As part of my IBM training this week, we learned data mining as a tool to manage data and make decisions that build your ROI.
How predictive analytics builds ROI
Helps connect data to effective action by drawing reliable conclusions about current conditions and future events.
Predictive analytics is a more proactive decision-making tool that drives forms of competitive advantage by analyzing patterns found in historical and current data as well as attitudinal and social media data.
Scenario 1. Let’s say you want to retain your current customers – 30. You decide to spend 30 $250/ customer to retain them ($7500). Despite your efforts 5 customers churn (leave the company). You now spend $1500 to bring your total customers back up to 30. Outcome: Spent $9000.
Scenario 2. Now, let’s take a look at this with predictive analytics to help predict WHICH customers are most likely to leave. I mean face it – some customers won’t leave for a variety of reasons. Maybe their firm pays for the service. Maybe they’re locked in by high switching costs …. Why send these folks offers worth $250, when they’re going to stay anyway.
Recommended for YouWebcast: A Week in the Life of an Agile Creative Team
Scenario 3. Now, you segment customers based on their likelihood of churning into high, medium, and low risk. You now spend $250 only for those 10 customers with a high or moderate risk of churning – for a total of only $2500. Even though you still lose 5 customers, you’ve saved $5000 just by not wasting resources to send an offer to customers who would never leave. You still incur the $1500 cost to replace your lost customers, so the total cost is
You can afford to sweeten the pot a little. Let’s say instead of spending $250 you decided to send a really killer offer costing $400 to the 5 customers with a high risk of leaving and a $300 offer to those with a moderate list of churning. You’ve increased your expense to $3500, but now only 2 churn (instead of the original 5). It now only costs $600 to recover from this loss of customers. Now you’ve saved $4,600 over not having the predictive information.
OK, you say. So, let’s choose option 2, since we’re financially better off with that option than either of the other 2. Well, maybe. Don’t forget that running a business is more than a simple financial exchange. Those extra 3 people you lost in scenario 2 talk. And, with the vast networks available through social media, those negative comments can have a HUGE effect on what other consumers decide to do. This can accelerate the loss of customers, increase the cost of acquiring new customers, or both. YIKES!
Benefits of predictive analytics
- Produces the highest customer satisfaction
- Projects delivered on time and under budget
- Returns the highest ROI
- 94% of firms using predictive analytics see a positive ROI in 10.7 months
Three basic pillars underscore predictive analytics:
Capture – detect and capture transactions, demographics, interactions, opinions, etc
Predict – develop predictive models, data mining, text analytics, social network analysis, and statistical analysis. Predicting is part art and part science so NO software tool can provide answers by itself. You need skilled analyst to make sense of your data.
Act – use predictions to create rules, optimization, and processes.
Some questions answered by predictive analytics are:
- Who are your best customers?
- Can we get more like them?
- What/ why do they buy?
- Why do they leave?
Using predictive analytics in a social media context, we can also assess:
- What posts (at what times, in which social channels) provide the most visits to our website? The most conversions?
- Does spending time addressing customer service on Facebook cost less relative to having customers call our customer help desk?
- Which social networks most effectively help us reach our target market relative to consumers NOT part of our target market?
Of course, this just scratches the surface of opportunities opened through the use of predictive analytics in managing your social media marketing campaigns.
IBM Modeler helps manage the entire process of predictive analytics using drag and drop tools that alleviate any coding. Really cool stuff if you can afford the $40K price tag. Of hire a consultant who has both access to the software and the skills necessary to make your data sing — all the way to the bank.
Wanna learn more. I’m attending IBM analytic training all week and I’ll blog about each successive day. Tomorrow is on data mining, which I think you’ll find applies more generally to social media marketing. So, be sure to bookmark this site so you can come back tomorrow. Or, better, sign up for my email newsletter to learn more about social media analytics and marketing strategy. If I’ve whetted your appetite to use more analytics to optimize your marketing, let me show you how Hausman and Associates can help with our unique virtual agency model that provides cutting-edge social media at a reasonable cost.