Getting the Appropriate Returns on Investing in Social Media Activity
Most (>98%) higher education institutions (HEI) actively use social media investing time, effort and money in maintaining their presence. Based on the roughly 8,500 self-identified, US-based, higher education social media specialists on LinkedIn, at least $50 million in annual salaries, alone, is spent on social media. Add in the large contingent of part-time and occasional higher education social media practitioners and investment likely doubles. Spending jumps again when advertising and promotional activities are included. US higher education social media activity is, conservatively, a multi-hundred million dollar a year activity.
It is always prudent to re-examine on-going maintenance investment programmes to ensure they continue to deliver the anticipated returns. For US HEIs, social media investment yields a significant tangible impact. US universities and colleges have generated 16 million tweets, amassed 7.8 million photographs on Flickr, produced 1.8 million Instagram posts, and created video content that has been viewed 2.2 billion times on YouTube.
Assessing returns on social media investments needs context. The channels can provide analytics and these can be compared with the target response or activity levels, buy a vital context is answering the question: ‘what do other schools do?’. Or, perhaps more helpfully – how do we compare against other schools like us?
This post reports the results of our survey of 4,245 US university and college home pages to determine which social media channels are accessible to visitors and to understand which combinations of social media are in use. The survey results help answer: what do other schools do?
A subsequent post will analyse and review activity data (tweets, posts, likes, pins etc.) for the principal social media channels the 4,245 surveyed US HEIs use. The post will also provide clear benchmarks for social media activity levels.
Why We Embarked on This Project
During a recent assignment, we audited the social media links for 30 websites at a higher education institution. The exercise highlighted two issues. First, links to social media channels (mainly, personal LinkedIn and Facebook pages) can proliferate when content editing is highly devolved potentially giving rise to reputational and other risks. Second, dozens of institutional, departmental and other official social channels are brought into existence, without it being clear what value their ongoing maintenance brings.
The project prompted the broader question of what social channels do higher education institutions use, what sort of activity levels do they actually achieve and what are appropriate expectations about audiences (likes and follows) and interactions (views)?
Data Collection Process
While the assignment prompting our research was carried out for a non-US client, we decided to use US university and college data. US institutions have a broad range of student populations, such that the results for smaller institutions can used as acceptable proxies for faculties, departments and research centres within larger US and non-US HEIs. For example, 250 of the surveyed institutions have 100 or fewer students, 90 have 50 or fewer students. If your social media responsibilities are at the faculty or department level, our detailed results data will be relevant.
To collect the data, we pointed our website quality assurance software service at the home pages of all the active colleges and universities listed in the US National Center for Education Statistics Integrated Postsecondary Data System (NCES IPEDS) database. We used the 2016 dataset, cleaning the data to reflect institutions that had closed, merged or changed name since the data was originally collected. We set 260 institutions as inactive and excluded a further 72 entities as purely administrative offices.
The resulting data covers 4,245 active US colleges and universities. To aid institutional comparison we cross-tabulated the active institutions with the US Department of Education's Office of Postsecondary Education’s Database of Accredited Postsecondary Institutions and Programs. The concatenation of these two datasets allows social media activity comparisons across institutions accredited by the same bodies.
We additionally updated and corrected the main institutional URLs in the IPEDS database to reflect migrations to HTTPS from HTTP and the ‘swapping back and forth’ between www. and non-www domain addresses.
From the clean set of institutional URLs, we checked each HEI’s home page to identify social media links visible to a human visitor through a recognised icon (image) or text link. If a link was commented out or otherwise not visible to a human visitor or screen reading software we exclude it.
Aside from religious educational institutions for whom social media is not, seemingly, a relevant communications channel, 99% of the surveyed sites include social media links on their home pages.
Following collection and cleaning (mainly re-constructing full URLs from relative ones) of the resulting list of social media URLs we visited the underlying social sites collecting publicly-available data about each account’s activity. Tumblr doesn’t publish statistics, so there are none in our dataset. There are, currently, no statistics for LinkedIn, as data collection entails being signed into a user account and collecting the data in that way that is inconsistent with LinkedIn’s terms of service.
We collected the data over a period of 10 days spanning mid-March 2018.
