What multifamily residents think about WI-FI

Researched and compiled by: Dom Beveridge & Arjun Kannan

Introduction

In the fall of 2024, 20for20 published a white paper in collaboration with RETTC entitled “The State of Multifamily Connectivity." The intent of that paper was to create a business-driven framework for making Wi-Fi implementation decisions in multifamily buildings.

At the time, most of the available content on the subject was either highly technical or overtly sales-oriented. What was missing was a straightforward guide that pulled together enough information for a decision-maker, whether on the operations or ownership side, to understand the parameters of a Wi-Fi decision. The paper also summarized key business drivers and legal considerations of which leaders needed to be aware.

In its conclusion, the paper noted a critical information gap in the industry’s data about connectivity services. There is little publicly available data that delivers insight into resident sentiment toward managed Wi-Fi.

That gap is more significant than it may appear. Improving Wi-Fi quality in a multifamily community involves substantial effort and financial investment, yet underwriting usually treats all types of Wi-Fi as if they were equivalent. Bulk-to-unit Wi-Fi, for example, is not the same service as bulk managed Wi-Fi, which includes amenities like ubiquitous individual networks, enhanced cell coverage, and dedicated support.

Despite these clear differences, owners and operators tend to underwrite both services as if they hold the same value. Investors tend to view the price of managed Wi-Fi as a function of the local market rather than of the technology itself.

The real question worth investigating, then, is this: What matters to residents about Wi-Fi service, and does this vary by service type? Understanding such distinctions should enable better, more data-informed investment decisions about connectivity.

Research Thesis

There are good reasons why this data does not yet exist. It is extremely difficult to discern how residents feel about specific aspects of their living experience. Sentiment analysis—such as it is in our industry—has traditionally depended on surveys, which are a blunt instrument for uncovering what residents truly think or feel.

A community can ask a resident to score their satisfaction, typically with a specific event (move-in, work order completion, etc.) on a numerical scale. But even if the question is well constructed (which is a big “if”), the best we can hope for is a trailing indicator of satisfaction with a given aspect of service.

A better approach is to ask open-ended questions that allow residents to describe what matters to them in their own words. The challenge, of course, is that it is infeasible to collect and analyze that kind of unstructured data at scale. At least it was, until now.

This study introduces a new survey methodology, designed to address exactly this problem. The study engaged more than 30,000 apartment residents, using a text-based AI chat capability to “interview” them about their Wi-Fi experience.

Experience with sentiment analysis in this industry predicts that this research methodology will outperform conventional data-gathering and analysis approaches, which are summarized in Figure 1.

Where surveys rely on pre-defined question-and-answer formats, the research described in this paper followed an AI-guided conversational structure. That changes a couple of things about the research. First, it means that questions are open-ended. Respondents can reply in any way they choose, rather than being limited to choosing answers selected by the questionnaire designer.

The AI dialogue structure allows the conversation to pursue additional insight that is unavailable in a static multiple-choice form. Previous research efforts indicated that the more engaging modality of AI chat leads consistently to response rates that are double those of conventional surveys.

Chat fixes some pervasive disconnects between the insight the researcher is trying to uncover and the experience of the respondent. First, surveys are presented as a list of questions, or a number of forms that a resident must scroll through to complete the survey. Any reader who has completed a satisfaction survey is familiar with the feeling of survey fatigue that each survey request inspires.

Chat asks one question at a time, and is intelligent enough to know whether or not to keep asking questions. That lowers the friction in the collection process. It also removes an unhelpful bias in traditional surveys: the implicit assumption that every fixed question matters equally. A 10-question survey assumes that the respondent will answer all 10 questions. But, in reality, each respondent is only likely to care about a small subset of the questions on the list. Responses, then, are not of equal value, but disparities in value are not visible in the data.

This research inverts the model from designer-led to respondent-led, marking a methodological shift in how the industry can measure sentiment. When AI can dynamically follow lines of questioning to explore topics residents actually care about, it reveals a broader range of interests and sentiments, both positive and negative.

The following sections explain the methodology behind the test, share the results, and draw conclusions about what they mean for both managed Wi-Fi providers and property management companies.

