Prerequisite

After Transactions and Product Bundles are loaded, create a Segment of Accounts where your selling scenario fits. Only Accounts that have transaction data in the system can get Ratings from Cross-Sell Models. (Any other Accounts in the segment will not be Rated, nor used in model training.)

 

To build highly-targeted ranking models, refine your Segment up front to include exactly the customers you want to put in the selling scenario. For example, if you want to sell a solution component to past purchasers of related products, select those past purchasers into your Segment.

 

Steps

Step 1

 

 

Use the left navigation bar to land on the Models page, then click Create Model.

 

Step 2

 

 

The first two tiles are cross-sell models.

Choose,

• Tile 1: First Purchase. Build a model to sell a Product Bundle to the customers that haven’t bought that Bundle before. Use this to sell durable products, or to break into new customer cohorts with consumable goods or renewable licenses. If some customers in the Segment have already bought it, they won’t get Rated.
  
• Tile 2: Repeat Purchase. Build a model to sell a Product Bundle to the customers that did buy a Product Bundle in the past, but now haven’t refilled within the expected timeframe. Use this to win back customers that may be refilling elsewhere, or to be proactive with a renewal or upgrade conversation. If some customers in the Segment have never bought, or have refilled too recently, they won’t get Rated.

 

Step 3

 

 

Search or browse for your Segment that includes the targets of your selling situation, then click the box for the Segment that has your target customers.

Step 4

 

 

Search or browse your Product Bundles, then click the box for the Product Bundle you want to sell. (You can choose several, but make sure they are similar enough to work interchangeably in a campaign--pricing, solutioning how you sell and so forth.)

 

For Repeat Purchase, you must select the timeframe that signifies a lapse so the model can find the right targets and signals for you.

 

Step 5

 

 

Click a tile to choose how you would like to prioritize:

• Likely to Buy (default). Build a model to rank based on buying propensity. This model uses classification modeling to predict buying probability.
• Likely Amount of Spend. Build a model to rank based on a combination of propensity and forecast revenue (also called ranking by Weighted Revenue). This model layers revenue regression with classification modeling to provide sales volume forecasting.

 

Step 6

 

 

Review the size of the training data and number of target Accounts. Click Next to build the model. You can also use the settings on this page to change the way your model will be trained. All selected Settings will be applied. When settings are used, the counts at right will be updated. (Settings never change the number of targets, only the data used in learning.)

 

You must have at least 50 First or Repeat Purchases to learn from. If you proceed with less, your model build will fail.

• Checkbox 1. This Setting removes purchases from learning that do not meet the criteria. Use this when Product Bundle pricing varies widely to focus the model on the business you are looking for.
• Checkbox 2. This Setting removes purchases from learning that do not meet the criteria. Use this when purchase quantity for the Product Bundle varies widely to focus the model on the business you are looking for.
• Checkbox 3. This Setting forces the model to look only at data with the selected recency. Use this when something significant changed in the past (with Account, Transaction or Product tracking) and you want to keep older data out of the learning.
• Similar Segment. This Setting uses separate filters for model training, but still Rates in the original Segment. Data Scientists use this to gain full control over training data.
• Similar Product. This Setting replaces the actual Product Bundle in your selling situation with a different Bundle just for learning. The actual Bundle is still used for selecting target Accounts. Data Scientists use this when there is insufficient data to learn on the actual Bundle, so a similar one with fuller history is used in place.

 

 

 

Step 7

 

 

Review your model job. Confirm all of your settings, and monitor job progress. You can also track the jobs and history by clicking on the gear at top right. Modeling jobs can take several hours to complete.

 

Step 8

 

 

You will receive an email when your job has completed.

 

Step 9

 

 

Log into Lattice to review your model and Activate Scoring.  Navigate to the Models page to find your new Model tile. You can name your model by clicking on the menu in the tile’s top right corner.

 

Step 10

 

 

Click on the Model tile to land on the details page for your model. The left navigation bar in this view links you to the Attributes, Performance, and Summary pages for your model. These pages provide you with lots of information you can use to review the quality of your model. See also “Get Started with Model Review” to learn more about the charts and data in these pages.

 

Step 11

 

 

Click Activate Scoring on the model dashboard to see what your model predicts for your target Accounts. You can see the Rating buckets, and change the bounds to tune your Ratings.

 

Grab x-axis tabs on the bucket separators to drag, add and remove the separators.

 

Lattice shows your model performance as Lift and Account Volume in each Rating bucket.

• For Likely to Buy models, Lift is the estimated conversion probability scale factor you will see in the Rating bucket, compared to all targets.
• For Likely Amount of Spend models, lift is the scale factor of weighted revenue in the Rating bucket, compared to all targets.
• Account Volume is the percentage of target Accounts in the Rating bucket.

 

Step 12

 

 

Click Publish Configuration to create a Scoring Action in your Process & Analyze queue. The next time you run Process & Analyze, Your Ratings and Scores will be updated in My Data. You can use them to make new Segments, Rules-Based Ratings, or Plays.