You can download a sample of the data gathered for 99 institutions here: https://www.eqafy.com/dataset-sample.html
The 4,245 institutions use 15,174 unique social media accounts. Facebook dominates US HEI social media, with Facebook representing almost 25% of all the accounts in use. Twitter represents just over 20%. Figure 1 shows the relative market shares and the data for each social channel should as Facebook accounts represent 23.5% of the 15,174 unique social media accounts discovered, etc. We note that seven of the 12 channels have minimal market share and there may be benefits in switching activity to one of the four main social channels.
Many institutions (mainly public colleges and universities and private for-profit colleges) operate from multiple locations. The IPEDS database assigns each location a unique ID and we have followed that practice in our dataset. Reflecting each institution’s specific circumstances, some HEIs use the same social media accounts for all locations, while others use distinct accounts for each location. We speculate that this decision is driven by the degree to which individual locations have a distinct identity and resource availability.
Account deployment or allocation practices also vary by channel. For example, different locations may use their own Twitter and Instagram accounts, but employ a common Facebook account for all locations. The total number of all accounts for all locations bumps the count up to 18,150 accounts in use.
Lesson One. How do the social channels we’ve chosen to use compare with those of our peer groups? Our data allows us to compare sizes of institutions by number of students, location (state), accreditation body, Carnegie classification, religious affiliation and a number of other factors. And, against all those factors four social media channels predominate: Facebook, Twitter, YouTube and Instagram. HEI social media specialists have gone where the audiences are and it may be time to cut loose from legacy social media channels, where possible.
The distribution of social media accounts across the all 4,245 US HE institutions is heavily weighted to the big four (Facebook, Twitter, YouTube and Instagram). In figure 2 we show the percentages of US 2-year and 4-year colleges and universities with each type of social media account. The bar chart should be read as 96% of 2-year and 4-year US HEIs with social media links on their home page, have a Facebook account, 86% have a Twitter account, etc.
Interestingly plotting the data for public, private non-profit and private for-profit institutions shows broadly the same distribution and concentration of social media power with Facebook, Twitter, YouTube and Instagram. While we believe our data capture for links to SnapChat is robust, we suspect that as a mobile-only social channel links to SnapChat are under-represented on websites.
Lesson Two: Social media publishing is highly decentralised. What does the data show for departmental-size organisations? We divided the 4,245 US HEIs in quintiles, based on total student population. The bottom quintile represented the 20% of institutions with a student population of 300 or fewer – that’s similar to departments or other sub-units at larger institutions. It turns out that these organisations favour the same four social media channels along with a strong residual interest in Google+. Twenty three percent of smaller schools still use Google+
Given the distribution shown in Figure 2, we also looked at the number of different channels that institutions choose to deploy on their websites and this is plotted in Figure 3. The data should be interpreted that 24.5% of all institutions surveyed placed links to four different social media channels on their website home pages.
Comparing the data from figures 2 and 3 suggests that US HEIs focus social media activity through a combination of Facebook, Twitter, YouTube and Instagram. This pattern seems to reflect the complementary nature of the four channels: Facebook postings, Twitter ‘breaking’ events/notifications, YouTube video and Instagram images.
Lesson Three. Social media follows a hype cycle. How has social media adoption actually shaken out in US higher education? From the initial wave of over-enthusiasm, through the pit of despondency and back to the plateau of dogged persistence, HEI social media has tracked its audience. Audiences don’t want to follow a dozen different channels, so activity is directed at, and likely reaches, target audiences through four. There’s likely a case to be made for ending activity outside the four main social channels unless it clearly addresses a specific requirement.
We further decomposed the data from figure 3 to determine the most popular combinations of social media channels for groups of three, four, five and six channels. The results are in table 1. As the Google+ user community is static or in decline it will be increasingly difficult to make a case for maintaining content on this channel and we would anticipate institutions consolidating back to four or five core social channels.
YouTube and Twitter make account creation dates readily available, providing the basis for analysing social media adoption patterns. Initial YouTube accounts were being created from July 2005 onwards, with the first US HEI Twitter accounts opening in March 2007.
Figure 4 plots the number of new accounts being created each month for Twitter and YouTube. While the two curves look radically different, the total number of Twitter accounts opened is only 17.6% more than for YouTube.
