You are on page 1of 24

Running Head: ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT

Understanding Organic Post Content and its Relationship

to Active and Subconscious Engagement on Instagram

Reid Cobb, Sophia Tasselmyer and Natalie Wright

Elon University
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
2

Abstract

Our study utilized eye tracking, a technology that monitors eye patterns to determine how

users interact with what they see. By gathering a series of random pictures from the Instagram

account of a local coffee shop, we had five participants go through the eye tracking exercise and

we analyzed the length of time the respondents looked at each image to determine which drew the

most attention. We then compared the length of time a user spends looking a post versus the active

engagement, which is measured by the number of likes and comments of an organic post, to see if

length of time looking at a post correlates with active engagement. Upon analyzing the data, we

found that just because a user is more drawn to a particular post visually, that does not indicate

there will not result in active post engagement.

​ ye-tracking, Instagram
Keywords: E
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
3

Understanding Organic Post Content and its Relationship

to Active and Subconscious Engagement on Instagram

Introduction

The demand of coffee is growing rapidly within the United States and the demand for

caffeine extends to college towns, like Elon University in Elon, North Carolina. Located in

Downtown Elon is The Oak House, one of the few coffee locations within walking distance on

campus that draws in hundreds of students each day. With a target demographic of “broke college

students,” the Oak House needs to stand out from its fellow competitors in the area and a unique

way businesses can promote their products is through social media.

Without large budgets set aside to advertise their products and services, small, local

businesses continue to rely on word-of-mouth and emerging social media opportunities to bring

awareness to their brand. For our research, we looked at how audiences interacted with the posts

they saw before themselves and examined what captured their attention. By utilizing eye tracking

software, we were able to measure a different metric that we can’t usually measure by likes or

comments. By gathering information about length of time a user looks at a photo, we can better

advise our client about what would be worth posting.

Literature Review
Eye-Tracking

Eye-tracking techniques and technology has improved significantly since the beginning

creation of the tool. Andrew Duchowski describes four generations of eye tracking technology,

starting with the first which is an eye-in-head measurement of the eye using scleral contact lense

and electro-oculography. The second generation is photo and video-oculography, then third is an
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
4

analog video combined with pupil reflection. Finally, the fourth generation of eye-tracking is an

advancement of the third because of the addition of computer vision techniques and Digital Signal

Processors (DSP) (Duchowski, 2017).

In order to be as accurate as possible when conducting eye-tracking, the technology needs

to be calibrated correctly for each individual participant. This is done by have the participant look

at a number of predetermined locations on the screen. During this time, the eye-tracker records the

the pupil center and corneal reflections for each of the locations (Halverson and Hornof, 2002).

With this data, the technology will be able to record the user’s fixations during the actual gazepoint

analysis.

Social Media

Social media has pushed its way into the forefront of nearly every person in America and

the impacts it’s had on daily life have been profound. The connections promoted through users

sharing marginal views have increased visibility of issues and shifted the balance of power into the

hands of the masses. Businesses have been able to turn to social media to generate insights, create

targeted product offerings, and stimulate demand (Gaitho, 2018). Among the most popular social

media platforms that currently exist today, Instagram is one of the top picks that businesses choose

to advertise on (Lyfe Marketing).

One billion people use Instagram and 500-million of these users are active every day, contributing

to the 95 million posts and 400 million stories that are uploaded per day (Clarke, 2018). The

videos, images, and captions that are added to Instagram each day make up what is called

user-generated content (UGC). ​UGC is defined in a journal published by Business Horizons as

“the sum of all ways in which people make use of social media ​and is usually applied to describe
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
5

the various forms of media content that are publicly available and created by end-users (Kaplan,

2009).” ​This content is becoming a source that individuals trust to receive information from, due to

the nature of social media bringing simple elements and messages together that consumers thus

relate to and appreciate (Drury, 2008). ​With so much content available at our fingertips, it’s up to

businesses to catch the attention of the users they’re attempting to target. ​Marketing within social

media isn’t about just delivering a message, but also anticipating a return on that deliverance.

