Discussion

---

The underlying premise of this study was that, among experienced users of a specific IT, habitual previous IT usage would increase their intended continued use of the IT in addition to the behavioral beliefs, PU and PEOU, identified by TAM. This addition to TAM highlights the contribution of this study, namely, that while continued use of an IT does depend, as shown by previous research (Bhattacherjee, 2001), on a continued reassessment of its usefulness and ease of use, it also depends on habit which is devoid of a rational assessment. This premise was supported. Habitual previous preference to use a specific IT directly and strongly increased user intentions to continue using the same IT again. The data show that habitual previous IT usage increases intentions for continued use also indirectly, through increased perceptions of the ease of use and usefulness of the IT.

These conclusions extend the current theory base of TAM. The addition of habit is important because it shows that while behavioral intentions are in part the product of a rational assessment, as TRA suggests and as TAM shows, behavioral intentions are also the product of habitual previous usage. In this case, people chose to continue using an IT not only because of their beliefs about the desirable outcomes of doing so, but also directly because that specific IT was the one they were used to using. This implies that past behavior, and not only current rational assessments, directly determine behavioral intentions. Compared with previous TAM research, this is an interesting conclusion because it shows that, in contrast to the underlying assumptions of TAM, it is not only reasoned action leads directly to behavioral intentions.

So what do the results mean in the context of TAM and TRA? Habit is not an aspect of either of these two theories. On the contrary, both allude to a reasoned assessment leading to behavioral intentions. While the data support this thesis, showing that PU and PEOU do increase intentions for continued use, the data also show that there is a significant element of habit behind behavioral intentions, habit that in itself explains 40% of the variance of intended behavior. In this respect, habit affects behavioral intentions also beyond its effect on the rational antecedents, PU and PEOU, lending support to the assertion that human behavior may be rational, as TRA postulates, but that this is not the whole story. Habit is an additional force that increases the behavioral intentions to continue using a technology once its use has become habitual.

The study also highlights some aspects of Shapiro & Varian's theory (1999), whereby information technology companies succeed among other things by locking in customers through high switching costs but also through the mental effort needed to learn a new IT. This mental investment in an IT can make the users entrenched in using that IT, even when possibly better IT are available. This process creates a positive information feedback loop in which a good product or technology standard gains more and more market share, not because they are the best around but because of the cognitive and other switching costs the customers have: in other words, because the clients are hooked on their habits. Supporting these propositions, this study shows that user intentions to continue using an e-vendor depends also on habit, that is on entrenched behavior and, by implication, on the avoidance of the mental effort involved in learning how to engage with a new website.

Before discussing the implications, it is necessary to note that the data were collected from experienced users in the context of using an IT for B2C e-commerce purchasing. Whether this is a special type of IT and whether these users are typical of other IT users, and so whether the results can be generalized to other IT, require additional study. Having said that, the data imply some interesting guidelines regarding IT adoption management in general, and e-commerce adoption management in particular. Assuming the model can be generalized, the data show that a central aspect of an IT that makes it appealing to users beyond its perceived usefulness and ease of use is simply their habitual use of it. Another limitation is that habit was conceptualized as dealing with overall beneficial activities and trustworthy vendors. In these cases habit deals with the inclination to automatically do what has always been done, without recourse to necessarily recalculating the benefits, i.e., the PU and PEOU of the IT. Extrapolating from research on continued use of an IT (Bhattacherjee, 2001), however, suggests that when the habit is no longer beneficial the user may, by reconsidering the usefulness and ease of use of the IT, change their behavior. That scenario was beyond the scope of this study and requires additional research. Additional research could also examine the role of habit in other scenarios, such as when there is a lack of free choice or alternatives.

These results suggest that there might be a barrier to entry (Porter, 1980) facing new e-commerce entrants, or any other organization attempting to introduce a new IT, because of user habit to return to the same IT, in this case website, that they are used to for other purposes and activities; this, beyond their rational assessments for doing so. In the specific case of e-commerce, it is possible to speculate that this barrier can be reduced through promotions, sales, and special coupons to attract new customers, much as it is done with other stores. Once customers start coming and get used to the site, they will, suggests the model, be inclined to continue coming because of habit. In the broader case of adopting a new IT, the model implies that resistance to the adoption of a new IT may be in part due to habitual usage with a previous system. If this holds across IT, then from a managerial point of view one way of increasing user acceptance of an IT, whether directly or through increased perceived usefulness and ease of use, is to get the users into the habit of using it. This also means, however, that replacing an old IT might be harder than otherwise envisioned because it is not enough just to rely on the users' rational assessment that the new IT is more useful and easier to use. Rather, it is necessary also to overcome the force of habit, which this study has shown to be of comparable strength to that of the perceived usefulness and ease of use of the IT.

In conclusion, the model suggests, and the data verify, that while it is possibly undeniable that behavioral intentions in general, and IT acceptance in particular, are the product of people's rational analysis of the outcomes, it is also true, as the data show, that at the same time the force of habit is a major factor behind behavioral intentions and should be included in models such as TAM.

---