Many enterprises have started to understand the importance of following social media conversations.
Those that are ahead of the curve are already using social media analytics and/or listening tools in one form or the other. Most are using very cheap or even free tools. Some may be using extremely expensive solution that comes as an option with an overall social media publishing platform.
If you already have some kind of a social media analytics tool in place (especially those that get listed in Google when you search for “social media analytics”), here are 7 reasons why you should re-evaluate your existing tool.
1. Is your social media analytics tool unnecessarily expensive or too cheap?
Social media analytics and listening tools offer various features with different pricing levels and there are even free tools available in the market. When social media analytics is not at the top of the agenda for you, you tend to adopt cheap or free tools to save costs. It is understandable, but you’ll get what you pay for. You’ll encounter limitations and unreliable results with cheap or free tools.
Free tools can be an evaluation version to eventually encourage you to purchase an upgraded plan (typically still very cheap). These tools usually offer you with only quantitative measurements such as number of likes, comments, shares or retweets. Free or cheap tools might also be an “add-on” on top of your expensive CRM (i.e. Microsoft Dynamics) or other comprehensive social media management and PR platforms (i.e. Meltwater). Since the main expertise of the tool provider is not in social media data analysis, rest assured, the analysis results are not very accurate.
On the other hand, you might be using an unnecessarily expensive social media analytics or social listening tool. Bigger and famous companies in the field might dare to charge you more because of their brand recognition and arrogance that the customer would simply pay, believing that big brands would have the best tool. Or … they might actually really have great analytics tool but the implementation method might be causing the running cost to go up. With the same amount of investment, you might be able to get a much better service (not just a stand-alone tool), from a company that is solely focused on doing social media data analytics.
Budget is always an issue in marketing. But if you are using unnecessarily expensive or cheap/free tool, it is time for you to re-evaluate what exactly you are getting; the following questions will help you with that.
2. Can your tool interpret unstructured data?
Have you ever come across the term “unstructured data"?
Unstructured data is the social media data that even the smartest computer technology cannot classify.
Texts, photos, icons, and videos are used in social media posts. Marketing professionals usually consider quantitative metrics (eg. number of likes, shares, comments) and easy-to-collect data (eg. occupations, ages, locations, etc.), which account for less than 10% of the entire social media data, as an important KPI of a marketing campaign. But, these figures alone will not provide you with the brand insight that is necessary to formulate your social media marketing strategy. It is of utmost importance to understand the true qualitative sentiment of your customers behind their posts and comments, which include images, videos, and ironic/sarcastic jokes.
Some social analytics tools have started to analyze photos/images by implementing “logo detection algorithm” and counting the amount of impressions. “Logo detection” is done by taking a sample image of a product that has a logo and the tool will identify similar logos in other images. So these tools can find images that are associated with a certain brand... but it still can’t understand the bigger context where images and texts are combined to have varying meanings.
Unlike “Logo detection" using a sample to look for the other similar logos in different images, there are no absolute samples for context analysis from images and videos since contexts can be diverse. To understand the context of each post, computer algorithms need to be extremely flexible - similar to a human mind. Unless your social media analytics tool uses human analysts to interpret unstructured data, results of the analysis you get just won’t be good enough to formulate a successful marketing strategy. So, comes the next question …
3. Does your tool use human analysts?
Like said earlier, If you rely only on computer technology alone, it is impossible to interpret unstructured data. Computer technology might be excellent at counting figures and providing you with quantitative results but it just can’t interpret human senses behind social posts, which is a requirement to get accurate sentiment analysis. Sure, there are ‘deep learning’, ‘machine learning’, and ‘artificial intelligence’... social analytics tools can improve algorithms they use to get better at understanding customer sentiment. We can teach computers to interpret different human emotional states such as happiness, anger, disgust, surprise, excitement, etc, by adding default meanings to common keywords, hashtags, idiomatic phrases, slangs or acronyms. But it just won’t be accurate and flexible enough to work like a true human brain.
Some providers have applied Natural Language Processing (the combination of artificial intelligence and computational linguistics) to their tools to decipher the social message similar to a human based on standardized linguistic rules. NLP is especially good at interpreting slangs and sarcasm. But, language usage changes rapidly, updating of NLP algorithms takes time, and NLP still has difficulty interpreting the bigger picture when sarcastic text is combined with images and videos. Combining human intelligence with any type of technology you use will provide you with the best and accurate results.
Here is an example:
A post shared by Susanne Rosenberg (@susanne_rosenberg88) on
A typical analytics tool will interpret this post as positive (“new”, “great job”) sentiment towards this brand that made the salad. However, if a human takes a look at the image that is posted, user @Susanne_Rosenberg88 is being extremely sarcastic, because … well, who wants to have hair in their salad! Accurate analysis of this sentiment = negative.
Actually, sentiment analysis provided by most analytics tools is totally fu**ed up. Read my other blog on this topic.
4. Does your tool exclude internal posts from sentiment analysis?
Typical social analytics tools pull data from both internal posts (posts by your company and its employees) and external posts (posts by other parties or your social audience in general).
Internal posts are usually positive, because those posts are promoting your brand. If you include internal posts into your sentiment analysis, you will get biased results. You need to make sure that all internal posts are excluded.
