Are All Social Media Metrics Created Equal?

There is more to analyzing social media posts than just counting the number of likes a post might have or seeing how many views your video may have received. The significant growth of the data analysis industry has extended to social media, and the demand for its implementation is crucial for executing digital marketing strategies. Optimization techniques in social media marketing are frequently implemented through the use of metrics.

Before diving deeper into the application of metrics in the world of social media, it would be appropriate to define what a metric is.

A metric can be generally defined as measurements used to evaluate performance. In the context of social media, metrics can give us insight to how well a particular post, channel, or campaign is performing. Some of the most commonly observed ones in measuring social media content are as follows:

Engagement	The number of interactions in a post. These interactions can come in the form of likes, comments, shares, etc.
Impressions	The number of times the content has appeared to users.
Reach	The number of unique users that saw the content.
Click-Through Rate (CTR)	Percentage of users who clicked on a link compared to its audience exposure. Usually divided over impressions.

The overall pool of metrics used in social media is vast and highly dependent on objectives, platforms, and the level of analysis. Today, we aim to answer the specific question of which social media metrics carry more weight when it comes to analyzing performance.

We will take a look at each of the most prominent metrics used to analyze social media performance and analyze them on a platform-by-platform basis.

How Social Platforms Weigh Metrics Differently

Engagement is likely the most commonly used metric to measure performance on every social platform. This metric generally accumulates all the different interactions that are unique to their respective platform to create a holistic measurement of content relevancy and how well the audiences are resonating with a post.

On Meta platforms, specifically Facebook and Instagram, engagement is considered a significant variable when it comes to prioritizing a post's ability to spread to a broader audience. Among different types of engagement, both platforms heavily prioritizes likes, comments, and shares when it comes to fueling this visibility. Engagement on Facebook is also thought to be more personal, considering the network's tendency for personal connections with friends and family.

On Instagram, as a platform with a greater emphasis on visuals and a robust influencer culture, engagement takes a similar form to that on Facebook (likes, comments, and shares). However, on this platform, engagement may be derived more from what the viewer finds visually appealing.

The metrics themselves, when it comes to engagement, may start to take a different form when we start comparing platforms such as LinkedIn and X (formerly Twitter).

For example, LinkedIn includes 'Post Clicks' and 'Follows' as additional metrics when calculating the total engagement for each piece of content. These 'Post Clicks' differ from Meta's 'Link Clicks' as they record not only direct user interactions with any given links but also interactions with any clickable elements within a post (such as images or profiles). The trade-off with this method of recording clicks is that it can inflate the click-through rate (CTR) and complicate efforts to standardize clicks across different platforms. This is an important factor to consider when understanding why total engagement may appear in larger quantities for this platform. When it comes to weighing various forms of engagement metrics, LinkedIn's 'Follows' can indicate a higher level of engagement, as they demonstrate users' intent to see more of your content in their feeds.

At its current stage of rebranding, X may be the most challenging platform to analyze. Formerly known as Twitter, the platform's users have become accustomed to its posts, which were normalized as 'Tweets,' aligning well with the platform's name. Due to its ongoing changes, X might undergo a renaming process that incorporates metrics along with post names. Content on X consistently prioritizes trending topics and timely information, resulting in a high level of engagement through the 'Retweet' metric. They are often used as a measure of how viral or shareable a tweet is, and they contribute to the overall engagement level of a tweet. Clicks on X may carry less weight compared to the previously mentioned platforms, as recent studies indicate that this platform has the lowest average percentage in click-through rate (CTR). This could be a result of the skewness towards higher impressions per tweet. Additionally, differences in ad performance between platforms might also contribute to this phenomenon.

Navigating this sea of metrics is a complex yet necessary task. With each platform having its unique metrics landscape, understanding the significance and weight of these metrics can provide valuable insights tailored to the context of each platform.

Which Metrics Should I Prioritize?

When evaluating which metrics to prioritize in regards to measuring success for a campaign or post, it's highly important to factor in key performance indicators (KPIs). KPIs help measure the success of content as they are quantifiable metrics that provide a clear understanding of how well a campaign or post is performing in relation to those goals. It's important to note that all KPIs are metrics, but not all metrics are KPIs.