Research Methodology

The objective of this research is to interview residents at scale to uncover their sentiments about Wi-Fi. The design of the research described in this document, therefore, follows the principles of a good interview. The AI is not simply asking questions, but gauging each participant’s engagement and interpreting their responses to each question before answering the next.

Another important consideration in this research is that, unlike a conventional survey, the questions are not all asked at the same time. Multiple threads of conversation were deployed over time, allowing data capture through multiple, natural interactions and later analyzed together.

Because the output of these conversations is composed entirely of words, rather than numerical scores, the analysis differs from that of a traditional survey. The analysis detects patterns of sentiment in responses, like whether a comment conveys positive, negative, or neutral feelings. For example, “I’ve never noticed it” and “It’s fine as long as it works” differ structurally but express the same mild satisfaction—or even indifference—toward the service.

The following subsections describe how data was gathered and analyzed in this study.

2.1 Engagement Framework

Figure 2 presents a conceptual framework showing how the engagement process works. The AI pursues answers through a structure of anchor questions and follow-up questions.

Anchor questions are typically short and designed to gauge the respondent’s interest in a topic. When a response suggests strong engagement, the AI follows up with exploratory questions to maximize insight. When the response seems disengaged, the AI moves on to a different anchor question until it detects engagement.

Anchor questions exist to identify which connectivity topics matter to this specific resident. Examples of anchor questions include:

• “How important is reliable internet to your daily life?”
• “Do you notice any differences in connectivity between your apartment and common areas?”
• “How has your internet service been working for you? Any issues or pleasant surprises?” 

Follow-up questions are only deployed when a resident shows engagement with an anchor question, and are designed to go deeper on selected topics. Examples include:

• “Can you tell me more about that experience?
• “How did that affect your work?”
• “How does this compare to Wi-Fi you’ve had before?”

Figure 3 presents an example of a conversation where an anchor question is followed by a series of follow-ups.

The example in Figure 3 demonstrates some key principles of this research methodology:

• The resident steers the conversation toward the topics that matter to them, revealing their priorities rather than choosing from pre-selected answers.
• Multi-turn exchanges can expose contradictions, e.g., if a resident were to comment that service is unreliable, then subsequently share that they have not experienced any outages, the conflicting data points can be resolved in real time.
• Signals such as awareness of alternative providers indicate risk of churn.
• Respondents can introduce outcome-focused language, e.g., “reliable for video calls,” rather than product-centric language, like “300 Mbps,” further indicating what they actually value.
• Answers distinguish between hygiene factors, like basic coverage, and differentiators such as reliability, which ultimately drive decisions.

2.2 The Power of Correlation

The point of sentiment analysis is to understand the kinds of decisions businesses could take differently to drive better commercial outcomes. In the case of multifamily housing, commercial outcomes are mostly to do with decisions to sign a new lease or a renewal. It is natural, therefore, to try to frame sentiment in terms of its impact on those decisions.

At the same time, surveys must balance the conflicting objectives of brevity and completeness. The need to maximize response rates creates a natural pressure to shorten surveys, often resulting in questions that try to find out too many things at once. Researchers exploring relationships between sentiment and renewals are particularly susceptible to this problem.

As Figure 4 demonstrates, a single question about an amenity can be interpreted and, hence, answered in many different ways by residents.

The broad range of interpretations demonstrated in Figure 4 should render the scores that the question would attract from residents as unreliable. They may, at best, provide a general indicator of sentiment toward the amenity. But they would not constitute a reliable insight into the relationship between the amenity and renewal decision.

One of the biggest challenges is that feelings toward an amenity and feelings about renewal may or may not have anything to do with one another. A much better way of understanding this dynamic is to research the two things separately, then look for patterns of correlation. Over time, the AI can pursue two independent lines of questioning, using the anchor and follow-up question structure described above.

Renewal Intent and Outcomes

To characterize renewal intent, it is helpful to understand two things: each resident’s sentiment toward a forthcoming renewal opportunity, and the outcome of renewal decisions. The combination of what residents plan to do and what they actually do enables us to interpret renewal sentiment reliably. Two separate data elements were gathered for this research:

1. Explicit Renewal Intent: Residents were asked directly about their likelihood to renew using a 5-point scale that captures switching intent. These questions were asked separately—both in time and context—from Wi-Fi satisfaction, reducing survey bias.