Also, if your influencers mention your brand in social media, typical social analytics tools analyze all of them automatically. Since you pay influencers to promote your brand, those posts usually have positive sentiment. In order to get true insights of your target audience’s sentiment, it is better that you exclude influencers’ posts from your analysis. Otherwise you would get an overly positive sentiment result, fooling you to believe that your marketing campaign was a success.
High-quality social media analytics service can conduct influencer analysis separately and sentiment analysis is best to be done on data generated by your customers and target audience.
5. Can your tool filter out irrelevant posts?
Keywords/hashtags tracking is a common approach applied in social analytics tools. Keywords related to a particular brand are predetermined before analyzing social data and the social analytics tool will start scanning and collecting them from a specific amount of social mentions. This method is quick, affordable and is an ideal choice if you want to know how many people mention your hashtags in a marketing campaign.
For example, NIKE is a global shoe brand with “Just do it" as their primary slogan. Rather using generic hashtags such as #shoes, #trainers or #sneakers, NIKE has run #justdoit campaign in different social media platforms because the hashtag is unique and easy to track.
Relevant #justdoit post:
However, some people use #justdoit to discuss other issues that has nothing to do with NIKE. Following posts are good examples of that. Typical social analytics tools using keywords/hashtags tracking will not be able to filter these irrelevant posts out. Only a high-quality social analytics service, that combines advanced computer technology with human intelligence, will be able to interpret them as irrelevant data or ‘spam’.
Is your social analytics tool good enough to distinguish whether a social post with hashtags you are tracking is relevant or not?
Also, most free tools are not able to show you the original postings from which they are basing their sentiment scores. So you won’t even know if your sentiment analysis is correct or not, and which posts are positive or negative.
High-quality analytics service allows complicated queries to be implemented to filter out irrelevant data and also allows you to see the original posts.
6. Does your tool provide true multi-language support?
If you are a global brand operating in different countries, you need to follow social media conversations in different social platforms and in varying languages.
Does your social analytics tool process data in French, German, Swedish, Chinese, or something else? How many languages can your tool interpret? And do you think that just because your analytics tool has so many languages they support, that would guarantee its quality?
Free social analytics tools usually do not offer any multi-language support, but some cheaper tools might claim that they can also analyze languages other than English. But usually, language interpretation is done with digital dictionaries within the tool. And these dictionaries end up translating the data word-by-word, just like Google Translate. If you have ever tried translating something in Google Translate, you should know how funny the outcome can be.
Listening to what your foreign audience is discussing in their language is a huge challenge, especially when you want to know how they feel about your brand. Similar to English, each language also has its grammar system, millions of words, phrases, slangs, acronyms, and the people using it have their own local culture, customs and habits, which will take you a long time to teach the computers to absorb to get rid of word-by-word translation and to be able to distinguish spelling mistakes, grammatical syntax errors, and sarcasm.
Also, the same issues I’ve discussed earlier in this blog apply to other languages - can computer technology interpret unstructured data, filter out irrelevant posts, and understand true customers’ sentiment in other languages?
There are also languages that are rare, i.e. Finnish, variations of Chinese, Catalan … can you trust your tool to analyze all languages as accurately as possible? And, what if there is more than one language used in a post? Again, combining human intelligence with computer technology is the key. Who, other than a local human analyst, would be able to understand social media data better?
7. Does your tool include any ‘service’ and ‘customization’?
Most social analytics tools are just that. Tools. Especially the cheaper ones. You get access to a dashboard. You have to define the topics you want to track. You need to basically set everything up yourself.
There may be small pop-up instruction and tip boxes when you hover the mouse over each tab of the dashboard. But you have a million other things to do than to figure out how to work with the dashboard.
Cheap tool providers usually don’t provide you with any service. On the other hand, high-quality service providers can set up the dashboard for you, and train you on how to use the dashboard. Trust me, this saves you a lot of time. Better yet, they can also provide you with customization and consultation service.
How can customization service help you?
- Optimized dashboard: By determining what you need to analyze in social media, such as engagement rate, quantitative metrics (the number of likes, comment, tweets, shares), sentiment analysis, etc. and which platforms (Facebook, Instagram, Twitter, etc.) you want to track, the best-of-breed service will set up a web-based dashboard containing just the information you want. Optimized dashboard will eliminate unnecessary features so it is easy for you to use it.
- Customized reports: Some dashboards may have a reporting feature allowing you to manually create your own reports by dragging and dropping the results from the other tabs. This can be a daunting task for marketing professionals who are busy with everything else. There are service providers who will prepare customized reports just for you. These reports are in a shape which you can use when presenting social media campaign results to your management team.
How can consulting service help you?
- Provide a broader view over your company's social media’s current state, which you might not have recognized yet.
- Provide alerts and suggestions when faced with communication crisis in social media.
- Provide you with highlights, pinpointing what is important and what is not, from the vast data.
- Provide you with suggestive path to follow, especially when your company is planning to penetrate foreign markets.
Does your social analytics tool cover these points? If yes, congratulations! You are working with the best-of-breed, high-quality social media analytics tool.
But if some of the above questions made you stop and think, evaluating some other alternatives is highly recommended. A high-quality and customized social media analytics service (instead of a mere ‘tool’) will provide you with the brand insight that can help in formulating successful social media and marketing campaigns.