With interactive technologies increasingly embedding into everyday life and spreading through urban space, many researchers in Human-Computer Interaction (HCI) are leaving their labs to perform experiments in real-life contexts to uncover design probes 1. In this context, the main approach adopted is Research In-the-wild (RITW), which explores current practices and behavior change by introducing novel technologies in situ 2. Although this approach has brought several benefits to HCI research, this movement has also increased the number of studies deprived of an explicit methodological foundation 3. Furthermore, the complexity of performing such in situ user studies, raised the question: Is it worth the hassle? 4. For these reasons, we are concerned in presenting a systematic methodology to support the rationale when conducting an in-the-wild study.Any research project requires a variety of important decisions, from the method and theoretical framing to the minutiae of operational detail. However, performing an in-the-wild study involves even more decisions with the arising of unforeseen events needing to be resolved as the research progresses 5. These issues and decisions are frequently omitted, as the researchers neglect to document their rationale 6. With in-the-wild studies, researchers can obtain insights and formulate new hypotheses by observing users using and adopting interactive technologies in situ, while eliciting new system requirements and general design implications for new versions of the prototype 1. Nonetheless, evaluating these insights often demand changes in the prototype, planning and conducting other experiments, which consume time and resources. Hence, those studies may leave unexplored issues and open questions, affecting the reliability of the inferences 7. To address these issues, we describe and discuss a mixed and systematic approach, named as On-the-fly Prototyping, for learning as much as possible with an evolutionary prototype and ensuring traceability of the changes and their impacts. On-the-fly prototyping approach shrinks the iteration cycle by narrowing the research focus. Thus, researchers can evaluate several insights or design variations during the investigation. This method extends the in-the-wild approach by including aspects of evolutionary design and allowing designers to improve the prototype as they learn from the evaluation in situ.In this paper, we discuss our lessons learned from a case study with a situated interactive public installation developed following the on-the-fly prototyping approach. From that, we indicate in which directions evolutionary design might be developed to accommodate and support in-the-wild studies, considering the design requirements for public interactive systems. The overall structure of this paper takes the form of six sections, including an introduction to the challenges in in-the-wild studies and related work, the description of the on-the-fly prototyping methodology, the report of a study case, the discussion of the lessons learned for adopting our methodology to design interactive systems for public spaces, and finally the conclusion with suggestions for future studies.