Privacy Study
This user study was exploratory, with the purpose of investigating the product opportunity gap in privacy, as it was at the time the most unexplored area for the product team (when the study was conducted, in winter 2024/2025).
Annika will also present this work as a poster at ACM CHI 2026 on Tuesday, April 14. Here you can find the paper: Annika Sabrina Schulz and Paulus Schoutsen. 2026. Building a Private Smart Home: User Motivations and Challenges. In Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems (CHI EA’26), April 13–17, 2026, Barcelona, Spain. ACM, New York, NY, USA, 5 pages. https://doi.org/10.1145/3772363.3798963
To approach the study design, we phrased the following three research questions:
RQ1: What measures do smart home users take to ensure privacy in their smart home?
RQ2: What aspects of smart home technology make users uncomfortable due to privacy concerns?
The study combined two methods: a task that participants were asked to prepare on their own, followed by a semi-structured interview.
The participants were asked to prepare a list of best practices for a private smart home. To make this assignment more tangible, it was presented with the following scenario, envisioning that this list is given to a person starting with setting up a smart home:
A close friend/family member/colleague is moving into their new home and finally has the chance to set up the smart home they’ve always wanted. They trust your advice and specifically ask how to protect privacy in their new smart home. What would you recommend that person to do?
Please help us create a list of best practices for that person.
The purpose of this two-folded design is that the participants are sensitized to the topic discussed later in an interview and have reflected on their own attitudes and behavior beforehand. Most primary smart home users have a technical professional background, suggesting they are used to approaching a topic with a solution-oriented mind. Advising an imaginary individual probably corresponds to a communication they are used to from their work environment, their experience setting up their own smart home, and their interaction with the Home Assistant community.
Furthermore, advising a fictional smart home set up might help to bring up practices and topics contradicting their behavior regarding privacy, which is to be expected to exist, according to earlier research describing the privacy paradox, which argues that people often say privacy is important to them. Still, the value is not reflected in their actions. Asking directly about their behavior and how it contradicts their values and beliefs might trigger embarrassment or shame, making participants feel uncomfortable sharing with the interviewer.
Additionally, preparing for the interview beforehand and reflecting on their attitudes and behavior enables the participants to be more in control of the information they share about themselves, allowing them to withhold topics they don’t want to talk about consciously.
The best practice list with the scenario was also shared with the community in a thread of the Home Assistant Forum (here), and the link was shared with the participants, allowing them to discuss the topic beforehand with other users and see how others approach it. Additionally, the post in the online forum informs the community about the current user research focus, encouraging the participants to contribute their opinions and collect additional data concerning the practical implementation of a private smart home.
We invited the participants to a video call within Google Meet. Initially, we asked the individuals to provide informed consent based on the User Research Agreement we shared with them.
We followed an interview guide for the conversation, including a short introduction covering demographic data about themselves, their living space, and smart home setup, and cohabitants if they have any. Then, we asked each person how they would rate the privacy of their smart home setup on a scale from zero to ten, with ten being the highest, and how they explained their rating.
Following that subjective rating, we explored how they feel about their smart home setup being in that state, why privacy is important to them, and who they are protecting their privacy from.
The topic “how to build a private smart home” referred to the best practice list they prepared in advance, how they assess privacy, and when they consider it in maintaining and expanding the smart home, if there are devices or technologies they are hesitating about, and how they know so much about privacy.
Towards the end, or if brought up in between, we discussed how cohabitants feel about privacy in the smart home and whether they have previously addressed the topic with visitors or guests.
The interviews were auto-transcribed by Dovetail and analyzed using Thematic Analysis. This method includes a systematic coding of all the transcripts with different labels (which can be seen in the videos of the participants). I focused on the analysis of the interview transcripts - the concrete best practices users suggested in the online forums are not considered in the analysis, as they did not include a lot of information about the users, but technical details about online privacy and security. For the creation of solutions, they can be considered in more detail if necessary.
The coding process helps the researcher interpret the participants' statements and understand the bigger context, ultimately leading to the creation of themes, describing the main insights of the interviews.
We created the following themes based one the interview data, which are described in detail in the following:
Theme 1: Privacy is a driver for outgrowing
Theme 2: Making the smart home private is not an end in itself
Theme 3: There are clear strategies for setting up and maintaining a private smart home
Taking care of the privacy of the smart home seems to be a driver for outgrowing. In other words, the decisions people made earlier regarding privacy, either consciously or unconsciously, don't feel good anymore for the people. The participants shared that they became aware of a vulnerability in their setup, and their motivation grew strong enough to abandon a device, service, or ecosystem to remove or replace it.