2. Longitudinal Renewal Data: Anonymized Property Management System data was used to track renewal outcomes 30, 60, and 90 days after the survey. This analysis enables us to estimate the gap between stated intent and actual behavior.

Where the question in Figure 4 creates a highly unreliable data set, the separate research of renewal intent and Wi-Fi sentiment described in this section yields two reliable datasets. It identifies cohorts of residents based on their sentiment toward the amenity, and a separate set of cohorts based on relative likelihood to renew. The correlation between the cohorts can be instructive. 

For example, if the residents who express dissatisfaction with the amenity correlate highly with the cohort of low-probability renewals, it should be a cause for concern. While that finding alone may not necessarily represent a cause-and-effect relationship, operators should generally invest in amenities that increase satisfaction and intent to renew. For that reason, the correlation of various findings about Wi-Fi service and renewal likelihood is a focus of the results summary in Section 4.

Contrast with Other Satisfaction Reports

Finally, it is important to consider what other forms of sentiment analysis are available to operators seeking to understand resident perceptions of Wi-Fi service. Beyond the event-based surveys already discussed, another potential source of insight comes from the satisfaction surveys that connectivity providers send to residents.

These survey types share many of the same shortcomings as the others mentioned earlier. Closed questions and multiple-choice formats limit what respondents can express compared to open-ended questions that allow them to describe how they actually feel.

Crucially, when responses can be organized by the types of sentiments they express, it becomes possible to identify the kinds of correlations discussed above. The combination of rich sentiment data, its connection to commercial outcomes, and results spanning multiple communities and providers offers far deeper insight than any single provider’s research could deliver.

Service types are becoming increasingly differentiated, and delineations between categories like “Retail,” “bulk to unit” and “bulk-managed” services more blurred. In this environment, it will become increasingly important to understand how residents feel about different service types. 

These differences, and the profile of respondents in this study, are the subject of the next section.

Research sample and high-level observations

Conducting the study described so far in this paper requires a large sample set and a high enough response rate to be able to draw meaningful conclusions from the results.

To ensure that this exploratory study met both standards, a sample of 30,000 units was drawn from ResiDesk’s customer base. The residents of these units already participate in text-based conversations with the same platform that conducted the study and were hence likelier to engage with the questions. The breakdown of the 30,000 units is shared below.

The service type breakdown was 18% Managed Wi-Fi and 82% Retail ISP. Nobody knows what proportion of US multifamily properties is currently using managed Wi-Fi services, but the consensus of reports places the number below 20%. On that basis, the split in responses is roughly consistent with the industry average.

The geographical spread of submarkets is divided mostly between Urban (12K) and Suburban (17K) markets, with 1K in ex-urban locations. This distribution provides a potentially important nuance: competition is usually the most intense in cities where residents have the most options. Suburban properties have limited alternatives, so the variety can disclose where service type matters most.

Property Class represented a distribution where just below half fell into Class B, with half that number falling into Class A and the balance in Class C. That is a useful breakdown in an industry where the conversation about connectivity tends to gravitate toward Class A. The breakdown of building types was unsurprisingly more than half garden style, with 15% high-rise and 30% mid-rise.

For a sample of 30k units, this represents a cross-section of the mainstream of US multifamily real estate. The response was substantial: of the 30k units contacted, the AI engaged with 11,614 total units in 11 states over a 90-day period from August through November 2025.

These residents’ pre-existing engagement with ResiDesk meant they were more likely to participate when asked—yielding an engagement rate of almost 40%. That is a multiple of typical survey response rates and almost double the expectation for an AI-driven survey. But what is more important than the response rate is the far higher quality of responses.

Why Engagement Rate Matters More Than Response Rate

Traditional survey research measures success by completion rates (benchmark: <9% poor, >15% excellent, >20% exceptional). Those metrics reflect completion quality but miss something much more important: engagement quality.

Section 1 described how this AI research technique overcomes a flaw in conventional surveys: that lists of questions assume all questions are of equal importance. Instead of asking all residents the same set of questions, most of which they don’t care about, the interview path adapts as residents self-select into the topics that matter to them.

That requires us to think differently about response rates. A 38% response rate does not mean that 38% of respondents provided responses on every line of questioning. It means they engaged with the subset of the interview that most reflected their opinions.

