Abstract
Using a sample of nonprofit human service organizations in South Central New York, this article analyzes the nature and extent of their Internet presence by examining their adoption of various Internet-based tools as well as their activity and visibility on the Internet. Nearly all organizations had websites but few provided interactive features, beyond the opportunity to donate. Many organizations also used at least one form of social media, although level of adoption, activity, and visibility varied. High capacity organizations and those with members had higher levels of Internet presence while those reliant on program service fees and government funding had lower levels. In addition, our results suggest distinct aspects of Internet presence are related and organizations can increase their visibility on the Internet by being more active.
Introduction
Recent research has deepened our understanding of the ways in which nonprofits use Internet-based tools, especially social media (e.g., Guo & Saxton, 2017; Lovejoy & Saxton, 2012), but far less is known about a more elementary phenomenon: Internet presence, or the extent to which nonprofit organizations are available, or present, to the public online. Internet presence matters as a construct because it provides important information about both the degree to which nonprofit organizations are using the Internet to advance their missions and the degree to which they view online engagement as valuable. Also, measuring Internet presence enables us to examine factors affecting its different dimensions. This knowledge provides necessary context for the growing body of work on how nonprofit organizations utilize social media and other Internet-based technologies, particularly variations in use among different types of nonprofit organizations.
Internet presence has three elements: adoption, activity, and visibility. Adoption refers to an organization’s choice to use different Internet-based technologies, activity measures how much it uses the tools it has adopted, and visibility is stakeholders’ awareness of the organization’s online activities. These three dimensions taken together capture the extent a nonprofit organization is using the Internet and is in a position to leverage its power to engage stakeholders and advance its mission. As implied, this definition incorporates presence both through websites and social media, and acknowledges the multiple places where nonprofit organizations can be present online. The conceptualization of Internet presence as a distinct construct is new and brings together phenomena that other researchers have largely considered individually, focusing on either one dimension of presence or one Internet-based tool, such as website adoption (McMahon, Seaman, & Lemley, 2015; Powell, Horvath & Brandtner, 2016), Facebook adoption and use (Maxwell & Carboni, 2016; Saxton & Waters, 2014), or Twitter adoption and use (Saxton & Guo, 2014).
After discussing the concept of Internet presence in greater depth in the next section, we explore it in one nonprofit subfield: human services. We study Internet presence across multiple dimensions and forms of technology in this subfield and provide data at two points in time to examine how Internet presence has evolved. We also identify organizational characteristics influencing Internet presence, highlighting variations across different types of organizations in the nonprofit sector.
Internet Presence
As noted, we define Internet presence as the extent to which nonprofit organizations are available, or present, to the public online; it has three dimensions: adoption, activity, and visibility. Organizations can choose to adopt a variety of Internet-based technologies, such as by creating a website, a Facebook page, or an account on Twitter or other social media platform. When an organization chooses to adopt an Internet-based tool, it makes Internet-based engagement with the public possible. In this way, adoption is the starting point for an organization’s online presence. Activity refers to how much organizations use Internet-based tools once they adopt them and can be measured by things like the frequency of tweets or Facebook posts. An organization that posts on Facebook one time per month has made a very different choice about how much to be present on the Internet than one that tweets or posts daily. The third dimension of Internet presence is visibility. Organizations may choose to adopt Internet-based technology and engage in activities such as posting on Facebook or tweeting; however, that activity does not mean the organization is visible to the public online. Organizations are only visible to stakeholders when stakeholders choose to interact with them online by following them or becoming a fan (on Twitter or Facebook, respectively), or by visiting their website. As such, visibility includes such things as the number of Facebook fans or Twitter followers an organization has, or the number of website visits it receives. Visibility matters because it is only through visibility that organizations can use the Internet to engage with stakeholders. An organization with four Twitter followers has much less visibility and potential for engagement than one with 4000 followers. The three elements of Internet presence provide a foundation for understanding the subsequent choices nonprofit organizations make about how to use Internet-based technology.
Nonprofit Organizations Online
The Internet presence of nonprofit organizations has not been a primary focus of past research. In the Web 1.0 era, scholars emphasized issues such as the structure and presentation of information on websites and organizations’ technology planning activities (Hackler & Saxton, 2007; Kang & Norton, 2004; Waters, 2007). More recent Web 2.0 era research examines strategic uses of Internet technology. While this body of work has led to a deeper understanding of how nonprofit organizations use specific Internet technologies, researchers have paid less attention to organizations’ overall Internet presence, which provides the foundation for organizations’ strategic choices.