Some participants shared that they will lose trust in more providers because of the occurring privacy breaches, which might make them abandon their devices, ecosystems, or services in the future. They understand that trusting a service for the moment is not a final decision, but they expect and fear disappointments and the need to reevaluate their decision. They have the same feelings towards Home Assistant or other Open Home projects, overseeing them and being open to abandoning the technology if it doesn’t align with their privacy values anymore.
The reasons to invest efforts into safeguarding privacy in the smart home vary but are mostly related to socio-technical developments happening over the last years/decades. For example, the participants have experienced and observed multiple privacy breaches.
Furthermore, many people understood how companies are capitalizing on their data.
Understanding how user tracking is being used for political purposes, such as to influence elections, and current changes in politics emphasized the relevance of protecting their own and the other household members’ privacy for the participants.
Additionally, current practices such as Artificial Intelligence (AI) being trained with human data triggered the participants’ wish to protect themselves better as a precaution against yet unknown future developments in data handling and technological advances.
While some privacy concerns might not be pressing enough for users to take action to make the smart home more private, some external influences increase the priority to protect their privacy. These external influences seem to be associated with events in the user’s personal life.
One event is, for example, when a new person moves in with the primary user in a household or, for other reasons, expresses the wish to the primary user to reconsider or remove some devices or services in the shared household.
In this process, it also becomes clear that some areas in the home and some data, typically the main areas of the house, the bedroom, the bathroom, and particularly children's data, are assessed as more protection-worthy, triggering a reevaluation of the installed smart home technology.
Especially more personal information concerning the person’s body, such as pictures of nudity or their skin color, characteristics that a person cannot change about themselves, is considered more private by some participants.
Nabu Casa's customer support also received several requests for more detailed information and clarification, as well as changes in data handling when the US government changed. These customer reactions emphasize that events in individuals' personal lives, including political changes, trigger users to take action to protect their privacy better.
However, people can only react to such events if they learn about and understand how they might affect their privacy. The participants in this user study are very privacy-aware and inform themselves continuously.
People less engaged with privacy topics in smart homes might not be aware of situations and scenarios in which they should pay more attention to protecting their privacy.
Our communication around Home Assistant from Nabu Casa and the Open Home Foundation, as well as discussions within the community, can serve educational purposes and influence people to take action for a more private smart home, convincing them it is "the right thing to do".
Another theme that prevailed in the interviews was that people do not start a smart home with privacy in mind - the aim is to achieve other purposes, such as automating daily tasks or enabling energy monitoring, ideally in a private way.
Several participants believe that a fully private, or local, smart home is not realistically achievable.
The participants shared that they feel like there are always compromises to make concerning privacy within a smart home, either accepting vulnerabilities, holding back in installing desired functionalities and devices, or taking on additional effort.
Social influence and interactions with other people can impact a person's decision about the privacy of the smart homes, influencing them to take more or fewer actions. For example, family and friends being judgmental about a person caring too much can impact the individual to talk about the topic, or also reduce their effort. In a social circle where everyone cares a lot about privacy in IoT systems, an individual might be influenced to act more.
Some participants, for example, clarified in the interview that they are “not tinfoil hats,” indicating that they wanted to explain their intentions and were concerned about being misunderstood and judged for their beliefs.
For most people, protecting privacy is not fun per se. It does not seem to fall into the category of enjoyable tinkering, potentially because the outcome is not very visible and might limit the functionality. Accordingly, protecting privacy is not a priority, considering the effort the individual believes it would require to set up the smart home privately.
This is especially true if it concerns correcting earlier decisions, abandoning a device or service, or changing configurations. Several participants referred to the effort of making the smart home more private, resembling decluttering and sorting the mess built up over time, and how they feel frustrated to perform those tasks.
Since privacy-related tasks seem to be often procrastinated on and accumulate over time, starting from scratch seems to be a viable solution for users. Several participants shared that they have started from scratch or at least considered doing so, among other reasons, to set up their smart home more privately. Such a significant undertaking requires time and energy and will disrupt the smart home's daily usage and the habits and routines of other cohabitants.
Decisions within the smart home affecting privacy are not always made consciously, assessing all the consequences and deciding for a less private option. They are made ad hoc due to convenience.
People are aware of their paradoxical behavior, which became visible, for example, in how they rated the privacy of the smart home.
Earlier research on online behavior has revealed discrepancies between users’ attitudes and actual behavior and defined it as the ‘privacy paradox’, describing that users “claim to be very concerned about their privacy [but| undertake very little to protect their personal data” (Barth & de Jong, 2017).
The authors who defined this user behavior like this systematically analyzed 32 peer-reviewed articles, which accounted for 35 theories on the privacy paradox. As a result, they describe that the paradoxical behavior results in some cases from a risk-benefit calculation guided by rationality, more often by a biased risk assessment, for example, following misleading heuristics or prioritizing immediate gratifications, and often, only little or no risk assessment takes place.