Reciprocity

Reciprocity among users on social media is a mutual exchange of a good or service where a

receiver typically feels obliged ​to act in response to receiving something from a giver. Purposeful

and persistent interactions have the potential to create greater shared benefits and even stronger

community dynamics which is the aim of reciprocity (Lewis, 2015). ​Profiles on social media serve

to reflect an individual’s interests based on who they follow and interact with, information that

Instagram utilizes when generating what images they display on a user’s Explore feed.​ Social

media user consumption is often viewed as an individual-centered act that is goal-oriented, which

is the premise behind researcher Peter Pirolli’s “information foraging theory” (Pirolli). This ​theory

aimed to explain the relationship between human and computer interaction and claims incredible

explanatory power for a wide range of intentional online behavior. What Pirolli fails to address in

his theory, which is argued by Pelaprat and Brown, is that not all information consumed is

goal-driven (Brown & Pelaprat, 2012).

Social Media Behaviors

Relatively few people have conducted research in an attempt to better understand the broad

scope of individual internet behavior and is typically discussed on a case-by-case basis, failing to
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
6

provide generalizable insights. Our research aims to further understand social media behaviors and

what motivates people when they see images and if there’s anything about a post that will grab

their attention more than others or if there is intention behind their engagement. The concept of

reciprocity, a common tool relied on by marketers, was developed to try and explain online social

behavior in the “rational choice theory (RCT),” but much like Pirolli’s theory, it only operates

under the assumption that people are using social media as a tool to gain something, either

immediately or within time (Pelaprat). Both theories explain social phenomena that get reduces to

issues of social cooperation, but don’t accurately depict why people behave the way they do on

social media.

Since Pirolli’s proposal of his theory in 1999, social media usage has increased

exponentially, with the average human consuming 34 gigabytes of data, which comes from the fact

that the average American adult spends 12 hours and 7 minutes a day consuming media, thanks to

multitasking (Bilton). According to the Deloitte Digital Democracy Survey, 90% of Americans are

engaging in alternate activities while watching television. Multitasking could be your brother

sitting in a room with the football game on as he’s scrolling through his phone, going from

Facebook to the text message bubble that dropped down, back to Facebook and then to his e-mails

because he didn’t check them yet today. The announcer in the background of the football game

will say something about a terrible play made by the Buffalo Bills and suddenly he’ll be drawn

back into the TV, even though he hadn’t looked up for the past five minutes.

Multitasking

Media consumption has drifted from passive to active, and with multitasking diminishing

task performance due to our finite cognitive abilities, getting messages across to consumers in an
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
7

effective way in which they’ll remember what they see is crucial to media advertising success

(Bardhi). When you walk into the room your brother is watching the game on, you remind him that

we’re out of bread and that he should pick some up on his way home from his friend’s house later

that night. With the noise of the football game, the ding of yet another text message, and a picture

of dream car on his phone screen all happening at once, you know that him remembering bread for

your morning toast is a long shot because of all the stimulus occurring at once. While our research

will have subjects in a controlled environment, we will not be able to test their memory of what

they’re seeing in terms of whether or not the media they see during our study is actually absorbed.

Successful Social Media Marketing

Remembering what the media is saying is among the four primary goals of media messages

which are as follows: to attract attention, to be remembered, to entertain, and to persuade (Ravaja).

For successful business marketing, it’s imperative you’re posting content that’s going to leave a

meaningful impression among users that in turn leads to a business transaction. A study conducted

by Microsoft Corp. found that a person’s average attention span has dropped to eight seconds since

2000, highlighting the effects of an increasingly digitized lifestyle on the brain and become heavy

multitasking individuals. Though this distraction by multiple streams of media has taken its toll on

the human ability to focus, our ability to multitask has improved (McSpadden, 2015). Even though

we aren’t spending as much time looking at each piece of information that is heard or said, that

hasn’t shut the door on our ability to absorb or remember. In cognitive psychology, information

that is unconsciously processed has been equated with subliminal information processing, meaning

it’s below an individual’s threshold for conscious perception (Bargh).