When thinking about a social media post, metrics can cover a wide range of engagements and impressions. Let's say we're creating a promotional post with the goal of attracting more users to visit a website through the link in the description. A good KPI to measure the success of our objective would be clicks (or CTR). Other forms of engagement, such as likes or comments, might not be weighted as heavily as they are usually not going to correlate to an increase in website traffic. Being able to align social media posts with an objective allows you to transform any of these metrics into a measurable KPI. These will greatly enhance your understanding of how effective your social media strategy is and whether you're reaching your goals.

Below are a few more examples of common social media objectives, along with some of their corresponding KPIs:

Objective: Website Traffic	KPIs: Clicks, Click-Through Rate (CTR)
Objective: Brand Awareness	KPIs: Reach, Impressions
Objective: Showcase New Product	KPIs: Likes, Comments, Shares
Objective: Build Community	KPIs: Page Likes, Follower Increase

Our experts at Random are ready to create data-backed digital campaigns for your business! Reach out to us below to learn more about how we can work together.

Random Chart of the Month - What Makes THE Top World University Rankings?

Brandon Kim

June 2023

With graduation season in full swing, this was the perfect opportunity to bring out our Random Chart of the Month to discuss the Times Higher Education (THE) World University Rankings!

The Data

THE is known for annually providing a comprehensive list of university rankings, which ranks institutions based on different factors. The formula for the ranking given to schools from all over the world in this list featured multiple factors such as teaching quality, research influence, international outlook, staff, and more.

Based on the recent scores provided by THE, the top three schools by rank include the University of Oxford at first, Harvard University at second, and a tie for third between the University of Cambridge and Stanford University.

This month's analysis will dig deeper into some significant scoring factors contributing to these schools climbing the ranking ladder.

We will use a similar modeling method from our previous blog last month when we analyzed CO2 emissions from vehicles. For the context of this university dataset, we will want to find some significant predictors that propel a university's overall score.

The Analysis

Contrary to the previous iterations of our Random Chart of the Month, we wanted to initiate our analysis portion before the visualization to contextualize the backbone of our modeling method.

When we worked on a multiple linear regression model from our last blog, we introduced independent and dependent variables, which both construct the contents of this data mining technique. As a refresher, independent variables are our set of predictors that influence the value we are trying to predict, THE's overall university score, also known as our dependent variable. We call it a Multiple Linear Regression because we have more than one independent variable, but any simple or multiple linear regression models will generally have one dependent variable.

When fitting the most optimal model, which sets THE's overall university score used for ranking the top schools as the target of our prediction, we can construct this index to define our variables:

Independent Variable(s)
Research	Derived from various factors reflecting the university's strength of research.
Citations	Measures impact of citations received by research publications affiliated with that university.
Industry Income	A higher value represents stronger connections and financial support from the industry (applied research, knowledge exchange, and technology transfer) for the university.
Student:Staff Ratio	Ratio of the number of students to academic staff or faculty members.
% International Students	The proportion of students that come from countries other than the country in which that university is located.
Dependent Variable
Overall University Score	THE's complete university score to rank their top schools annually.

As we progress with our analysis, we will be able to conclude the statistical impact that these sets of independent variables have in contributing to a high overall university score and what component allows schools such as Oxford, Harvard, Cambridge, and Stanford to score at its peak.

The Chart

Having utilized a ranked bar chart in our previous blog, which used a similar modeling method, we wanted to showcase another visualization that can help understand the level of impact a set of predictors can have on what we are trying to predict. This time, we constructed another bar chart that targets the significant, independent variables we mentioned in our analysis and their per-point impact on the university score:

After verifying the necessary assumptions required to run a multiple linear regression model on this data, we can estimate coefficients for each of the predictors we used for modeling THE's university scores.

The values at the apex of each bar represents the coefficients we estimated for the predictors. One interpretation in our example would be: for every percent increase in a school's international students proportion, we expect that school's overall score to increase by 4.921 (while keeping other variables constant). If we were to write out the full estimated model as an equation, we would construct this:

From this equation, it becomes a matter of plugging in values for these variables to estimate where a school might end up in its overall score.