For more details on the privacy paradox, see here:
Barth, S., & de Jong, M. D. T. (2017). The privacy paradox – Investigating discrepancies between expressed privacy concerns and actual online behavior – A systematic literature review. Telematics and Informatics, 34(7), 1038–1058. https://doi.org/10.1016/j.tele.2017.04.013
The Home Assistant’s Community Survey 2024 also illustrated the privacy paradox (see figure), where people rated the importance of privacy (M=4.61) and its practical implementation (3.71) in their smart home on a scale from 1-5.
This user study emphasized that there are strategies for setting up and maintaining a private smart home. The people I talked to in the interviews were all people with high technological affinity and skills; however, they don’t seem to follow their own advice. Considering the backgrounds of the participants here, we can assume that the level of user-provider privacy being implemented in smart homes that are set up and maintained by less knowledgeable primary users is lower.
Individuals' struggles to set up and maintain the privacy of their smart homes seem to stem from the complexity of the topic and the impression that the measures cannot achieve the desired result. Open and local protocols appear to users to be more private, but some doubts remain. To be confident about their decision, individuals would like to understand the working principles of the protocols in the specific devices in detail.
Additionally, several participants reported setting up multiple networks of one protocol, for example, for wifi, often for testing purposes before they expose it to the general network.
Security and privacy are often interwoven. Security often comes up in discussions about privacy and introduces additional complexity, blowing up the topic and potentially overwhelming users with technical terms, definitions, and details.
Furthermore, the participants shared that they often understand clouds to be untrustworthy as they maintain a connection to an external server that is not justified for the functionality a device provides. Some participants use external tools in addition to Home Assistant to investigate the behavior of their connected devices, mostly to detect “whether they call home” and to stop them from doing so.
Consequently, the provider often introduces restrictions to the device’s functionality, such as disabling the configuration of colored light when the cloud connection is cut off. This consequence leaves users frustrated about their purchases, feeling that the result did not justify their effort to set up a device privately. Experiencing such disappointments more often or hearing about them leads to users becoming dull with providers invading their privacy.
Some participants suggested concrete solutions that Home Assistant could implement as a feature to support users in setting up and maintaining a more private smart home.
This includes, for example
-
initiating a regular review of the data being tracked by Home Assistant
-
Telling users how private their smart home is and suggesting actions to improve,
-
Visualizing how much the server is talking to the outside
-
watching all smart home networks - or overall networks within the home - and pointing out privacy and security vulnerabilities
-
Motivating people to take action to improve their privacy
-
-
suggesting privacy-respecting devices with an “approved by” label
This user study involving participants with high privacy sensitivity and knowledge revealed that the reasons they think about privacy in their smart homes refer to their general knowledge about online privacy. The concerns do not refer to smart home technology specifically but to the experiences they made in larger contexts - understanding patterns such as user data being capitalized and political elections being influenced makes the users act cautiously in other areas of their digital life as well.
Most participants don’t have the impression that their data is very personal and protection-worthy—except in some areas such as bathrooms or data about children—but they rather want to protect their privacy out of principle.
The primary users seem to think about provider-user privacy first and foremost. On this topic, they are actively educating themselves, staying up to date about current political effects, and at least in their minds, seem to review their level of privacy regularly, being aware of potential vulnerabilities. Closing gaps and replacing a device or setting it up in a more private manner often has low priority and does not count as one of the favorite tasks of the users - they would appreciate help with the facilitation within Home Assistant.
Reach: Privacy is a cornerstone of Home Assistant. Therefore, it reaches all of our users and also potential users in our marketing materials.
As one of the core values of the Open Home Foundation, privacy impacts all users and all Open Home projects. A big share of the Home Assistant community is probably highly aligned with this value already, while for others there is a big potential of the Open Home Foundation and Home Assistant community acting as educator on online privacy.
Impact: "Privacy is when the product does not do something." If we execute this opportunity correctly, the effect would be subtle and users should not be aware of it.
This might be true for users who are not interested in privacy. For users who are currently investing a lot of effort in researching, setting up, and maintaining their smart home to be private, an opportunity is to help them regain a lot of their time for other purposes. Additionally, executing this opportunity visibly would reassure people about the privacy status of their smart home, taking off the mental load to remember and worry about the vulnerabilities in their smart home system. For people who are not that privacy-conscious there is an opportunity to illustrate the choices and compromises they accept about data handling and make them reflect, leading potentially to people making more conscious choices about privacy in the future.
To discuss the results and ideate on potential solutions, please engage on the ux-design repo of the Open Home Foundation!