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
8

Despite the lack of consensus on what qualities define unconscious processes, several

theorists have postulated that the unconscious is the source of behavioral impulses, which in turn

generate a conscious result (Bargh). When you spoke to your brother, he didn’t give any indication

he even knew you were in the room but following your request, you asked him to repeat what you

said, and he was surprisingly able to recall verbatim what you weren’t even certain he heard.

Bargh and Morsella proposed that unconscious processes are defined “in terms of their

unintentional nature and the inherent lack of awareness is of the influence and effect of the

triggering stimuli (because nearly all naturally occurring stimuli are supraliminal)” (78). When

someone instinctively grabs their phone when they have a free moment, as 71% of people between

the ages of 18-25 have reported doing, they are almost guaranteed to be multitasking until the next

moment outside that digital world brings their nose out of their phone. Businesses are tasked with

the difficult responsibility of not only forcing their presence, but also leaving an impact.

Our study is going to focus on Instagram and the power of the unconscious and conscious

thought processes. When users are on Instagram, we are curious about what they’re most drawn to

visually and we will be able to explore that even more through the utilization of Elon University’s

eye tracking system. The way that eye tracking works is each eye data observation is translated

into a set of pixel coordinates and from there, the presence or absence of eye data points in

different screen areas can be examined (Eye Tracking). There hasn’t been enough previous

research to generate an idea of what our results may turn up.

This study presented the following research question:

RQ: How does organic post content affect the relationship between active and

subconscious engagement on Instagram?


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
9

For the purpose of our study, organic posts are considered any unpaid Instagram post from

@elonoakhouse. Additionally, content is defined by three mutually exclusive and exhaustive

categories: people, products and text. A post that has a human face is placed in the people

category. If the product has a product that is sold by the Oak House, it is categorized as a product.

And finally, a post that includes text in the form of a digital graphic or a picture of a sign or flyer,

this post is considered part of the text category. If a post includes a combination of the categories,

the item that fills the larger percentage of the photo will determine the content category.

Methodology
Data Collection

This study required three data sources. The research included an eye-tracking session, a

survey and a collection of the engagements for the posts that were used during the eye-tracking

portion. The three research methods were completed by five participants who were students at

Elon University.

First, the students entered the room with the computer and Gazepoint eye tracking

technology. They began by adjusting their position at the computer and completing a series of

calibration activities. Once the setup was completed, the participants were given 30 seconds to

look at a six by five gridlike display of Instagram posts from @elonoakhouse. The grid consisted

of six columns and five rows of square posts from Instagram and is provided in Appendix A. The

posts were chosen randomly and the post placement on the grid was random. So in the provided 30

seconds, participants were able to look at the 30 displayed pictures. During this time period, the

participants were asked to not touch the mouse, but to only view the content on the screen.
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
10

With eye-tracking it is important to understand that there are two assumptions that are

made when analyzing the gathered data. The first assumption is that fixation is the same as

attention. This means that as the participants eyes move, it is assumed that this is because their

attention is moved. The second assumption is that gaze is the same as focus. So, this means that

what the participant is looking at, is the same as what they are thinking about. These assumptions

ignore the fact that participants might “zone out” during the eye-tracking process because of

disinterest or other internal distractions that were beyond the scope of this study. But, this was not

a huge concern for the study because the eyetracking portion was only 30 seconds long. Finally, a

follow-up survey was used in order to help provide support to the eye-tracking portion, for the

assumptions that were made.

Upon completion of the eye-tracking portion of the investigation, participants were asked

to complete a survey that included seven questions. The first section of questions was used to get

information about the participant’s use and experience with Instagram, and the second section was

used to get more details about their experience during the eye-tracking. Participants were asked

about the content that they saw and the post that they remembered most vividly.

