Predicting Future Sales

This article outlines how Emcien can be used to analyze CRM data, like the data that is maintained in salesforce.com, to understand past successes and failures, identify the factors that lead to those outcomes, and apply those results to maximize future sales.

 

Align your data with a use case

This walkthrough uses the bank_sales.csv data set. To apply this same analysis to your own sales data you will need sufficient CRM data.

sales-data.zip

 

This demonstration uses a curated and open data set. If your sales data is not as complete or clean as the example set, you may still be able to implement sales analysis. Contact Emcien for help with your specific data challenges or complete automation solutions.

Optimize your data.

In order for Emcien to identify patterns of successful and unsuccessful sales, it is important that the data being analyzed includes dependent variables, or the outcomes you want to predict. For most CRM systems the desired outcome is a concrete outcome of a successful sale. Other CRM events might be more complex, like identifying the factors that lead to a shorter sales cycle.

 

In predicting successful sales, it can be helpful to determine what additional data can improve the analysis. Adding data in Emcien is as simple as joining additional data, such as weather data, as the process of analysis is very forgiving.  

 

To further enhance the analysis, or to predict more complex outcomes like shorter sales cycles, it is necessary to derive data from existing sources. For example, if a timestamp is an existing element in the data, it will add depth to the discovered patterns to derive features like seasonality, time of day (e.g. early morning or night shift), geographic region, to the analysis. With multiple timed events you can also derive a sales cycle time. Reference the Derived Data page and the Data Prep page for more guidance on data preparation, or you can contact the Emcien Services team for assistance.

 

Upload your data for analysis.

You can upload data files to Emcien using SFTP and your preferred FTP client, such as FileZilla or Cyberduck.

To connect to Emcien using your FTP client, use the following credentials:

  • Host: feeds.emcien.com
  • Username: {Your Emcien Feeds server Username}
  • Password: {Your Emcien Feeds server Password}

Begin your analysis.

On the Home page click Analyze Data.

  1. Using your preferred Internet Browser, navigate to the Emcien Sign In page:

  2. On the Home page click Sales Data.

 

The Analyze Data button will bring up all of your uploaded data. Select your data set.

Select a Project Folder, and if necessary rename the Report. For this data set we will be predicting the Category “Sale”.

Once your project is organized, click Start Analysis.

The load screen will take you through each stage of the analysis.

When the analysis is complete you will hear a chime and the View Analysis button will be highlighted in green. Just click the button to see your results.

See your results.

By clicking the “View Analysis” button you are accessing the precomputed results of your analysis. From the Emcien Dashboard Page you can explore the connections in your analyzed data, but the current objective is to determine the factors that signal the outcome of a positive sale.

In the right hand column simply click on the blue button marked “Sale” to see the Category Details Page.

 

Within the Sale Category are two items, Sale:Yes and Sale:No. Click on the light blue “yes” button to see what items connect to a successful sales outcome. By default these items are ordered by the strength of their connection with the Category Sale and the Item yes. There are many options for exploring the connections with this outcome, but the most helpful will be View Predictors link.

 

Clicking on View Predictors will display the predictors of positive sales, combining the frequency and probability of the connection between each item, or even between complex groups of items and your selected outcome.

 

The software displays every one of the 1,891 patterns that have meaningful connections to the successful sales outcome. Now that the historical data has been captured and analyzed you can plan future sales ventures that will maximize the effectiveness of your sales force and their campaigns.

 

To take this analysis even further, the Emcien Services team can implement end to end integrations to apply these predictions to new leads and feed directly into your CRM system, even making predictions in real time.

Apply predictive patterns to new data.

Now that the patterns that identify successful sales opportunities have been identified, Emcien’s real-time prediction engine can be used to evaluate and qualify leads as they progress through the sales pipeline. This application of our two-tiered architecture can be demonstrated by clicking on Predictions in the menu at the top of the screen. Simply upload the file of new or incoming data to the same FTP address.

Once your new sales prospects data has been uploaded, click on the green Predict button and select the file. You can let the software append (Prediction) to the existing file name, or you can create a new name for your prediction.

 

Then just click Predict and the Predictions module will apply your analyzed rules to the incoming data. The new data will be scored with all of the rules that justify each prediction and output to a machine and human-readable .CSV file.

  

Used inside of the Emcien application these predictions can be very useful, but with the help of our Services team you will be able to automate this workflow. Through our automation tools, Emcien can create a custom data flow that will detect new prospects, use the latest analysis to generate a predictive score, and insert that predictive score directly into your CRM system. Contact the Emcien Services team at [email protected] to find out how.