A few statistics about the 11,614 responses demonstrate their quality and depth. The conversations achieved 3.99 average exchanges per resident, with 62% reaching more than three exchanges and 38% reaching 5+ exchanges. Some conversations even reached 30+ exchanges when residents had urgent issues to explain.

The anchor-and-follow-up structure described in Section 2.1 creates genuine and detailed responses to the most salient questions. The self-selection of topics means that the research should have a dramatically higher signal-to-noise ratio.

3.1 Response Segmentation

The conversational structure of this research yields a data set that supports an unusually broad range of analysis and interpretation. The major insights from the research are shared in detail in the next section. But first, it is helpful to understand some broad characteristics of the responses in the data set.

Respondent Attitudes to Wi-Fi

The more actively a respondent engages with a line of questioning, the stronger we can assume their opinion to be. Yet, in the case of Wi-Fi, relatively low engagement is generally a good thing. Across the whole population of respondents, just over a third fell into this neutral sentiment category. The following statements typify each of the three attitude segments:

• 35% Neutral/Satisfied: “Wi-Fi works fine, never thought about it”
• 48% Engaged Shoppers: “Wi-Fi matters, I’m looking to get the best service I can”
• 17% Very Negative: “Ready to switch/leave.” 

As we progress through the findings of this study, it is important to understand how we separate signal from noise in the data. Section 4 will frame specific findings from the analysis and their potential implications. But first, it is important to understand some of the broad characteristics of the comments that respondents shared.

Sentiment Distribution

For example, a negative comment in a conversation does not mean that the respondent’s overall sentiment is negative. Negative and positive comments coexist in the same conversation. Figure 6 demonstrates a counterintuitive feature of these responses. It segments respondents based on whether or not they commented that their service had been unreliable.

What is surprising is that the positive–neutral–negative sentiment distribution of comments is almost identical for residents who characterized their service as “unreliable” and those who did not. That said, all “complaints” are not created equal. While the incidence of complaints reported in Figure 6 is similar both when a respondent has reported unreliable service and when they haven’t, the sentiment those complaints reflect is quite different.

Section 4 explores the sentiment distribution more deeply, including its correlation with churn rates. An important piece of context is that residents with different types of Wi-Fi service tend to complain about different things.

To tease out some of the differences in sentiment, we can segment comments by topic. Figure 6 splits the responses by service type, with separate totals for respondents accessing Wi-Fi through either “Retail” or “Managed” services. While Figure 6 shows how the incidence of complaints is similar across all respondents, Figure 7 exposes significant differences in what those complaints are about.

Speed performance and outages featured much more frequently in conversations with retail users compared to the managed Wi-Fi cohort. That may explain why a relatively high proportion of respondents mentioned alternative providers during their interviews. “Alternatives” in Figure 7 covers a range of comments from, e.g., “I wish I could have Verizon like my old place” to “I can’t look at other options for internet” to “Xfinity seems cool.”

The frequent presence of alternative providers and their ubiquitous marketing likely affect the perception of residents who have selected and entered a contract with a retail provider. Complaints from residents whose access to Wi-Fi is through a single provider may be less inclined to think about their Wi-Fi through the lens of how it compares with other service types.

Complaint levels are similar between the service type cohorts, but they seem likelier to trigger switching behavior for retail customers than for managed Wi-Fi users. A bigger question for multifamily operators is how these disparities affect resident experience more broadly, which is the subject of the next section.

Major findings

With more than eleven thousand responses—some spanning thirty separate interactions—this study yields an unusually large dataset. It creates the luxury problem of having many possible lines of inquiry to pursue. In reviewing the most interesting correlation patterns, we chose to frame the findings around three first-principles questions.

The first is simple: does Wi-Fi quality matter? In exploring that question, we evaluated popular preconceptions about whether residents see connectivity as a basic utility or as a service whose value they actively recognize.

The next question is how Wi-Fi matters, examining the commercial impact of the sentiments revealed in the data. Finally, we draw conclusions about what residents mean when they talk about their experience with Wi-Fi.

4.1 Does Wi-Fi Quality Matter?

The premise for this research is that little reliable information exists about what residents think about Wi-Fi. In the absence of solid data, a wide range of assumptions tends to fill the gap. Before tackling questions about how impactful Wi-Fi is and what influence it may have on the multifamily experience, we must start with the simplest one: Does the quality of Wi-Fi even matter?

