in examining the interrelationships existing among a greater number of variables. It is a process that involves testing the appropriateness of the samples using Bartlett’s test of sphericity and measuring the sampling adequacy using Kaiser–Meyer–Olkin (KMO). Bartlett’s test requires the items to be statistically significant, while any value above 0.6, as stated by Tabachnick and Fidell (2007), is considered acceptable. The results of the tests are shown in Table 4.

Table 4.

Kaiser–Meyer–Olkin—Bartlett Test Results.

Kaiser–Meyer–Olkin Measure of Sampling Adequacy.		0.914
Bartlett’s test of sphericity	Approx. χ²	4367.255
Bartlett’s test of sphericity	df	136

Source: The authors.

As the results for both the tests adhere to the laid standards, it can be stated that the sample is adequate for conducting the study. The process of EFA involves an extraction method coupled with a rotation. For this data structure, Principal Axis factoring is used as it enables the determination of latent constructs, which are often difficult to measure. The principal axis factoring is used along with the orthogonal rotation of the varimax. The varimax rotation enables the maximising of the sum of variances that influence the factor loadings. The EFA as a whole indicated the presence of five factors with an eigenvalue of greater than 1, explaining 78.9% of the variance. The results for common method bias (CMB) are analysed using Harman’s single factor score, where all the items are recorded in one common factor. As the cut-off variance for the single factor is less than 50% (Eichhorn, 2014) and the value generated from this data set is 29.83%; there is no CMB reported. Hence, with adequate sampling and a CMB-free data structure, CFA analysis is initiated.

Confirmatory Factor Analysis

The process of CFA helps in confirming the identified factors in the EFA concerning the data structure. It helps in determining the effective depiction of the factors based on the data collected for conducting the study. As explained by Hair et al. (2006), CFA helps in determining how the constructs considered in the study can be represented through a smaller number of constructs. The process of CFA involves testing the items in the questionnaire for validity and reliability. The reliability is measured through Cronbach’s alpha value; while the validity is calculated using average variance extracted (AVE).

The Cronbach’s alpha value for the mentioned constructs stands at 0.775, and as mentioned by Hair et al. (2006), any value above 0.7 is considered acceptable for the study. Moving on to the calculation of AVE, it is calculated as the sum of the squared standardised factor loadings, which is divided by the number of items, and a value of more than 0.4 for each construct is required (Fornell & Larcker, 1981). The AVE for PE, EE, SI, FC and BI stands at 0.509, 0.446, 0.442, 0.425 and 0.421, respectively. The validity and reliability of the items considered for the study are thus met.

In the next step of CFA, a measurement model for all the constructs is created. The model would enable us to detect the insignificant factors corresponding to the constructs. Moreover, the model fit can be determined using the popular model fit indices includes seven model fit indices that are found to be used extensively in the past literature to serve the purpose. The first one is the χ²/df ratio, or CMIN/df, which indicates the difference between the expected and observed covariance matrices. However, because one characteristic of CMIN/df is that it is sensitive to sample sizes (Schumacker & Lomax, 2004), other indices are used as well to determine the data fit. The second index is named the goodness-of-fit index (GFI) and is used to determine overall model fit. Then there is the comparative-fit-index (CFI), which represents the overall improvement of the proposed model over the independent model where the observed variables are not correlated (Byrne, 2006). There is Bentler and Bonett’s fit index (NFI or TLI), which determines the convergent validity of the questionnaire; another index for model fit is the relative-fit-index (RFI), which is a derivative of the NFI, while the incremental-fit-index (IFI) explains the sample size as well as parsimony issues of the NFI. Lastly, there are the two indices, namely root mean square error of approximation (RMSEA) and root mean square residual (RMR). RMSEA measures the error of approximation in the sample and is associated with the residuals in the model; on the other hand, RMR represents the value of the average residual that is derived from the fitting of the variance-covariance matrix for the model. This GFI has an acceptable range of values corresponding to which the theoretical model can be considered to have a good fit. The measurement model generated from the CFA conducted for the UTAUT model is represented in Figure 2.

