Frequently Asked Questions
1. Do all Accounts in my Segment get Ratings?
Sometimes, but not always. If you built the Segment so that all Accounts in it are right for the selling situation, then all Accounts will be Rated. On the other hand, you can also use Cross-sell Ratings to surface the right targets for your selling situation from within a larger Segment.
The following Accounts in a Segment will not get Cross-Sell Scores or Ratings:
- Accounts that do not have any transactions in the system
- Accounts that do not fit the selling situation, e.g., for First Purchase Accounts that have already purchased will not get Ratings
When exporting Ratings, Accounts that are not rated will have an empty Rating.
2. When do my Ratings get updated?
Each time the Process & Analyze job runs, all Active Ratings are refreshed:
- Target Accounts within the Segment are updated
- The updated targets are given new Scores and Ratings
- The new Scores and Ratings are available in My Data
- The Model dashboard will show the updated Ratings in the Lift chart.
If your updated Ratings show a performance decline, try building a new Model so that your Model is also up to date with your data.
3. When can I update my Rating buckets?
You can update your Rating buckets any time. Keep in mind that the new buckets will be applied after the next Process & Analyze job completes.
Update Rating buckets to adjust the number of targets receiving top Ratings. This way, you can get exactly the campaign membership you need to optimize budget and capacity.
4. What does Lift mean in Atlas?
Lift is a measure of how top Ratings will perform against the full set of targets.
- 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.
5. How can I use my Ratings in workflows?
There are several important ways you can use Ratings:
- Use Ratings to create new highly targeted Segments by adding Lattice Ratings attributes to Segment Queries
- Use Ratings within the Rating rules of a Rules-Based Model to combine predictive dimensions within a campaign
- Use Ratings within a Play to launch a sales campaign through BIS.
For 1&2, note that Ratings may have one level of dependency. Deeper nesting is not supported.
6. How does Atlas compute buying propensity (Likely to Buy)?
Lattice uses binary classification modeling to predict buying probability. The underlying algorithm is a random forest.
7. Does Atlas do Revenue Modeling / What does it mean to Rate for Likely Amount of Spend?
To layer revenue modeling in with propensity, choose Likely Amount of Spend for Prioritization in the Create Model wizard. The underlying algorithm is a random forest for propensity, with logistic regression for revenue forecasting. Scores from the two models are multiplied to generate a weighted revenue forecast for prioritization.
8. Can I decide how I want to set up training the model?
Yes. In the Training step of the Cross-sell Model Create wizard, several settings can be used alone or together to engineer the training data in whatever way the situation calls for.
These Settings are,
- 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.
You must have at least 50 First or Repeat Purchases to learn from. If you proceed with less, your model build will fail.