In recent years, researchers have devoted increasing attention to the way in which nonprofit organizations participate online. These studies have included considerable information about the three dimensions of Internet presence: adoption, activity, and visibility. None of these three elements, nor Internet presence as a construct, are the focus of these studies, but researchers have provided these data as supplementary material. Table 1 lists 15 research articles and what they tell us about Internet presence in nonprofit organizations. All include adoption information; some provide data about the level of activity and a few report visibility numbers. Most studies focus on just one form of Internet technology. Nonetheless, they provide a useful overview about the current state of knowledge. The table notes the adoption rate (the percentage of the sample using a particular technology) and both the frequency of activity and the level of visibility, if available. We have also included the date of publication as well as when the data were collected—to correct any misimpression the lag in publication may give about adoption of technologies at different points in time.
Internet Presence of Nonprofit Organizations in Recent Research.
Although it is difficult to draw definitive conclusions from these studies, we can make several observations. While adoption rates have increased over time, they vary across different categories of nonprofit organizations and by technology. For example, few independent foundations use Facebook, 29% in data collected in 2013 (Foundation Center, 2013) and 11% in 2014 (Maxwell & Carboni, 2016). By contrast, many community foundations have adopted Facebook, with 72% using Facebook in 2014 (Maxwell & Carboni, 2016). Twitter is particularly popular among advocacy organizations: both Guo and Saxton (2014) and Saxton, Niyirora, Guo, and Waters (2015) report adoption rates of 80% or greater for this subgroup. At the same time, two studies of human service organizations report very different levels of social media adoption, particularly for Twitter, 6% in one (Campbell, Lambright, & Wells, 2014), and 72% in the other (Young, 2017), but also for Facebook, 54% and 98%, respectively. Finally, social media use in the largest nonprofit organizations has grown over time, and these organizations have the highest adoption rates. Studies using the Nonprofit Times 100 (a list of the largest non-education nonprofit organizations in the United States) showed an increase in Facebook adoption from 65% in 2009 (Nah & Saxton, 2013) to nearly universal adoption (97%) in 2013 (Saxton & Waters, 2014). In studies that reported on multiple forms of social media, Facebook was consistently the most popular.
Research on Internet activity and visibility is limited to two subfields, advocacy organizations and foundations. Those data show advocacy organizations were more active on Twitter than foundations and generally had more Twitter followers. Taken together, these findings suggest variation in Internet presence, and use of these tools appears to differ among organizational subgroups, although more systematic study is needed.
Research about the predictors of Internet presence is even more limited. One notable exception is work by Nah and Saxton (2013) which examines the factors influencing the adoption of Facebook and Twitter and their frequency of use (equivalent to our category “activity”). While their results varied both by application type (Facebook or Twitter) and between adoption and frequency, several predictors were more consistent than others. Dependence on public support; program efficiency, the ratio of program expenses to total expenses; program service fees; and “website reach” defined in terms of the “degree of influence of the organization’s website” (p. 301) were positively associated with both adoption and frequency variables. Also, whether the organization had members was positively associated with Facebook adoption. Finally, status as a membership organization was negatively associated with a combined variable of Twitter and Facebook activity, and reliance on private fundraising dollars was negatively associated with frequency of Twitter use.
To understand Internet presence better as a foundational concept in nonprofits’ use of Internet-based technologies, we focused our research in one subfield, human service funders and providers, based in one region of the United States and investigated
the extent to which they had an Internet presence and offered the public opportunities to engage with them online,
how their adoption of Internet-based tools had changed over time, and
the factors affecting different aspects of Internet presence.
Research Methods
Funders and providers are the primary institutional actors in the nonprofit human services field which is why we collected data from both groups in a six-county area in South Central New York. This region includes small and medium-sized cities as well as suburban and rural communities. The county populations range from approximately 50,000 to 200,000. We gathered information on how these organizations utilized the Internet during two time periods: March-April 2012 and October 2014-November 2015 (n = 163 and n = 157, respectively). We also collected considerable background information on them to identify factors impacting various aspects of Internet presence.
The nonprofit funder portion of our sample includes all local United Way chapters and community foundations as well as the major local private foundations, all of which fund human service nonprofit organizations. To identify nonprofit providers, we asked each of the three types of nonprofit funders and county departments in the region involved in the delivery of human services for lists of the providers they funded. Our sample for both waves consisted of 24 funders (14 private foundations and 10 public charities). We had 139 nonprofit providers in our 2012 sample and six fewer in 2014-2015 because two mergers had taken place and four organizations had ceased to operate.