Data Cleaning and Analysis

After the eye-tracking and survey were completed by all of the participants, the data

analysis portion of the study was started. In order to begin, the grid image used for the eye-tracking

portion of the study was labeled so that each post had a unique letter and number combination.

This was necessary for easy discussion and use of the data. The columns were labeled with letters

from A to F and the rows were labeled with numbers one to five, which is shown in Appendix B.

Using this key, the number of likes and comments for each post ID were collected into a
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
11

spreadsheet by looking directly at the Instagram metrics for all of the posts. Additionally, a column

was added to categorize each post with a content category (people, products and text). This was the

only data collected from Instagram.

Next, the data provided from the Gazepoint tracking was cleaned and analyzed. The User

Fixation files and videos were the only output files that were used. From the fixation data, the

columns for FPOGX (the x-coordinate of the fixation as a percentage of the screen width), FPOGY

(the y-coordinate) and FPOGD (the duration of the fixation) were used. These coordinates and the

videos for each participant were used to determine which image on the grid that the participants

were looking at. It is important to note that these assumptions for which image the participant

looked at might not always be completely accurate because the Gazepoint device relies on the

participant’s calibration test. For example, on occasion the point of fixation would be on the small

line in between images, so coders used their best judgement for what image the participant was

looking at. Also, note that the recording of the participants’ eyes blocks image A1 in the top left

corner, so the coder’s best judgement was used whenever the gazepoint visual path went to that

portion of the grid. The resulting cleaned data consisted of the visual path order, the x and

y-coordinates, the duration of fixation and the image ID for each of the five participants. With this

data, an aggregation for the duration of each image was conducted using a sumif equation for each

participant so the total duration of fixation for each image was calculated. This was combined with

the Instagram data to create a master spreadsheet of all necessary data.


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
12

Findings/Results

Figure 1: Instagram Post Data and Eye-Tracking Results

The scatterplot shown above (fig. 1) displays the results for the eye-tracking experiment.

On the Y axis, there is the average time watched for each image, and on the X axis is the amount

of engagements (likes and comments) for each image from Instagram. The purpose of the graph is

to see if more interesting photos (more watch time) correlate with more engagements. Evidently,

the trendline only has a slight slope upward, indicating that there is almost no correlation between

the two variables. Image A3, which got almost not engagements, had almost the same amount of

watch time as image D1, which had one of the most engagements of the group.

When each of the images was ranked from 1-30 according to time watched and

engagements, there was also a discrepancy. Taking the rank for time watched, i.e. #5, and then
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
13

subtracting it by its rank of engagements, i.e. #24, and then finding the absolute value reveals the

difference in ranks. In an ideal scenario, there would be no difference between the two numbers.

However, we found there was an average discrepancy of 11.267 placements of rank (out of 30

ranks possible).

In addition to this, we matched up an image’s time and engagement rank, with the rank that

it would be in an ideal scenario (Actual Time/Eng.). This means for an image with the time

ranking of 15, and the engagement ranking of 8, we would match them with the engagement

ranking of 15, and the time ranking of 8, just for comparison’s sake. The numbers we matched

them with are the same numbers they already have, but taken out of order to create the ‘ideal’

ranking. We also had a theoretical ranking, which is the same concept except instead of using the

numbers that already existed, we created them by finding the difference between the minimum and

maximum in each category and dividing it by 30 (the number of images). Then once we found the

average difference between each rank, we multiplied it by the inverse rank of that number (1

would be 30 and 2 would be 29, etc) to get the theoretical rank. From there we placed them with

the ideal rankings.