Based on the magnitude of each coefficient, we can see that a change in the proportion of international students in a school impacts the overall score the highest. Internationalization appears to matter a great deal when considering an increase to a school's score. This aligned with schools ranked in the top ten, which averaged about 30% of their student base coming from out-of-country.

Another variable worth noting is the school's student-to-staff ratio, which seems to carry a negative coefficient value in our model. This indicates an inverse effect on a higher overall score when that ratio increases - signifying that a smaller student-to-staff ratio is better for a more significant score. A smaller ratio could indicate an increased individualized learning experience, mentorships, and more.

Although these variables proved to be rather significant in deciphering how THE constructs their university scoring, plenty of other factors likely establish the top schools' tier. They are still great indicators of why you would expect some of these institutions to make the top in various rankings consistently.

Continue to tune into our ˜Random Chart of the Month' each month! We're experts at analyzing all kinds of data, but especially social media. Let us help you build a social media campaign backed by data and results. Reach out to us below!

Random Chart of the Month - Which 2022 Vehicles had the Most CO2 Emissions?

Brandon Kim

April 2023

As we begin to close the curtains on Earth Month, this is an excellent opportunity to bring in April's iteration of Random Chart of the Month. This time, we will analyze the publicly known Fuel Economy data for vehicle manufacturers in 2022 provided by the Environmental Protection Agency (EPA).

The Data

Since the mid-1970s, the United States government has been collecting and publishing Fuel Economy data to discern the availability/cost of oil and the environmental impact on emissions. The EPA measures and tests fuel economy and emissions, which is then publicly available on their government website.

We will be looking specifically at the 2022 iteration of the Fuel Economy data. It covers a wide range of variables for each recorded vehicle for that year, including CO2 (carbon dioxide) emissions, which we are highly interested in this analysis. We can visualize this data relating to how much vehicles emit based on various features such as brand and fuel type. We can then perform some data mining techniques to see what other vehicle components play significant roles in how much CO2 a car emits.

The Chart

When evaluating this extensive dataset, three different variables measured CO2 emissions by each vehicle: city, highway, and combined. We also had a categorical variable naming which manufacturer belonged to that vehicle. With this information, we were able to discover, by a multiple bar chart, which manufacturer averaged the most CO2 emissions (for both city and highway):

In 2022, Rolls-Royce averaged the highest among all manufacturers in this data based on these averages for combined CO2 emissions for both city and highway driving cycles. On the lower end, we had Mitsubishi Motors Corporation averaging the least combined CO2 emissions.

The Analysis

Although it is fun to see how different manufacturers compare in a plot for CO2 emissions, there are further questions we can answer when we evaluate this large government data set. Now that we have a general quantification of where different manufacturers averaged CO2 emissions for their 2022 models, we can also look into other factors about those vehicles contributing to more significant emissions.

This is where we utilized some data mining techniques to fit a statistical model, specifically a multiple linear regression. Multiple linear regression is a statistical method that models relationships between a dependent variable and two or more independent variables. It is essential throughout the entire modeling process and interpretation to understand what these variables mean:

Tying into the Fuel Economy data, our dependent variable in question would be the combined CO2 emissions by the vehicle. We want to analyze and find out which other features in a vehicle included in the data play a statistical significance in contributing to how much a car pollutes. After meeting certain statistical assumptions for this modeling technique (we will likely cover these techniques in a future analysis blog, so stay tuned!), we can fit a multiple linear regression model to determine those predictors.

After running this model, we concluded a handful of predictors that impact combined CO2 emissions. One significant feature in a vehicle that strongly correlated with CO2 emission was the vehicle engine's method for air aspiration. Across all vehicle brands, cars with the supercharged process for air aspiration seemed to average higher levels of combined CO2 emission. This is likely related to how this engine type requires more fuel for its combustion process, increasing the overall amount of pollutants. Turbocharged engines averaged the lowest combined CO2 emissions among the different air aspiration descriptions.

Another significant predictor for CO2 emissions, to no one's surprise, is the vehicle's class description. This feature directly relates to the size of the vehicle, which should broadly impact mass and, ultimately, how much power is required to move the car. Among all class descriptions, vans averaged the highest combined CO2 emission, while compact cars emitted the least.

