1.Metrics: Metrics Bring to New Light to Events
Prepared by: Amar P. Natasuwarna
A metric usually brings a different light to an event.
Metrics is creating a different angle of view of the facts happening in the world.
It is easier to reach new ideas, opinions, or insights on the events or facts.
The metrics see the character as a very methodic person as if such personality would indicate the cause to other aspects of his life.
Metrics change the way we see even the most common of things.
The use of metrics in that scene is a work of genius storytelling.
4.A Deck of Metrics
Ideal metrics can put together into sets, collections, or decks to make it even easier to work with.
Some tools give many dashboards to work with and use them as our decks.
A deck can be a good starting point for our work on any project.
Create themed decks, ready for high-level performance analysis, a few KPIs, or detailed content investigation with content filtering and different views on interactions.
A deck can become a live dashboard in one project, or a linear storytelling report in another.
The idea is that such decks can save us a lot of time, giving us a head start into new projects or goals.
Building and collecting metrics can become as exciting as collecting any of the common collectibles in a social media analytics kind of context.
Conversations on this topic are usually very interesting and become more, so when we meet colleagues that are into customization of metrics and building new ways of approaching the use of certain data.
This level of connection to metrics is easier to understand once we have experienced working with custom metrics and manipulating data sources toward the precise goals of our projects
6.Default and Custom Metrics
The metrics are available to us before we decide to create our own, called default.
We find many similarities across different tools.
Usually, an analyst quickly becomes familiar with the default metrics across the social media analytics landscape. The social media analytics market is not, by any means, fully standardized, but professionals everywhere try to make that happen as much as possible.
Some tools give us the power to create our own metrics.
Some custom metrics are very simple, in other tools it can be more complex, with the need for a special professional services team to be involved and extra fees added for the service.
Private-level data is only available to users that have administrator access (admin access) to the social media channels.
With the manager access a user can pull the private-level information into metrics.
Private-level metrics are defined by the networks as sensitive or carrying proprietary information.
It is usually not visible from public access in any way.
Examples of private-level metrics are reach, impressions, information on investment to promote content, and most demographics.
Available to everyone, as the name suggests.
It is visible from public access into any social media network.
If we are scrolling through our timeline on Facebook, and the post of a friend comes up, we usually can see how many interactions it has, the types of interactions, and the interactions themselves. This is public information, or public-level data.
Eventually, certain networks will only deliver data on certain types of profiles, such as businesses and not individuals.
Issues with privacy of individuals can also come into play.
9.Metric Categories: Divide and Conquer
How to turn something complex into something simple to understand and remember? Divide it into parts and group those parts under categories.
The term analysis comes from the process of dividing something into its parts to understand the details.
While separating the universe of metrics into categories, we find that most metrics fit well into a given category, which is extremely helpful for us to quickly find and remember them all.
A few metrics, however, fall off from any category.
These are usually metrics that involve formulas that mix many metrics together such as machine learning
12.Graph Types: Data Is There, But Does It Look Good?
13.The Effectiveness of a Chart Type
Period of analysis
The number of channels or sources in display
The number of metrics displayed in multidimensional graphs
The size of the display surface (some metrics look good on huge screens but not on a standard screen or on a mobile device)
14.Example 1: A Simple Change in Chart Type (1)
15.Example 1: A Simple Change in Chart Type (2)
16.Example 2: Adding an Extra Dimension to a New Chart Type (1)
17.Example 2: Adding an Extra Dimension to a New Chart Type (2)
18.Default Metrics Capabilities
A tool tries to cover as much as it can under the default section.
If it is not a very comprehensive analytics tool, it tries to cover enough to answer for the concept behind the tool.
They have enough default metrics to help the user see how the posts are performing.
The main objective is to deliver relevant metrics, so usually they offer many metrics, and the best tools offer the greatest amount of metrics.
Having more metrics is a safer bet when choosing a tool If the investment is in the same price range.
19.Flexibility is important in the ever-changing world of digital marketing
It is very rare that default metrics will cover the needs of a project in a perfect manner.
As we become more experienced, and we learn of what data is available, we begin to have new ideas into metrics we want to use and even experiment with.
There is a trend in customization all around the digital world.
Specific social media analytics could continue to use this knowledge in projects involving other digital data.
It is also fun to have the ability to test theories that we have or that we come across.
Custom Metrics Capabilities
20.An Interesting and Simple Example
Once upon a time, a trend of blog posts throughout the Internet indicated that Instagram performance needed at least ten hashtags per post for it to work. Could that be true?
Taking into account different brands, profiles, audiences, and all the different dynamics on different content, my doubt only grew even further. I then set out to prove this theory with the use of data.
So I created a custom metric to look at that. My goal was not to create another rule, on the contrary, it was just to be free of such rule and open my mind to different factors when it comes to the use of hashtags.
The metric was average interactions per post by average hashtags per post
Example 3: An Interesting Chart (1)
Example 3: An Interesting Chart (2)
Example 3: An Interesting Chart (3)
Example 3: An Interesting Chart (4)
Example 3: An Interesting Chart (5)
Custom Metrics In Paid vs. Organic Analyses
The Use of Machine Learning
It is important to highlight that we are looking at a competitive benchmarking approach in Figure 11-14, an external analysis.
This means we are comparing the results of pages that do not belong to us, that we do not manage.
These results are obtained by a machine learning process, which analyzes each post to see if it shows signs of being promoted or not.
If detected as a promotion, it is added to the list of paid posts and all the interactions from it are added there as well. So what we are looking at in Figure 11-14 is an estimate of what is potentially paid.
This is still very useful, because machine learning processes can reach high accuracy rates. The result is still an estimate, so we can ponder our analyses and data-based decisions.
Interactions vs. Paid Posts Graphs
Paid Interactions vs. Paid Posts Graphs
Paid and Organic Post vs Interactions
Special thanks to all the people who made and released these awesome resources for free:
Presentation template by SlidesCarnival
Photographs by Unsplash
Illustrations by Undraw.co