Industry views about Wi-Fi range from “just a utility” to a differentiated product that delivers a markedly better service. Because there is such a wide variety of Wi-Fi offerings, there is also a broad range of resident sentiment. Figure 8 summarizes those sentiments when viewed through some of the industry’s most common preconceptions.

Connection speed is a good example. “Gig speeds” are prominent in marketing communications, but that does not reflect how much speed most people need for daily use. Higher-speed capacity naturally future-proofs communities. The question is whether residents value it. The data suggests they do—up to a point. Baseline speeds for video calls and gaming appear sufficient for most residents, and overperformance beyond this point goes largely unnoticed.

A similar pattern appears in customer service. A core value proposition of managed Wi-Fi is the availability of dedicated support for a building-wide network—far better than treating each unit as a separate customer in a retail model. Yet the absence of problems matters more than access to high-quality support that can fix them. Residents do not want to have to talk to anyone to resolve issues. The factors that prevent breakdowns have a huge impact on avoiding negative sentiment, but the upside of excellent service teams is often unseen.

Many of the most advanced Wi-Fi platforms offer premium features designed to extend uninterrupted connectivity. These include ubiquitous property-wide network access rather than unit-bound Wi-Fi, and cellular assist to prevent dropped calls in weak-signal areas. It is obvious how these features improve resident experience.

However, the data suggest that they matter only to the extent that they prevent problems. Residents expect strong signal everywhere and treat it as a hygiene factor. Failures create negative sentiment; success, despite its technical complexity, does not materially elevate perceived value.

Reliability follows a similar pattern. Extremely high uptime is a natural differentiator of high-quality building networks, but residents do not think about uptime in percentage terms. Responses reference outage frequencies: once per month, once per week, and so on. The grace period for outages appears to be limited to one per month.

Multiple outages undermine confidence in the network and, in the most unsatisfactory cases, prompt people to consider moving out of the community.

A frequent framing of the question of resident perception is to ask if residents see Wi-Fi as a value-adding amenity or as something more like a utility. The short answer, based on this analysis, is that Wi-Fi is like a utility when it works well, and more like an amenity when it doesn’t.

4.2 How Does Wi-Fi Quality Matter?

The findings so far tell a clear story: success in avoiding downside service issues does more to shift resident sentiment than the upside of premium services. The logical next question is how that negative sentiment affects commercial outcomes for multifamily operators.

In particular, we want to understand how sentiment toward Wi-Fi affects economic behavior. As Section 2 explained, the main resident decision Wi-Fi sentiment can influence is whether a resident renews their lease.

Figure 9 shows the results of the most interesting analysis relating to outages. Across the population surveyed, the average churn rate—the share of residents who chose not to renew—was 44%. When we break down the types of complaints described in the interviews, we see clear signs that residents with different experiences fall into distinct renewal cohorts.

When residents express positive or neutral sentiment about Wi-Fi, their churn rate is effectively unchanged, remaining within a percentage point of the population average.

The relationship between Wi-Fi sentiment and renewal is asymmetric. Positive feelings about Wi-Fi do not appear to increase renewal likelihood: few residents renew leases because they are pleasantly surprised by how good the Wi-Fi is. However, the dynamic flips when outages become a problem.

Residents who complain about outages—typically anything from one per month upward—belong to a cohort whose average churn rises from 44% to 58%. The gap widens further when reliability issues compound, with churn reaching 83% when residents experience weekly outages.

The results of this analysis do not by themselves constitute a measured cause-and-effect relationship; the research does not isolate outages as the reason these residents move out. But the fact that renewal likelihood is so much lower for the same people reporting internet problems suggests that operators should consider inadequate control of outages as a risk to their business.

The “Complaint Profile” Thesis

Section 3 showed that managed and retail residents display nearly identical sentiment distributions—47.5 percent positive overall—yet their churn rates differ markedly. The difference lies not in how often residents complain, but in what their complaints signal.

Residents using retail Wi-Fi direct their negative sentiment toward reliability, speed, and pricing. Speed and pricing concerns may push them to shop for different vendors, but outages create a stronger response: when service fails during work or essential tasks, residents start to consider moving altogether. Complaints about retail Wi-Fi, in other words, are likelier to activate churn.