In both waves, we noted whether organizations had adopted a variety of Internet-based tools. In addition, we collected supplemental data on their use of these tools in our second wave. Table 2 summarizes the data collected by wave. Based on a review of organizational websites, we also noted whether organizations served vulnerable populations (i.e., any group for whom confidentiality is a primary concern such as victims of domestic violence). We analyzed this data using descriptive statistics and chi-square tests.
Internet Data Collected by Wave.
We obtained the 2013 and 2014 IRS 990 returns for 118 organizations in our sample as a source of additional data. Utilizing data from both years, we calculated the average percentage of total revenues organizations received from each of the following sources: (a) program service fees, (b) government grants, and (c) fundraising (the sum of revenue from fundraising events and “all other contributions”). In addition, we calculated each organization’s average total number of employees as a proxy for size (2013 and 2014 data) as well as age based on founding year (2014 data only) and recorded whether it was a membership organization (2014 data only). Several organizations were private foundations or completed EZ forms and thus did not report data for some variables. We left any missing values “blank” in our dataset. The 990 data were used along with dummy variables for status as a funder, affiliation with a parent organization, and focus on a vulnerable client population as independent variables in a series of regressions predicting various aspects of Internet presence.
We selected these independent variables for several reasons. Earlier research identified program service fees, membership status, and working with a vulnerable client population as factors influencing Internet and social media presence (Campbell et al., 2014; Nah & Saxton, 2013). In addition, because organizations can use the Internet to engage stakeholders, we selected variables consistent with that role and the resource dependence they reflect. For example, organizations reliant on fundraising are likely to view individual donors as key stakeholders and may use the Internet to connect with them. In addition, we hypothesize that some organizations have greater capacity to use Internet-based tools, such as organizations that are larger, older, or affiliated with a parent entity. Finally, we included a dummy variable to control for the functional differences between nonprofit funders and providers.
We conducted logistical regressions using adoption of the following social media as dependent variables: (a) Facebook, (b) Twitter, (c) YouTube, and (d) blogs. 1 We could not conduct a comparable regression with website adoption as a dependent variable because only three organizations without websites had complete data. In addition, we ran a logistic regression examining what predicts how many forms of social media an organization has by creating a “combined social media” dummy variable, clustering the standard errors by organization. Each nonprofit had four observations in the dataset for this analysis and received a “1” for the combined measure for each form of social media they had. For example, an organization with Facebook and Twitter only had two observations where the combined social media variable was recorded as “1” and two observations where it was recorded as a “0.”
We also used ordinary least squares (OLS) regressions to explore factors influencing (a) the mean number of monthly Facebook posts (a proxy for activity) and (b) the number of Facebook fans (a proxy for visibility). Our dependent variables for the activity and visibility models were the natural log of Facebook posts and fans, respectively, to correct for skewness. In addition to the standard independent variables, we included the mean number of posts as a predictor in the visibility model, hypothesizing that more activity would result in greater visibility. Drawing on Nah and Saxton’s (2013) finding that website reach (analogous to our concept of visibility) was positively related to social media activity, we were interested in exploring whether different aspects of Internet presence were related. There were too few organizations using Twitter with complete data to do similar analyses for activity and visibility on this platform. Given the small sample size for most of the regressions, we used a less conservative threshold and considered a p value lower than .1 to be significant.
Findings
Nearly all organizations used an Internet-based tool, though the popularity of different tools varied. By 2014-2015, 94% (n = 147) had a website. Just one of the organizations without a website had a Facebook page, indicating these organizations generally did not use Facebook as a website substitute. Of those with websites, most had sites providing one or more mechanisms the public could use to contact them: 99% (n = 145) included a general phone number; 82% (n = 120) posted a general email address; many fewer (39%, n = 58) posted email addresses for specific staff. Several also had features designed to facilitate stakeholder engagement, although the percentage using many of these features was minimal: 72% (n = 97, excluding private foundations 2 ) provided an online donation option, 19% (n = 28) let individuals subscribe to receive email updates, 1% (n = 2) let the public comment on news updates, 1% (n = 1) had an online discussion board, and 1% (n = 1) let website visitors take a survey. Table 3 summarizes key adoption information for the entire sample.
Comparison of Internet Presence in Waves 1 and 2.
