skip to Main Content
Social Analytics

Social Analytics Changes the Game for Marketers with Machine Learning

Savvy marketers are more metrics driven than ever. The culture shift to online communications and commerce allows marketers to make data-driven decisions; a far cry from considering newspaper circulation to estimate the impact of an advertising investment or crossing your fingers on a pricey direct mail campaign.

Today, marketers are focused on online activities where audiences can be curated and where ROI can be measured through clicks, views, and form-fills.  Finding your audience with proper keyword and hashtag selection is a mix of art and science, and monitoring your competition is a daily activity.  Layer on aspects of machine learning and artificial intelligence though, and you get to a whole new level of accuracy, allowing marketers the opportunity to demonstrate an even higher revenue impact for their organization.   Here are three examples of how machine learning can improve organic social promotion, online advertising, and competitive intelligence.

Organic Social Promotion

If you’re a marketer working to promote a diabetes product related to testing one’s blood sugar, you’ll want to know what hashtags would be the best to organically promote your company’s tweets.    To determine this, you’ll analyze the reach and importance of the hashtags.  The hashtag #diabetes has a broad reach (about 213 million tweets in the past two months).   Now consider how to further focus the audience to those interested in blood sugar within the diabetes-related sphere.  You’d want to find tweets that are semantically similar but add enough impact to warrant the use of the additional characters.   Should you add #glucose or #hypoglycemia?  After determining that the semantic similarity qualifies it as appropriate to use, consider the impressions and net impressions to determine how important each hashtag is.

  • #glucose is 81% semantically similar, so it’s certainly interesting to your audience, but it only adds about 48,000 impressions so that’s a very small change over 213 million for #diabetes alone.
  • #hypoglycemia is 78% semantically similar, also very interesting to your audience, but it would add about 1.3 million impressions…. Far more than #glucose. Of the two, this is the clear winner.

Twitter Advertising

If you are promoting a drug therapy for depression, you’ll want to carefully consider the keywords you buy from Twitter.  “Depression” is the obvious one but what about “Anxiety” or “Mental Health”?

  • “Depression” will get you about 200 million impressions over a two-month period. That’s a significant number.
  • Adding “Anxiety” increases that by only around 37 million, or 18.5%.
  • “Mental health,” on the other hand, is hugely important, contributing an even larger audience than “Depression” with over 555 million impressions in 60 days. Clearly adding Mental Health to a campaign will broaden your company’s audience and would be the appropriate choice in this use case.

Competitive Research

If you’re working for a company that developed a new voice activated product that competes with Microsoft in the IoT space, you’ll want to know what Microsoft is talking about in social media.  What keywords and hashtags are they using and what sentiments do people feel about the company?

  • Microsoft’s top keyword and top hashtag are the same: AI / #AI, which is included in about 27% of their tweets.
  • #iot is one of their top 5 hashtags and #language is in their top 20.
  • Microsoft’s tweets make up about 5% of all tweets about #iot.
  • Sentiments around Microsoft are positive, including “achieve,” “empower,” “support,” “helping,” “better,” and “easier” in their top 10.

The Status Quo is Not Good Enough

Enterprise companies have access to Twitter representatives who can provide insights not available to the public, albeit from a biased sales rep who must be reached by phone or email, a time-consuming process.  Worse than that, agencies and smaller companies are making do with limited information.  Social media monitoring companies are not able to call out incorrect co-occurrence like #Pisces from #Cancer for a pharmaceutical company.  Northern Light has created a machine learning based platform, Social Analytics, to get marketers the real-time information they need to make intelligent choices for organic post promotion, advertising, and competitive research.    Contact us for a demo today.

Back To Top