5 Things to know before getting Einstein Analytics

The idea of being able to correctly predict things has been something of fascination to humans for quite some time; whether it’s ambiguous predictions of future events by Nostradamus or an octopus predicting World Cup results, the ability to make such projections has always been a major talking point. Businesses can also benefit from predictions that help them understand the probability of an event in simple terms – albeit they would rightfully want something more robust than a 500-year-old book or the appetite of a 2-year-old octopus.

Salesforce’s Einstein Analytics provides you with the opportunity to make statistical predictions through Einstein Discovery stories. In case this is the first you’ve heard of these products:

  • Einstein Analytics is an application built on top of Salesforce, that acts as a cloud-based analytics platform which allows you to get insights on your business in a very straightforward manner.
  • Einstein Discovery is part of Einstein Analytics; it enables you to discover patterns in your data, help you understand why things happened and what to do next, and make accurate predictions using its AI-powered analytics.

Use cases for this include using past data to predict which sales lead is most likely to lead to eventually convert into a client or forecasting how well your company is likely to do financially over an upcoming period.

Before diving straight into using the product, here are five things worth knowing about Einstein Analytics:

1. You can use a free Developer Edition to test things out first

If you are new to Salesforce and are fearful of potentially messing something up, worry not. You can test things out on a Developer Edition (Tableau CRM-enabled) to get to grips with the product and understand what it can do to help your business.

It is worth noting that the Developer Edition will have some different limitations to the full fat version – for example, Einstein Analytics will allow you to have datasets with up to 20 million rows; in the Developer Edition, you will be limited to only 100,000. However, this should not detract in any way from your learning journey.

Should you decide to take up our instructor-led Einstein Analytics training course through the iMovo Academy, it is ideal to have this access to a Tableau CRM-enabled Developer Edition.

2. This is not a replacement for human resources – it is an enhancement to their performance

You may see mention of Einstein Discovery being like having an additional data scientist within your team, though this does not mean that it should be a replacement for human resources. The tool itself certainly provides a lot of value but will need initial guidance from someone who understands its intricacies and the way the business works.

In a way, this can be explained better through a comparison to the film Ratatouille: Alfredo Linguini (human) can offer value to the kitchen but can only truly be at his best when guided by a talented chef in the form of Remy (rat). Similarly, you will need someone with an understanding of your business and of statistical modelling to really use Einstein Discovery to its maximum potential.

3. Caters for varying levels of ability

Einstein Analytics offers you an entire analytics platform at your feet, and one that can be used by people of varying degrees of ability. The best example of this is found in the creation of dashboards, where queries (used to create visualisations) can be made in a variety of ways. The most straightforward method is simply choosing fields from your dataset to aggregate and group, but for more complex ideas, you may require someone with more knowledge of queries to create your visualisation via Salesforce Analytics Query Language (SAQL). This ensures that users will be able to get better insights into their business while hardly ever leaving their comfort zone.

4. The technology is only as good as the data

It is worth reminding people that while this is an incredibly powerful tool, it cannot transform tap water into Merlot or Chardonnay. The quality of your data needs to be high to ensure that predictions made will be meaningful, and the size of your data must also be of a sufficient quantity.

Should you clear these hurdles and decide to create a story with Einstein Discovery, you will come across a model evaluation section which will inform you of the model’s ability to accurately explain variation and any recommended changes that should be made prior to deploying the model. Additionally, the choice of whether to deploy the model or not will depend on the model’s strength – this is where it becomes useful to have someone with statistical knowledge up their sleeve.

5. Provides you with explanations that suit your role

While working on your story in Einstein Discovery, there are a variety of things that you can get insights about: when and why something happened, what is likely to happen based on certain factors, and things that can be done to improve your chances of a successful outcome.

Prior to model deployment, all of this and more can be seen in a collection of charts with concise text explanations. You also have the option the view the R code behind the model generated if you are that way inclined.

After model deployment, a scorecard will help to explain the leading causes of an outcome, as well as ways that the predicted outcome can possibly be improved upon.


If you are interested in implementing Einstein Analytics to gain a better understanding of your business, then contact us at [email protected] or on +356 22488300 to schedule a call with one of our data experts to help you understand how to make better use of this technology or if you’re interested in getting training on this system, you may book for a consultation or one of our training courses. We cannot offer you rats in hats to control your every move, but we can offer you help from experienced and knowledgeable individuals who will help you empower your business.