Get in the Game: How to Get Started with Predictive Analytics

There is something undeniably aspirational about predictive analytics. It invites decision-makers at all levels to think in terms of grand transformation, to envision themselves unlocking the secrets of the business universe. And that is awesome. I would never discourage anyone from thinking big. 

I do however consider it an inefficient way to start applying predictive analytics to your business problems. So many analytics undertakings bog down in overthinking the big picture.

As hockey great Wayne Gretzky once said, “You miss 100% of the shots you don’t take.” I agree. Predictive analytics is a full-contact sport. While it is brimming with potential to enable massive change, it is not magic. Scoring goals, both in hockey and business, does not just happen because you visualize it. Analytics is a tool, and like a hockey stick or a baseball bat, you're going to have to put it to work.

So instead of spending your time anticipating what you might someday try, here’s an alternative; a 90-day plan that lets you start taking shots right away. In fact, you can take the first step as soon as you’re done reading this! 

The best way to get started with predictive analytics is to think about your business problems, identify a pressing use case, and let it fly. Start by identifying basic principles that matter to your business: the things you think about every day. After all, what greater satisfaction is there than finding an answer to a problem that already has your attention? 

Start by simply jotting down a few business processes or use cases that matter, and choose the right technology to make it happen. It is important to make implementation and iteration as self-service as possible: that’s the key to lasting success. Highly intermediated technology often dies under its own weight. Self-service BI proved this point dramatically, and now it is the time for self-service AI. For example, if you are in customer service, you probably think a lot about attrition, predicted customer satisfaction and optimizing the productivity of calls.

Write down a description of insight that defines what you want to achieve. Think in terms of a simple framework: I want to MAXIMIZE customer satisfaction; I want to MINIMIZE attrition. Easy, right? Equipped with this, we can look more closely. 

Einstein Discovery come standard in Einstein Analytics Plus : Complete Augmented Analytics Platform

Now let's consider how predictive analytics connects the dots leading to business outcomes:

  • Insights: Operationalize predictive & prescriptive analytics at scale surfacing actionable intelligence at every point of decisioning. Example: Predicting risk of customer attrition in the business process for relationship management.
  • Process Automation: Automate workflows using the results of a predictive AI model. This could include anything from suggesting actions to a service agent to fully automated execution. Example: Pre-approving RMAs if the customer satisfaction score is low.
  • Enhancing Business Applications: Creating a digital experience that shows what happened (Descriptive analytics) and predicts the results of a range of next steps (Predictive Analytics).  Example: Providing a salesperson with natural language recommendations for how to reduce time to close.

Can you match your intent with one of these AI-powered predictive analytics capabilities? Getting this right is crucial. When we hear about businesses struggling with AI-powered analytics, it is often because they haven’t taken the time to figure out what they’re after. Formulate clear questions related to specific business challenges.

It’s only at this point, with a clear idea of the challenge and the shape of the answer, that I recommend thinking about technology. Analytics is most effective when integrated with the business process you want to improve. In an intelligent predictive AI experience, the right data automatically flows to users who need it in a context that helps them know what to do with it.

With a Augmented Analytics Platform like Einstein Analytics from Salesforce, everyone, not only data scientists, can take a Gretzky-style slapshot. The ability to automatically create models based on data from the service cloud, means that those close to the action (lets call them Business Scientists) can create predictive models and put them to work. That’s how you get from your head to a goal in just 90 days.

The net is right in front of you. Take your best shot!

Sarath C B.

Director | Enterprise Applications | GTM Applications Leader

5 年

Thanks for the write-up, Ketan, well done! One of the questions many ask is - what type of data do Salesforce or any AI solution provider have access to for training the data models? And the answer is simple - You, "As a customer" is in full control to create the stories aka use cases without any sensitive data. Ketan has perfectly stated it, pick a business problem that you are planning to solve and let it fly. AI is the future and let's maximize the potential by putting it to work.

Michael Cademartori

VP of Sales, WW @Anaconda, Inc. | Experienced Sales Executive | Enterprise SaaS | CRM, AI/ML, Analytics, B2B Sales, High Growth

5 年

Well done Ketan!

要查看或添加评论,请登录

Ketan Karkhanis的更多文章

社区洞察

其他会员也浏览了