quintly Blog  /   Research & Trends
4 MINUTE READING TIME
Research & Trends

Seasonal and Trend Analysis in Social Media Analytics

by Alex Goncalves on January 31, 2018

Seasons are a big part of marketing and business in general. Certain periods of the year can reflect positively or negatively on the results of business. The reasons behind that will vary from one industry to another, and can also change depending on the territory, product category, or even if a sudden new technological advancement is out. A seasonal analysis, or an analysis of a longer period to detect any potential trends in the market is therefore a big part of analytics. When it comes to social media analytics specifically, this can be exciting, since we have access to a good amount of competitive information at our disposal.

Being practical about it, let’s then jump right into it and look into a few interesting ways of exploring such analysis in social media using an analytics platform such as quintly.

1. Choose the best period:

Every social media analysis starts with the choice of a good period. When going for a seasonal or trend analysis there are a few points that can help us be sure we are looking into the most relevant data.

1.1 - Base the selection on something that is relevant for your strategy specifically: Many times we choose periods that are common to everyone else, such as the “first quarter of a year” for example, or even just “monthly.” While this can be interesting for a general analysis, being more specific to what is going on with our business can prove to be a lot more relevant and strategic. Campaigns, product launches and crisis will rarely follow a common pattern with the rest of the world. Even if our financial reports must follow certain patterns, we don’t need to follow such regulations when it comes to marketing and business analytics.

1.2 - If a trend is spotted, extend the period until you can see all of it: If a certain graph is showing us a clear trend that is interrupted either in the beginning or end of the chart, extend your period so that you can see all of it. When does such trend begin? When does it end? Many times the turning points of a trend will reveal excellent insights. The following graph is an example of such incomplete view of a trend.

Screen Shot 2018-01-31 at 15.00.59.png

1.3 - Choose the best graph for the job: Don’t settle for the default graph that the platform suggests. Especially in a platform like quintly, in which you can choose any graph type you want and customise metrics as you see fit, substitute the graphs until you find the ideal one. A line chart can be a good option to detect turning points on trends, or the intensity of growth/fall curves, but if we move into comparing periods it is possible that a column chart can be a better fit in some cases. Be open minded about it, what matters is that you can quickly and easily spot the key points from what the data is trying to show.

The following two graphs are an example of a good moment to change from the line chart (default) to a column chart for a potentially more interesting or even just complementary visualization.

Screen Shot 2018-01-31 at 15.00.44.png

 

Screen Shot 2018-01-31 at 15.00.32.png

2. Create and detect specific triggers for specific campaigns:

To make things even easier when we know that we are starting a new campaign, we can create specific triggers to be used for analysis down the line. If we are searching for competitive insights, we can detect such triggers created by our competition and track them more closely.

Triggers can be any hashtag, keyword or expression that will repeat itself throughout the campaign. We are able to detect such triggers and use them as a key to the choice of period or the selection of content pieces for performance analysis.

To learn more about filtering and working with such triggers, this article explores the topic a little further. Facebook Analytics

3. Separate the analysis by content type and different types of reactions:

Once we have an ideal period and we filtered out the best data, we can further dissect the results. Separating content by different types is a good start. Then we can look into the types of reactions. A few points are relevant when jumping into this kind of dissection:

3.1 - Consistency? Do we have any kind of consistency in performance by content type? Or does it vary by the type of product, different social network, or even post frequency or time of day? When we understand the consistency we can better understand if the trend might be following a certain type of content specifically, or if the trend is broader than that.

3.2 - Reactions by content type: Pushing further into the content type, we can separate the different reactions on different content types. This is an extension of the study for consistency, but it can reveal very interesting insights so it is worth mentioning separately. A key point to remember here is that “comments,” when it comes to reactions, must be looked into with care. People might be expressing very negative feelings about what they see. An estimate of the sentiment of comments can be a great support for this analysis.

3.3 - Reaction types independently: Beyond content type, how do reactions perform in general? Different types of reaction can point out to trends of their own. Especially when there is a new feature released by a social network which can give users the ability to react in different ways, the audience might create a trend simply by making good use of what is new.

This post shows a study made specifically around the use of emojis.

In Conclusion:

Analytics needs a little bit of work from our analyst mind to deliver the results we need. We can reach such results by starting with an objective such as we did here: “detecting potential trends.” From there we can shape the platform to show us what is behind the data and reveal the insights we need to shape our success going forward.

There are usually more than a few ways to perform certain analyses as well, so the examples we bring here are just one set of options to be explored. You can go further with your studies by exploring more of the analytics platform and its features, and also by sharing your objectives with the team behind the analytics provider. Contacting an experienced analyst or team member in your provider can save you a huge amount of time when unsure of the best way to perform certain analysis.

If you haven’t yet, start a free trial with quintly right now.

Related Categories

Research & Trends

Alex Goncalves

Alex was part of the client & partner relations team at quintly, and author of the book Social Media Analytics Strategy, Apress 2018. By Alex’s calculations he has spoken to more than 9,760 people about social media, for more than 24,400 total hours of conversation in more than ten years in the field. Being passionate about the topic, Alex enjoys the idea that the world is better when united by the Social revolution.

Join the conversation. Leave us a comment below!

Get in touch today to start analyzing and optimize your social media performance.


Request a Demo