Research: Digital POS Adoption · Rafif Dzaky Daniswara

The question

Most MSME digitalization research looks outward: e-commerce, social media, QRIS and digital payments. A point-of-sale app is different. It's an internal tool for recording transactions, tracking stock and reading sales. This pilot asks what drives MSMEs in Bandung Raya to adopt one. It applies UTAUT2 to that internal setting, and it checks whether the survey instrument actually measures what it claims to before any large survey is run.

Sample and method

50 MSME practitioners with current or past POS use: 31 in Kota Bandung, 19 in Kabupaten Bandung. 64% owners, 36% staff, ages 19–60. Most used Moka POS.
A 37-item questionnaire adapted from prior UTAUT2 studies. It used a 4-point Likert scale, so there was no neutral midpoint to hide in.
Nine constructs: performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value, habit, behavioral intention and use behavior. That gives ten hypothesized paths.
PLS-SEM in SmartPLS 4.0. The focus is the measurement model: indicator reliability, internal consistency and convergent validity. The structural model is read only descriptively.

Refining the instrument

Four items fell below the 0.708 outer-loading threshold. Each also fit the POS setting poorly, so they were removed:

Item	Loading	Why it didn't fit
`PE4`	0.707	fewer calculation errors reads as a basic expectation, not a performance gain
`SI1`	0.678	broad social pressure matters less than business relevance to owners
`PV2`	0.694	overlapped with the other value items
`HM3`	0.625	an entertainment angle that functional, efficiency-driven MSMEs don’t weigh

After re-estimating with the 33 remaining items, every loading clears 0.708. Every construct has Cronbach's alpha and composite reliability above 0.70 and AVE above 0.50.

Refined construct reliability (Table V)

Construct	Cronbach's α	ρc	AVE
Performance Expectancy	0.850	0.899	0.690
Effort Expectancy	0.839	0.892	0.674
Social Influence	0.708	0.836	0.631
Facilitating Conditions	0.817	0.879	0.644
Hedonic Motivation	0.728	0.847	0.649
Price Value	0.802	0.884	0.718
Habit	0.853	0.901	0.695
Behavioral Intention	0.815	0.879	0.646
Use Behavior	0.842	0.894	0.677

What the model shows

The refined model explains 84.4% of the variance in behavioral intention and54.8% in actual use. Three factors stand out on effect size:

Path	Relationship	f²
SI → BI	Social influence → intention	0.204
HM → BI	Hedonic motivation → intention	0.164
PV → BI	Price value → intention	0.145
BI → UB	Intention → use	0.113
HT → UB	Habit → use	0.046
EE → BI	Effort expectancy → intention	0.039
FC → UB	Facilitating conditions → use	0.018
PE → BI	Performance expectancy → intention	0.014
FC → BI	Facilitating conditions → intention	0.001
HT → BI	Habit → intention	0.001

Social influence (0.204), hedonic motivation (0.164) and price value (0.145) had the largest effects on intention. Performance expectancy, effort expectancy, facilitating conditions and habit had small ones.

What it suggests

One reading is Herzberg's hygiene–motivator split. In a maturing POS market, "it works and it's easy" is already expected. What still separates adopters is what peers use, whether using it feels good, and whether the price feels fair.

For POS providers and local digitalization programs, the tentative direction is: lean on social proof (peer testimonials, referrals), make onboarding feel easy and rewarding, and communicate costs and savings concretely instead of making broad productivity claims.

Limits

A pilot with 50 respondents. The structural paths are described, not formally tested.
Discriminant validity wasn't assessed, because HTMT is unstable at this sample size.
No demographic moderators (age, gender, experience). Those belong in the full-scale study.

Why it's on a DevOps portfolio

It's a different kind of evidence: framing a question, checking the measurement before trusting the results, and explaining the findings clearly enough to win Best Presenter.

Explaining Digital POS Adoption Among MSMEs: A Pilot Study Using UTAUT2

Docs · the paper, in short