Technology & Data
- Big Data
- Tech & Innovation
How to Get Your Sales and Marketing Teams to Work in HarmonyContent Marketing for Midsized Companies: Whom to Target, What to CreateAtri Chatterjee of Act-On Software on the New Generation of MarketersMarketing Automation: What It Is and Why You Need to Know
- Social Tools
Join us September 15th in Atlanta for The Employee Advocacy Summit and learn how to unleash the power of your employees.
Post your event here and we'll share it with our community. If one of our members is featured, we'll promote as well on their profile.
- Marketplace & Webinars
The SMT Marketplace
Your resource for exclusive content and insights from Social Media Today, and opportunities to reach our community of professionals.
The Social Business Book Club brings you books, discussions, and insights from today's to business thought leaders.
Join interactive talks and and panel discussions with leading thinkers and practitioners on social media and networked business, or browse the catalogue of recorded sessions - all completely free.
Reach Social Media Today's community of marketing and communications professionals in an editor-approved context with a native advertising package.
Prescriptive Analytics in Social Media: Optimize Your Processes
Posted on May 21st 2013
I’m attending a week of analytics training this week generously provided by IBM.
First, let’s take a look at why metrics matter in running your social media marketing:
- spend – a huge amount of consumer spending is motivated by social media. And mobile shoppers spend, on average, 21% more than those on other computer devices.
- organizations using analytics to monitor their business achieve 2.2 times better ROI than those who don’t.
- using advanced analytics, including predictive and prescriptive analytics takes skills businesses need. 60% of businesses need MORE skills in business analytics.
You’ve heard me talk about the difference between descriptive, predictive, and prescriptive data on this site before.
Descriptive analytics, such as #visitors, #fans, and even virality estimates are useful, but are limited in their ability to help you make decisions. Predictive analytics go a little farther and predict outcomes. For instance, you can predict your # of sales based on visits to your site if you have prior data to build the model. Prescriptive data goes even farther to help you optimize your returns by finding the BEST solution to a business problem.
Today, we learned about prescriptive analytics used to optimize your business processes — and your profits.
Many businesses face problems that require them to optimize their processes given constraints faced. Prescriptive analytics help you allocate scares resources, optimize market outcomes or limit risks. Some examples include:
- With a given demand, how much of each product should you manufacture to optimize your profit, given each incurs set costs and returns a set profit?
- Should you make or buy products and how much of each will optimize your profits?
- If you have members who reflect certain patterns of buying, how can you optimize your offering to each segment.
The outcome of optimizing your processes is it determines the best possible plans, provides alternatives and tradeoffs for various solutions, and can respond to changes in your strategy and/or environment to determine the effect of these changes on your returns.
In social media, you can use these optimization tools to determine how much to spend on advertising (or incur other costs) on each platform, for instance.
The specific optimization tool we learned was CPLEX from IBM. But, there are other solutions for optimizing decision-making exist.
What optimization does
Optimization uses data, objectives you want to optimize and constraints on resources to determine how the resources should be allocated. Let’s take a simple problem.
Let’s say you set a goal to respond to all comments posted on your social media platforms within 6 hours.
You’ll need data reflecting the average number of comments achieved in each social media platform. You can even arrange this by hours of the day and days of the week. Estimate the amount of time necessary to respond to each comment. Now, define your resources — the number of people you have available and their hours. You can even define which employees are able to complete which tasks — ie. maybe some of only able to respond to comments on brand1 while others can respond to comments on brands1-3 (of course, you’ll also need data on the relative number of comments related to each brand). Or, maybe you have employees who only address customer service issues and other who deal with other types of comments.
Using prescriptive analytics, you can now determine how to deploy your staff to meet your goal of responding to comments within 6 hours. If you don’t have enough or have too many staff dedicated to this task.
You can do the same thing with content creation — you can optimize the number and/or type of posts that optimize your market returns based on issues such as the costs associated with creating content and returns from creating content such as increased sales, improved SEO, etc.
What do you think?
Of course, this is a VERY different way of approaching social media analytics. It may not be everyone’s cup of tea, but I hope you see how prescriptive analytics helps run your business and optimize your outcomes. If nothing else, I hope you’ve gained an appreciation of metrics beyond simple descriptive analytics in making the most of your social media marketing.