Managed Wi-Fi users still complain, but about lower-stakes issues such as device setup, coverage gaps, or modest performance limitations. Reliable service removes the outage risk that drives switching and predictable pricing eliminates surprise fee triggers. They express a similar volume of negative sentiment, yet far fewer of those complaints point to conditions that lead to churn.

The broader lesson is simple: residents want the internet to feel reliable. When sentiment is linked to actual renewal behavior, reliability emerges as the decisive factor. Managed and retail residents complain at similar rates, but the nature of those complaints differs—and only one set of issues drives people to leave.

4.3 What “Good WiFi” Means to Residents

While outages dominate negative sentiment about internet service, not all commentary focused on service shortfalls. Many residents described what “good” Wi-Fi looks like, and some clear trends emerged.

First, Residents do not talk about Wi-Fi the way suppliers do: they rarely mention cabling, specifications, speeds, or devices. Their narrative is about outcomes—speed as it relates to work productivity, predictable pricing, and peace of mind. Price predictability, in particular, appears highly salient. This is almost entirely a feature of retail rather than managed Wi-Fi.

What we did not see was much discussion of value for money. Wi-Fi pricing is competitive, and it does not appear expensive enough for residents to feel overcharged. What they dislike intensely is when prices change suddenly and without warning, typically at the end of promotional periods.

The discussion of speed and utility revealed another pattern. Positive descriptions of Wi-Fi often cited specific use cases and benefits, and the language closely matched that used in fiber advertising in their neighborhoods. Sample quotes included:

• “My new internet is 10-100× faster. That makes a difference when you make a living from home with video calls all day”
• “I can work while my kids are watching CocoMelon”
• “Work from home is only possible when I always have internet”
• “We’d love if management worked with a fiber provider.” 

After years of major telecom companies spending heavily to promote fiber networks, it should not be surprising when residents absorb those messages. It appears to affect how residents think about the service, referencing the terms the advertising provides. For service providers and operators attempting to promote their fiber-backed services, this suggests an opportunity to align language to the attributes of a service offer that is increasingly well-understood by multifamily residents.

Conclusion

Strong value propositions appeal to two human motivators: avoiding pain and seeking pleasure. Across this study, one theme dominates: Wi-Fi matters most when it fails. When it works, it behaves like a utility: essential, invisible, and taken for granted. Success produces silence. When it breaks, Wi-Fi becomes an amenity: visible, emotionally resonant and inferior to alternatives.

This dynamic explains why the identical sentiment distributions in Section 3.3 mask very different outcomes. Positive sentiment does little to increase renewal intent, while negative sentiment sharply depresses it. Providers often talk about speed and coverage, but residents reward dependability: whether the service supports work, feels predictable, and can be relied on without thought.

The same pattern underpins the churn paradox. Managed and retail residents report almost identical sentiment, yet their behavior diverges. For residents using retail Wi-Fi, negative sentiment has a direct path to action. Awareness of fiber options and frustration with outages or surprise fees lead to comparison-shopping, and churn accelerates as reliability failures accumulate.

This study has identified the issues that increase the risk of resident churn. They are the very risks that managed Wi-Fi architecture is designed to mitigate. Reliability engineering and dedicated support prevent outage cascades and ensure high enough speeds. Predictable monthly pricing removes fee shocks.

As Section 3.3 demonstrated, managed Wi-Fi users still complain, but their frustrations tend to involve lower-stakes issues. As a result, churn remains stable even though sentiment volume is similar. The sentiment is the same; the behavioral mechanics are not.

When residents describe what they want—”fast, reliable, affordable internet, etc.”—they are echoing a message shaped by the industry’s largest marketing budgets. Managed Wi-Fi delivers on these requirements, but it does not “own” the narrative framing that shapes residents’ expectations. The opportunity is communicative rather than technical: to translate the utility of managed service into the amenity language residents have been taught to recognize.

Above all, this study has highlighted the range of understanding that can be unlocked by the power of open-ended questions. This short, limited-scope study has clarified some intuitions and challenged others. We now know a little more about the impact of connectivity on multifamily communities, and we know how to extend and deepen similar levels of insight throughout the resident experience.