Post Average Time Engagements


A1 1.223 59
A2 0.932 28
A3 0.886 16
A4 0.789 216
A5 0.553 116
B1 1.394 233
B2 1.501 51
B3 0.825 140
B4 0.615 54
B5 0.524 79
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
14

C1 1.401 38
C2 1.006 73
C3 0.977 73
C4 0.340 79
C5 0.592 26
D1 1.009 210
D2 0.928 54
D3 1.346 27
D4 1.165 164
D5 0.485 92
E1 0.628 89
E2 0.566 75
E3 0.786 20
E4 0.340 87
E5 0.505 80
F1 1.404 118
F2 0.615 114
F3 0.861 163
F4 0.566 104
F5 0.835 185
Table 1: Average Time and Engagements for Each Post
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
15

Post Average
B2 1.500986
F1 1.403982
C1 1.400862
B1 1.394138
D3 1.345812
A1 1.222678
D4 1.1647
D1 1.00911
C2 1.00589
C3 0.97692
A2 0.931796
D2 0.928374
A3 0.886388
F3 0.860526
F5 0.83485
B3 0.824972
A4 0.789378
E3 0.786052
E1 0.627538
F2 0.614692
B4 0.61451
C5 0.5919
E2 0.566136
F4 0.566066
A5 0.553204
B5 0.524068
E5 0.50458
D5 0.485152
C4 0.339712
E4 0.33952
Table 2: Average Time Ordered by Most Time Watched
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
16

Post Engagements
B1 233
A4 216
D1 210
F5 185
D4 164
F3 163
B3 140
F1 118
A5 116
F2 114
F4 104
D5 92
E1 89
E4 87
E5 80
B5 79
C4 79
E2 75
C2 73
C3 73
A1 59
B4 54
D2 54
B2 51
C1 38
A2 28
D3 27
C5 26
E3 20
A3 16
Table 3: Ordered by Most Engagements
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
17

Figure 2: Slopegraph Comparing the Two Ranks


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
18

Average Engagem Time Actual Actual Theo. Theo.


Post Time ents Rank Eng Rank Time Eng. Time Eng.
A1 1.223 59 6 21 0.615 163 0.387 180.833
A2 0.932 28 11 26 0.524 104 0.194 144.667
A3 0.886 16 13 30 0.340 89 0.039 130.200
A4 0.789 216 17 2 1.404 79 1.123 101.267
A5 0.553 116 25 9 1.006 38 0.852 43.400
B1 1.394 233 4 1 1.501 185 1.161 195.300
B2 1.501 51 1 24 0.566 233 0.271 217.000
B3 0.825 140 16 7 1.165 79 0.929 108.500
B4 0.615 54 21 22 0.592 59 0.348 72.333
B5 0.524 79 26 16 0.825 28 0.581 36.167
C1 1.401 38 3 25 0.553 210 0.232 202.533
C2 1.006 73 9 19 0.628 116 0.465 159.133
C3 0.977 73 10 20 0.615 114 0.426 151.900
C4 0.340 79 29 17 0.789 20 0.542 14.467
C5 0.592 26 22 28 0.485 54 0.116 65.100
D1 1.009 210 8 3 1.401 118 1.084 166.367
D2 0.928 54 12 23 0.566 92 0.310 137.433
D3 1.346 27 5 27 0.505 164 0.155 188.067
D4 1.165 164 7 5 1.346 140 1.007 173.600
D5 0.485 92 28 12 0.928 26 0.736 21.700
E1 0.628 89 19 13 0.886 73 0.697 86.800
E2 0.566 75 23 18 0.786 54 0.503 57.867
E3 0.786 20 18 29 0.340 75 0.077 94.033
E4 0.340 87 30 14 0.861 16 0.658 7.233
E5 0.505 80 27 15 0.835 27 0.619 28.933
F1 1.404 118 2 8 1.009 216 0.890 209.767
F2 0.615 114 20 10 0.977 73 0.813 79.567
F3 0.861 163 14 6 1.223 87 0.968 122.967
F4 0.566 104 24 11 0.932 51 0.774 50.633
F5 0.835 185 15 4 1.394 80 1.045 115.733
Table 4: Results for Actual Time/Eng. And Theoretical Time/Eng.
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
19

Survey Results
The survey portion of the investigation was used to get an understanding of the

participants’ relationship with Instagram and to understand how familiar they were with the

platform. Additionally, the survey was used to understand any lingering images from the provided

grid image. The results of the survey showed that the majority of participants were familiar with

Instagram before beginning the eye-tracking, and that each of the participants remembered

different types of content.