Figure 2.

Source: The authors.

Table 5.

Model Fit Indices from the Measurement Model.

Model Fit Indices	Accepted Value	Value	Status
CMIN/df	<3	2.496	Good fit
GFI	>0.9	0.966	Good fit
NFI/TLI	>0.9	0.947	Good fit
IFI	>0.9	0.967	Good fit
CFI	>0.9	0.967	Good fit
RMSEA	<0.06	0.043	Good fit
RMR	<0.05	0.027	Good fit

Source: The authors.

The model fit indices for the measurement model show a good fit with a χ² value of 229.659 at a significant probability level of 0.000. A CMIN/df value of less than 3 represents a fairly good estimation of model fit even with such a large sample size.

Structural Equation Modelling

SEM allows the researcher to simultaneously determine the relationships existing among the dependent and independent variables in a study. The structural model allows a wholesome picture of the entire process. The model fit indices generated during the process help in determining how well the model is capable of determining the desired relationships. The total variance explained by the model as well as all the significant independent variables can be very well explained by this process. Thus, the structural model generated from the initial analysis is shown in Figure 3.

Figure 3.

Source: The authors.

The path coefficients generated in the model have a χ² value of 176.879 at a probability level of 0.000. The significance level of each path is mentioned in Table 6 below along with their standardised coefficients.

Table 6.

Significance of the Paths in the Structural Model 1.

Path	Standardised Coefficients	p-Value
PE ↓ BI	0.21	0.000
EE ↓ BI	0.16	0.000
SI ↓ BI	0.03	0.199
FC ↓ BI	0.28	0.000

Source: The authors.

From the estimates generated, it is seen that the path from SI to BI is insignificant and hence we move on to refining the model by removing the insignificant path. Out of the four hypotheses formulated, it is only H_0c that cannot thus be rejected.

The model is hereby refined by removing the variable corresponding to social influence. Once the model is refined, the analysis moves on to determine the model fit indices along with the estimates of all the constructs in the model.

Table 7.

Model Fit Indices for Structural Model 2.

Model Fit Indices	Values
GFI	0.928
NFI/TLI	0.867
IFI	0.869
CFI	0.868
RMR	0.092
RMSEA	0.267

Source: The authors.

The model fit indices in Tables 5, 6 and 7 show a good fit as the GFI value scored more than 0.9. The standardised coefficients are interpreted as the amount of variance in the dependent variable for each unit change in the independent variable. BI corresponds to a 21.8% change in itself for a unit change in the PE. Similarly, a unit change in EE causes 0.166% or 16.6% variance in BI in the respondents and a 28.8% variance due to FC while using the incremental service innovations. Also, the squared multiple correlation values for BI are estimated at 0.316, which indicates that the predictors of BI account for 31.6% of the variance in behavioural intention. The model thus, as a whole, explains 31.6% variance in the behavioural intention of adopting incremental innovations in digital platforms.

An interesting finding to note here is that the moderators for the study, namely—age, gender, voluntariness to use and internet experience, have not showcased any kind of moderating effects on the relationships amongst these variables. The process has been conducted using multiple linear regression and no significant findings can thus be reported from the same.

The structural model thus formulated from the UTAUT model adopted for determining the adoption of incremental innovations in digital platforms can be represented simply as:

Figure 4.

Source: The authors.

Demographic Relationships

As evident from Figure 4, the four moderators have not shown any effect on the relationships among the three independent variables on the behavioural intention to adopt. However, it would be interesting to note if any of these five variables considered in the study were influenced by the demographics of the respondents recorded through the primary data collection. One-way ANOVA and t-tests would be used for this purpose.

Table 8.

Summary of One-Way ANOVA.

Dependent Variable	Age		Education Level
Dependent Variable	χ ²	p	χ ²	p
Performance expectancy	10.25	0.001	2.51	0.285
Effort expectancy	9.00	0.003	2.10	0.350
Social influence	5.11	0.024	12.2	0.002
Facilitating conditions	6.97	0.008	0.07	0.964
Behavioural intention	5.99	0.014	2.50	0.285

Source: The authors.

