Humans are unable to effectively handle machines that have poor interface and interaction patterns. The Automated Teller Machine (ATM) is one of the machines that is most frequently operated by a wide spectrum of people. Hence, it is common to apply Human-Computer Interaction (HCI) technology in the design of ATMs. User-ATM interaction, however, has barriers such as a user's working memory limit, lower literacy level, or lack of awareness and access. Moreover, the diversity of interfaces hurt user motivation, perception, and experience significantly. For example, each bank has its own ATM design targeting its customers only and, thus, when clients of other banks come across such a machine, they suffer from cognitive burden and commit many errors. To this end, we conduct an analytical survey of ATMs used in Pakistan through user analysis and heuristics analysis. We figure out the influence of design diversity and user perception on ATM users' cognitive stress with the cognitive walkthrough analysis on the survey. We also uncover a realism-usability gap by reviewing the sampled ATM screens from the viewpoint of HCI heuristics. Finally, we suggest the guidelines on the cross-ATM design, characterized by the standardized fast-cash menu and the fail-safe mechanism for novices. This study demonstrates that users are reluctant to switch out of their comfort zones, due to learning costs and a lack of motivation. Therefore, we should design public systems in a very standardized way, which becomes more crucial for a user's cognitive relief as interface and design patterns are becoming more diverse in our daily lives.