The majority of participants have a personal Instagram account so they are familiar with

what is offered on the platform and how it is used. Additionally, two of the five participants have

access to an Instagram account for a business or organization, which shows that they have another

relationship with the type of content posted. Overall, they enjoyed the posts from @elonoakhouse

because they liked seeing the products offered by the shop, and they enjoyed seeing familiar faces

in the posts. Three of the participants said that the most memorable posts were the ones that

included products and two of the participants mentioned that they liked posts with filters and

contrasting colors. Overall each of the participants felt like they had differing experiences when

viewing the content.

Conclusion
The purpose of this study was to determine the relationship between and active and

subconscious engagement by analyzing organic post content on Elon Oak House’s Instagram

account. In order to do so the engagements (likes and comments) from the selected Instagram

posts, were compared to the duration of fixations for each image obtained during an eye-tracking

investigation. The final result was inconclusive and did not show any connection between the

number of comments and likes and the duration of look at the post.
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
20

The main limitation on this study was the sampling size. Because of time and resources,

only five participants were able to be analyzed. This limits the total amount of data. The other

limitation was the set up of the eye-tracking process. For ease of analysis, a grid of images was

used do that the using would not have to scroll in order to see each image. For future research, and

when technology becomes more advanced, it would be best to match the normal habit of scrolling

through an Instagram feed on a mobile device to look at the images. This would best mimic how

Instagram is used and would result in more realistic findings.


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
21

References
Bargh, John A., Morsella, Ezequiel. (2008). The Unconscious Mind: Perspectives on

Psychological Science. ​Association for Psychological Science,​ 3(1).

Brown, B. & Pelaprat E. (2012). Reciprocity: Understanding Online Social Relations. Retrieved

from doi:10.5210/fm.v17i10.3324

Clarke, T. (2018) “24 Instagram Statistics That Matter to Marketers in 2019.” ​Hootsuite Social

Media Management.​ Retrieved from blog.hootsuite.com/instagram-statistics/.

Drury, G. (2008) “Opinion piece: Social media: Should marketers engage and how can it be done

effectively?” ​Journal of Direct, Data and Digital Marketing Practice, ​9(3). Retrieved from

https://doi.org/10.1057/palgrave.dddmp.4350096

Duchowski, A. T. (2017). ​Eye Tracking Methodology: Theory and Practice. ​Switzerland:

Springer International Publishing AG.

Gaitho, M. (2018) “What is the Real Impact of Social Media?” ​Simpli Learn.​ ​Retrieved from

https://www.simplilearn.com/real-impact-social-media-article

Halverson T. & Hornof A. J. (2002). Cleaning up Systematic Error in Eye-Tracking Data by

Using Required Fixation Locations. ​Behavior Research Methods, Instruments &

Computers, 34,​ 592 - 604.

Kaplan, A. M. & Haenlein, M. (2009). “Users of the world, unite! The challenges and

opportunities of Social Media.” Retrieved from doi: 10.1016/j.bushor.2009.09.003

Lewis, S. C. (2015). “Reciprocity as a Key Concept for Social Media and Society.” ​Social Media

Society​, 1(1). Retrieved from doi:10.1177/2056305115580339


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
22

McSpadden, K. (2015). “You Now Have A Shorter Attention Span Than a Goldfish.” Time

Health.

Retrieved from http://time.com/3858309/attention-spans-goldfish/


ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
23

Appendix A
Grid Image Used During the Gazepoint Eye-Tracking Study
ACTIVE & SUBCONSCIOUS INSTAGRAM ENGAGEMENT
24

Appendix B
Labeled Grid Image Used During Analysis

You might also like