Table 9.

Summary of t-Tests.

Dependent Variable	Gender			Earning Status			Place of Residence
Dependent Variable	Statistic	df.	p	Statistic	df.	p	Statistic	df.	p
Performance expectancy	0.437	819	0.662	−0.630	819	0.529	1.359	819	0.175
Effort expectancy	0.583	819	0.56	−0.219	819	0.002	−0.378	819	0.70
Social influence	0.624	819	0.533	−0.643	819	0.52	−1.570	819	0.117
Facilitating conditions	−0.917	819	0.844	−2.29	819	0.027	−1.573	819	0.11
Behavioural intention	−2.570	819	0.010	−4.25	819	<0.001	−0.060	819	0.947

Source: The authors.

Tables 8 and 9 show the results from the demographic analysis done across the five variables in the UTAUT model. The place of residence of the respondent did not show any statistically significant association with the variables considered. However, for the other variables, significant mean score differences are observed. The mean scores are significantly different when it comes to the variable of behavioural intention to adopt incremental innovation for gender, earning status and the age group of the respondents. For earning status, the mean score differences for the respondents having an earning of their own and those who do not earn exist in terms of EE and FC. The difference in the mean score of responses based on age is significant in terms of all five variables. Based on the education level of the respondents’ significant differences in terms of SI are observed.

To determine the exact levels at which the mean differences are significant, a pairwise post hoc analysis was conducted for the age group and education levels of respondents.

Table 10.

Pairwise Comparisons—SI.

		W	p
Undergraduate	Graduate	4.78	0.002
Undergraduate	Postgraduate	1.27	0.641
Graduate	Postgraduate	−3.35	0.047

Source: The authors.

As shown in Table 10, considering the demographics of education level and the variable of social influence, it is among the graduate and postgraduate respondents that a significant mean difference is observed. However, for every variable in the age group, it is between the age groups of above and below 35 years that significant mean differences are observed.

Results and Discussion

The key objective of the study was to identify the factors that influence digital users’ adoption of incremental service innovations by using the UTAUT model as the conceptual framework. The analysis of the study is conducted in three stages, where in the first stage, the items considered in the questionnaire are identified and confirmed with the data structure using exploratory and CFA. In the next stage, a model is constructed using SEM to determine the amount of variance explained by the significant factors at hand. Finally, the five standard variables of the UTAUT model are used to find demographic associations. There have been several revelations made regarding the adoption of incremental innovations in digital platforms. The model highlighted the insignificant role of social influence in determining the behavioural intention to adopt. This can be interpreted as a positive indication as digital users who are inclined towards the adoption of incremental innovations are not dependent on their social support groups. Digital users indeed look for factors based on performance measurement, the effort required to use them, and the facilities supporting their use at each level. The variables are further not seen to have any significant difference regarding the adoption of incremental innovations based on their place of residence. This might be a hint towards the reducing digital divide among the urban and non-urban areas in India. With internet penetration levels growing in the country, awareness of such frequent innovations can be expected to further grow in the future.

Theoretical Implications

The CFA conducted assured the model fit of the data; the validity and reliability were also attained. The initial structural model revealed that out of the four constructs: PE, EE, SI and FC, SI does not have a significant impact on the behavioural intention to adopt an incremental innovation. This signifies that the user of a digital platform is not influenced by his/ her social support system in adopting such an innovation. Peer pressure or conditions of social acceptance do not make a user adopt an incremental innovation. On the other hand, out of the three significant influencers, facilitating conditions correspond the most to the behavioural intention of the adoption of incremental innovation, followed by performance expectancy and effort expectancy. A total of 31.6 % variance in the behavioural intention to adopt incremental service innovations in digital platforms is estimated by the UTAUT model, that is, by the three constructs—PE, EE and FC. The four moderators in the original UTAUT model, namely—age, gender, voluntariness to use and internet experience, are tested to see if any of them has significant influence. Interestingly, none of them are seen to have any influence on the relationships.