We speculate that similar growth patterns exist for the other social channels US HEIs use – perhaps without Twitter’s dramatic adoption spike. The curves suggest that early adopters experiment with the new medium leading to a surge of interest as awareness spreads followed by a slow decline in new account creation.
Experimentation seems an appropriate response to new social channels. We note the relatively large number of ‘abandoned’ social media experiments we encountered as we cleaned the raw data, hinting at neglect rather than creativity. Barring the need to comply with recordkeeping regulations, closing no longer-maintained social media pages would be consistent with the practice of ensuring website content is up to date.
Lesson Four. Some institutions seem to have runaway success with social media. Who are they and what does success look like? Success measures include the number of likes or followers that an HEI social media account attracts. For example, Harvard University’s Facebook page has attracted 5.2 million likes, almost three times more than the number two HEI. On Twitter, Stanford University has over 590,00 followers, which puts it in second spot. On Instagram, Coastal Carolina University is higher education’s number one poster with over 5,800 posts. On YouTube the second most subscribe higher education channel is Berklee College of Music with just over 500,000 subscribers. In the next post we’ll examine activity levels in more detail.
What Else We Learned Along the Way
Help Visitors Find Your Social Channels
The majority of higher education institutions use social media icons placed in the footer of their main home page to direct visitors to their social channels. A small number (< 10%) places social media links in page headers. We would be interested in knowing which position generates a higher click through rate.
Some sites have chosen to ‘hide’ the actual URL of their social media channels via an internal redirection process. It is not entirely clear why this is a good idea and we found this implementation particularly prone to resulting in ‘broken links’, as discussed below.
A small, but perverse, group of institutions forces site visitors to click through to social media directory pages and then select from a list of social media channels. In our opinion, this approach makes visitors work too hard to find an institution’s main social media channels. We completely understand the need for social media directories (and our work suggests most institutional directories are incomplete), but a more natural flow is to browse rather than dive directly into an index or table of contents.
A small number of sites are confused over the difference between social sharing and following, which typically reveals itself with “Add-this” or “Share-this” social media buttons, but no links to the institution’s own social media accounts that visitors may wish to follow.
There is confusion over the use of “channels” and “users” in YouTube, some institutions opt for one others opting for the other. It likely makes more sense for individuals that you want to engage with your organisation to be able to see all of the ‘channels’ you have available rather than being forced to search for all the relevant channels.
Broken, Dead and Legacy Social Content
Our best estimate is that 2%-3% of social media links are broken – clicking on a link either generates an 404 error or more likely leads to a message on the destination social media site that the material sought is no longer available. In this exercise we only looked at home pages, but our client work confirms that the problem of dead/broken social media links gets worse the deeper visitors venture into a website.
Dead or broken links are ‘complemented’ by links to legacy social media channels. Channels no longer being actively maintained that should be archived or for which the content could be transferred to an actively supported channel.
As noted above, four channels dominate higher education social media – Facebook, Instagram, Twitter and YouTube. It may be difficult, but a consistent naming convention across these channels makes it easier for them to found, followed and shared.
Before carrying out this exercise we would have questioned the need to declare that an account is an institution’s ‘official’ account. No more. There is little to prevent parody, tribute or other types of account being created and visitors need help understanding that they have reached an institution’s actively maintained account.
Social SEO tip. Use the institution’s official name somewhere on each social media account. You may know that everyone calls your school “PDSA”, but we don’t. We can’t find you on Instagram or YouTube, as a result. And, that’s not the outcome marketing and communications teams are looking for.
About the Data
For each social account we discovered, we collected the relevant metrics, where a social channel makes these visible to human website visitors.
The associated statistics are accurate of mid-March 2018, when they were collected. The data is dynamic as accounts add or lose followers and make additional posts, so there will be a difference between the data we captured and today’s number of posts, pins, followers or followings, but in aggregate the differences are likely to be immaterial.
We collected a total of 67,800 data points. We’ve diligently checked the data, but please feel free to bring any omissions or errors to our attention.
If you've read this far, you may be interested in downloading sample data for 99 institutions here: https://www.eqafy.com/dataset-sample.html
The complete dataset is available for purchase as a downloadable Excel workbook.
eQAfy helps higher education institutions manage their web estates and helps websites be error-free and
deliver their intended user experience and meet their marketing and communications objectives