Although not as common as having a website, nearly three-quarters used at least one form of social media by 2014-2015. But only 13% had adopted three and 2% utilized all four. There were differences by groups (mean for private foundations = .14, public charity funders = 2.3, providers = 1.29, p < .001 for the Kruskal–Wallis test). As shown in Table 3, Facebook was the most popular; 69% had an account (n = 108). This count does not include “About Pages” created by other Facebook users. A quarter (n = 39) utilized Twitter. A modest percentage had a YouTube channel (18%, n = 29). Just 13% of organizations (n = 20) used a blog. Of those with blogs, only half allowed public comment. Also, as illustrated by Table 3, we found small increases in the percentage using websites and each of the social media tools between waves. Based on chi-square tests, organizations were only significantly more likely to have adopted YouTube in Wave 2 compared with Wave 1 (p < .01). Five organizations with Facebook and one with Twitter in Wave 1 no longer had these accounts in Wave 2.
Because Facebook and Twitter were the most popular forms of social media, we analyzed organizations’ level of activity and visibility on these platforms and adoption by year. Among organizations using Facebook, the median number of fans was 315 although the number varied considerably, from three to 4,951. Organizations had a median of six posts per month, indicating that many were posting on a somewhat regular basis, but this category showed a high level of variation too, with a range of 0 to 48 posts per month. Compared with Facebook, fewer people interacted with organizations on Twitter: the median organization had 155 followers. However, organizations used Twitter more frequently than Facebook and had a median of 8.67 tweets per month (range = 0-350). Table 4 provides a count of organizations adopting Facebook and Twitter by year. As this table illustrates, most organizations using these tools had their current accounts by 2012 (74% of those with Facebook; 67% of those with Twitter). Ten organizations closed their Facebook accounts and opened up new ones between waves (i.e., the establishment date recorded in Wave 2 was after we collected data in Wave 1). In some cases, this change appears related to a merger or name change. In other cases, the impetus was unclear.
Count of Wave 2 Organizations Adopting Facebook and Twitter by Year.
Note. A Facebook adoption date is not listed for two organizations because they had “individual” profiles rather than organizational pages.
In addition, we assessed factors influencing various aspects of Internet presence. Tables 5 and 6 detail descriptive statistics for variables in our regression analyses and summarize our results, respectively. 3 Entities with parent organizations were more likely to have Facebook (p < .1) and Twitter (p < .05). In addition, older organizations were more likely to adopt Twitter (p < .05). By contrast, organizations less reliant on program service fees were more likely to use YouTube (p < .1) and blogs (p < .05). Generally consistent with the prior models, entities with parents, membership organizations, and those receiving fewer program service fees adopted more forms of social media (p < .05 for all). The factors influencing visibility were different than those affecting adoption. Organization with more employees (p < .1) and less revenue from government grants (p < .1) had more Facebook fans. Organizations posting more frequently were also more visible (p < .001). The OLS results for the visibility model should be interpreted cautiously: while the distribution of the logged dependent variable was much more symmetrical, there were still outliers. As shown in Table 6, the activity model explained very little variation in the dependent variable, and only reliance on program service fees was marginally significant.
Descriptive Statistics for Regression Variables.
Note. The dependent and mean post variables are based on 2014-2015 data, and the 990 data are from 2013-2014.
One organization reported negative numbers for other revenue and investment income.
Factors Influencing Social Media Adoption and Facebook Visibility.
Note. Dependent variables are based on 2014-2015 data. Cell entries are unstandardized parameter estimates. (Odds ratios and standard errors are in parentheses for the logistic and OLS regressions, respectively.)
p < .1. **p < .05. ***p < .01.
Discussion
This article develops the concept of Internet presence and analyzes its three dimensions by examining human service organizations’ use of multiple forms of Internet-based technology. The variations in Internet presence among our sample we found suggest human service organizations have differing perceptions about the extent of the Internet’s mission-enhancing benefits. Websites provide a base level of Internet presence, and nearly all organizations we studied had adopted them. But beyond websites, organizations made a wide range of different choices about how to be present online. The most obvious point is that social media adoption varies by type of organization, although we also found wide variation in activity and visibility on Facebook and Twitter within our sample. The difference in adoption rates between the two types of funders we studied stands out. Public charity funders on average had adopted significantly more forms of social media (2.30) compared with private foundations (.14). This result is consistent with the different levels of Facebook adoption Maxwell and Carboni (2016) found when comparing these two groups. Having an Internet presence is more important for public charity funders because they rely on the public to sustain them financially and as a result have to do more direct outreach (including outreach online) than private foundations. This interpretation also seems appropriate for the providers we studied, though the average number of social media tools they had adopted (1.29) was somewhat lower. The extremely low level of social media adoption by private foundations suggests they do not view Internet-based technology as essential in advancing their mission, in sharp contrast to the other types of human service organizations we studied.