Practical Implications

Incremental innovations are taking place in almost every digital platform, whether it is in the banking sector, wallets, travel services, or medical providers. One of the many modifications that result from incremental innovations is the change in the layout structures of the digital platforms. As in India, where the number of internet users is increasing constantly, such frequent changes in these platforms might serve as a hindrance to their digital learning experience. This study helps in identifying the factors that might help in determining the key points to focus on to help increase these innovations. With the social influence impact not being significant, it becomes apparent that it is the benefits arising from the innovations that make the users avail of the modifications. Consequently, there is a chance of the incremental innovations in digital platforms getting adopted if there are performance benefits, less effort invested in learning to operate the innovations, and enough facilities to enhance the users’ experience of using the innovations. These significant effects of the PE, EE and FC impact on the behavioural intention to use incremental innovations in digital platforms. Focus on the age group of the users designing the incremental innovations can help marketers receive their share of the market amongst the age group above 35 years as well.

Conclusion and Recommendation

The main aim of the study has been to identify the factors that influence the impact on the adoption of incremental innovations. As the world is moving towards a completely digitised state, it is important to understand the adoption of an integral aspect of it, that is, incremental innovations. The study clearly highlights the role played by the utility-based parameters, such as performance measures, effort required and the support services to use these innovations. The lack of impact of social influence on adoption shows that users would be willing to adopt these innovations only when they are optimum for their daily uses.

Hence, the focus of the marketers and policy makers must be to highlight the benefits of adopting these innovations rather than the upgradation of oneself in the modern world through their implementation. Further, in terms of research in the future, the UTAUT model is used in the case of this study; however, it is important to investigate the robustness of the UTAUT model in the future regarding the adoption of incremental innovations in different aspects and cultures. Further, adding relative factors such as trust and other psychological factors into the model could throw additional insights into the topic.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Dorothy Dutta

References

Agostini

L.,

Galati

, & Gastaldi.

(2020). The digitalization of the innovation process: Challenges and opportunities from a management perspective. European Journal of Innovation Management , 23(1), 1–12. https://doi.org/10.1108/EJIM-11-2019-0330

Ajzen

(1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes , 50(2). https://doi.org/10.1016/0749-5978(91)90020-T

Akesson

, & Akram

(2011). Value network transformation by digital service innovation in the vehicle industry. PACIS 2011 Proceedings . https://aisel.aisnet.org/pacis2011/12/

Alshehri

M. A.

(2013). Using the UTAUT model to determine factors affecting acceptance and use of e-government services in the kingdom of Saudi Arabia . Griffith Research Online.

Arain

A. A.,

Hussain

Z.,

Rizvi

W. H.

, & Vighio

M. S.

(2019). Extending UTAUT2 toward acceptance of mobile learning in the context of higher education. Universal Access in the Information Society , 18(3), 659–673. https://doi.org/10.1007/s10209-019-00685-8

Byrne

(2006). In search of a ‘Good Mix’: ‘Race’, class, gender and practices of mothering. Sociology , 40(6), 1001–1017. https://doi.org/10.1177/0038038506069841

Chan

A.,

F. M.

, & Pine

(1998). Service innovation in Hong Kong: Attitudes and practice. Service Industries Journal , 18(2), 112–124. https://doi.org/10.1080/02642069800000021

Davis

F. D.

(1989). Perceived usefulness, perceived ease of use, and user acceptance of information. MIS Quarterly , 13(3), 319–340. https://www-jstor-org-443.web.bisu.edu.cn/stable/249008

Djafarova

, & Bowes

(2021). ‘Instagram made Me buy it’: Generation Z impulse purchases in fashion industry. Journal of Retailing and Consumer Services , 59. https://doi.org/10.1016/j.jretconser.2020.102345

10.

Eichhorn

B. R.

(2014). Common method variance techniques . Cleveland State University. https://www.lexjansen.com/mwsug/2014/AA/MWSUG-2014-AA11.pdf

11.

Fornell

, & Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research , 18(1), 39.

12.