Although we could name some significant predictors for CO2 emissions, some factors outside the Fuel Economy data can contribute to substantial emissions. Some non-recordable variables include a person's driving style, overall traffic conditions, vehicle maintenance, etc. A special shoutout to vehicles marked with the ˜electricity' fuel usage, which averaged zero CO2 emissions. #SustainabilityRocks

Random Chart of the Month - Top Instant Noodles by the Ramen Expert

As we all prepare to look back on the passing colder season (or perhaps our not-too-distant college days), there's no better time to introduce our first Random Chart of the Month, featuring instant noodles!

Specifically, we are looking to visualize and analyze a large dataset provided by The Ramen Rater, a seasoned expert who has reviewed instant noodle brands from around the world for over 20 years. His dataset, known as The Big List, contains over 4,000 entries of the instant noodles' variety, brand, style, country of origin, as well as a personal rating on a numerical scale of 0 through 5. This dataset is ongoing and continually updated, so ramen enthusiasts should go check out The Ramen Rater and all his cool content!

The Chart

With the sheer amount of instant noodles entries alongside the categorization in the dataset, a prominent analysis question was: which countries do it better? Taking the variables on the expert rating of the instant noodles as well as their origin, we were able to generate the following geographical heat map:

This geographical chart measures, by country, the average ratings of all instant noodle brands and codes them into a color scale. The darker green signifies the higher end of the 0-5 rating scale, while the light yellow is the lower.

The Analysis

At first glance, we can see a significant increase in the amount of countries coded with darker green when we evaluate the eastern side of the world map. It probably doesn't surprise most of us to see more of that darker green populate around Southeast Asia, as that area marks the origin of instant noodles (Japan, in 1958).

When ranking the average ratings among countries with at least 30 reviews, instant noodles branded from Malaysia came out on top with the highest average rating of 4.21. This also aligns with a YouTube video by The Ramen Rater, where he gave a 2022 ranking of the top 10 instant noodles of all time, with Malaysian-branded instant noodles appearing five times. South Korea and Japan closely follow with 3.88 and 3.86, respectively.

The three most prevalent countries regarding total reviews in this list were Japan, South Korea, and the United States. These three countries combined branded about half of the 4,298 instant noodle entries in this iteration of The Big List. Japan, the inventor of it all, leads all countries with the most entries with 845 instant noodle ratings. The large variability by the sheer number of instant noodles branded in Japan makes their average rating of 3.86 rather impressive.

Knowing the powerhouse that is Southeast Asia when it comes to instant noodles, there's another analysis question we can consider: what are the best packaging types for instant noodles?

This same dataset categorized each instant noodle with a variable relating to its packaging type, for instance, whether they came in a pack, box, bowl, etc. The three most dominant types were instant noodles packaged in packs, cups, and bowls (about 93% of all instant noodles in this list were from these categories). Averaging out the same rating scale by packaging type, instant noodles in packs average a rating of 3.82, followed by bowls at 3.69, then cups at 3.47. We also constructed a 95% confidence interval for the ratings of each packaging type to really compare the three means:

Confidence intervals are a way to show some accuracy in our estimation. In statistics, we often use these intervals to make an educated guess about a larger population or, for relevancy, a larger population of noodles. A practical interpretation of a 95% confidence interval is: if we ask our ramen expert to rate the same number of instant noodles (falling in the Pack category) a hundred times, we expect about 95 of our average ratings to fall between 3.78 and 3.86.

When deciding on the differences in averages with this method, we generally look for any overlaps between each interval. If such overlaps existed, we could conclude that one packaging type wasn't too significantly different from the other based on average ratings by our ramen expert. For our data, we can see averages for the packaging type of packs were slightly higher than bowls and even more so than cups. We also have a smaller margin of error for the ratings of instant noodle packs, considering almost half of the noodles populated this packaging type.

Using our ramen expert's list, we visualized and learned of higher review ratings from Southeast Asian-brand instant noodles. We also supplemented these findings by concluding that the packaging type of packs is generally favorable among the same metric. To our delight, it is clear that the world of instant noodles is varied, and there is something for everyone on the globe to try. But hardcore instant noodle enthusiasts may consider at least trying some of those Malaysian-brand noodle packs or even seek the veteran opinions of The Ramen Rater.