While our descriptive statistics provide insight into the extent to which human service organizations have an Internet presence, our regression results contribute to our understanding of the organizational characteristics affecting this presence. Our results suggest that two organizational characteristics positively affect Internet presence: organizational capacity and having members. High-capacity organizations have higher levels of Internet presence, as indicated by the positive associations between (a) several dependent variables (Facebook, Twitter and overall social media adoption) and having a parent organization, (b) Twitter adoption and organizational age, and (c) the number of employees and the number of Facebook fans (our visibility measure). Furthermore, the positive association between membership and overall social media adoption, as Nah and Saxton (2013) also found, suggests that membership organizations see Internet-based technologies as important mission-advancing tools, perhaps because of the centrality of communication with members in these types of organizations.
Two other organizational characteristics were associated with lower levels of Internet presence: program service fees (with YouTube, blog, and overall social media adoption and with activity) and level of government funding (with visibility). One way to interpret these findings is to see them as indicative of the nature of the work these organizations do and how they interact with their beneficiaries. For example, organizations reliant on program service fees (such as self-pay, private insurance, Medicaid, and Medicare) serve many individual clients. These clients may lack choice when selecting a provider due to thin markets in the communities we studied, or they may be directed by referral sources to specific providers. In both situations, provider organizations may not need a high level of Internet presence to recruit clients. In short, human service organizations with high levels of program service fees or government grants may not see Internet presence as essential to advancing mission goals. 4
Considering the organizational characteristics that did not impact Internet presence as expected is also interesting. For example, many researchers have emphasized that Internet-based technology has created tremendous potential for stakeholder engagement Yet our results indicate that the kinds of organizations we might expect to prioritize this kind of stakeholder engagement have not chosen to do so. We found no relationship between revenue from private fundraising and any dimension of Internet presence, even though a high level of Internet presence would facilitate that engagement.
Generally consistent with Nah and Saxton (2013), our regression results indicate that distinct aspects of Internet presence are related: the more an organization posted on Facebook, the more fans it had. This suggests organizations can increase their visibility on the Internet by being more active. Taken together, our findings reinforce the idea that studying Internet presence reveals important insights about how organizations choose to participate online.
Finally, we have contrasted our three-element definition of Internet presence with earlier research emphasizing those elements (adoption, activity, and visibility) individually. Our findings demonstrate the value of Internet presence as a construct. Understanding variations in human service organizations’ decisions to adopt Internet-based technology, their use of it and their visibility to stakeholders, in addition to variations in the factors affecting Internet presence, provides a more nuanced view of how this group of organizations participates online.
Although it makes several contributions, this study has some limitations. Our sample only includes human service organizations in small cities and rural areas, possibly limiting the generalizability of our findings. We also may have omitted some human service providers in the region due to our sampling approach. In addition, our sample size is relatively small so the power of our many of our regressions is low. Finally, the limited number of organizations without websites and with Twitter prevented us from doing additional inferential analyses.
Our results suggest several areas for future research. Using other subpopulations of nonprofit organizations, scholars should examine whether the factors impacting Internet presence vary across the sector. Studying the Internet presence of specific subgroups in our sample separately, such as public charity funders which had adopted more social media platforms than the other two types of organizations, could provide important insights as well. In addition, scholars should further examine the factors impacting activity as our model did not do a good job explaining variation in the frequency of Facebook posts. Experimenting with alternative measures of activity on different platforms such as the frequency of videos posts on YouTube, blog updates, and website visits would be worthwhile. More research is needed on whether the factors influencing adoption and visibility in our models apply to other types of Internet-based tools. Finally, the data we collected for this study do not include any information taken directly from the staff of nonprofit organizations who have made the choices reflected in our findings about Internet presence. In 2007, Hackler and Saxton assessed the choices nonprofit sector leaders made with respect to Internet presence; given the considerable changes that have taken place since then, a follow-up study would contribute to our understanding of the choices leaders are making today.
Understanding Internet presence as a phenomenon is important in its own right. It provides insights into the extent to which nonprofit organizations view Internet-based tools as valuable and believe online engagement will help them achieve their missions. In addition, Internet presence provides a context for other research about how nonprofit organizations are using Internet-based tools to achieve important goals.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.