Garcia

, & Roger

(2002). A critical look at technological innovation typology and innovativeness terminology: A literature review. The Journal of Product Innovation Management , 19(2), 110–132. https://doi.org/10.1111/1540-5885.1920110

13.

Gunasinghe

A.,

Hamid

J. A.,

Khatibi

, & Azam

S. M. F.

(2019). The adequacy of UTAUT-3 in interpreting academician’s adoption to e-Learning in higher education environments. Interactive Technology and Smart Education , 17(1), 86–106. https://doi.org/10.1108/ITSE-05-2019-0020

14.

Gupta

K. P.,

Manrai

, & Goel

(2017). Factors influencing adoption of payments banks by Indian customers: Extending UTAUT with perceived credibility. Journal of Asia Business Studies , 1(2), 173–195. https://doi.org/10.1108/JABS-07-2017-0111

15.

Hair

J. F.

Black

Babin

Anderson

Tatham

(2006). Multivariate data analysis (6th Ed.). Pearson Education. https://doi.org/10.1080/19447013008687143

16.

Hurmelinna-Laukkanen

P.,

Sainio

L.-M.

, & Jauhiainen

(2008). Appropriability regime for radical and incremental innovations. R&D Management , 38(3), 278–289. https://doi.org/10.1111/j.1467-9310.2008.00512.x-i1

17.

Keelery

(2020, July 7). Internet usage in India—Statistics and facts . Statista. https://www.statista.com/topics/2157/internet-usage-in-india/

18.

Kirzner

I. M.

(1997). Entrepreneurial discovery and the competitive market process: An Austrian approach. Journal of Economic Literature , 35(1), 60–85.

19.

Mansoori

K. A.

(2017). Use of a modified UTAUT model to investigate Emirati Citizens’ adoption of e-Government in Abu Dhabi . University of Wollongong Research Online. http://ro.uow.edu.au/theses1/20

20.

Nambisan

, Lyytinen

, Majchrzak

, & Song

(2017). Digital innovation management: Reinventing innovation management research in a digital world. MIS Quarterly , 41(1), 223–238.

21.

OKD Institute of Social Change and Development. (2014). Assam Human Development Report . Guwahati: Planning and Development Department Government of Assam.

22.

Pappenheim

B. C.

(2016). Incremental innovation effects on consumer perception. Science Journal of Business Management . https://doi.org/10.7237/sjbm/272

23.

Pisano

G. P.

(2014, June 10). In defense of routine innovation . Harvard Business Review. https://hbr.org/2014/06/in-defense-of-routine-innovation

24.

Rautela

, & Mavale

(2016). Incremental innovation for sustainable growth in restaurant businesses: Global practices for the growth of local business. Annual Research Journal of SCMS , 4, 80–87.

25.

Schumacker

R.E.

, & Lomax

R. G.

(2004). A beginner’s guide to structural equation modeling (2nd Ed.). Lawrence Erlbaum Associates.

26.

Sorwar

, & Hoque

(2014). Understanding factors influencing the adoption of mHealth by the elderly: An extension of the UTAUT model. International Journal of Medical Informatics , 101, 75–84. https://doi.org/10.1016/j.ijmedinf.2017.02.002

27.

Sundbo

(1997). Management of innovation in services. The Service Industries Journal , 17(3), 432–455.

28.

Tabachnick

B. G.

, & Fidell

L. S.

(2007). Using multivarite statistics (5th Ed.). Pearson Education.

29.

Varshney

B.,

Kumar

P.,

Sapre

, & Varshney

(2014). Demographic profile of the internet-using population of India. Management and Labour Studies , 39(4), 423–437. https://doi.org/10.1177/0258042X15578018

30.

Venkatesh

V.,

Morris

M. G.,

Davis

G. B.

, & Davis

F. D.

(2003). User acceptance of information technology: Toward a unified view. MIS Quarterly , 27(3), 425–478. https://doi.org/10.2307/30036540

31.

Yoo

, Henfridsson

, & Lyytinen

(2010). The new organizing logic of digital innovation: An agenda for information systems research. Information Systems Research , 21(4), 724–735. https://doi.org/10.1287/isre.1100.0322