How to analyze a fit model

Learn how to analyze a fit model correctly, understanding how predictive it is and how to take advantage of everything Infer has to offer.

Hello, in this video we will be looking at how to analyze a Fit model. To view the model you need to login to your infer portal and select the models tab on the left hand menu.

Here you will see a list of all the models that are deployed for you and some details showing what they are. This includes the type of entity that the model is scoring, whether they are fit or behavioral models and when they were created In this case, we will be looking at the Lead to Opportunity model, so click on the name in the list and it will open the model view.

The screen now changes to show an overview of the model, with a description of the model itself and the details of how the model has been set up. Below this there is the visual of the model evaluation. This shows you a sliding scale of scored leads going from the highest score to the lowest and grouped into 4 “buckets” from A-D. Each line is a lead that has been scored by the model and they are segregated by the thresholds that have been set for each group. The green lines are leads that have converted and the grey lines are ones that have not. The ideal scenario is that most of the green lines are in the A and B groups and the grey are mostly in D. The A group is the leads with the highest scores and those that would best match the criteria to result in a won opportunity. The table below shows the conversion rate for each group and the sales funnel to demonstrate the probability and worth of the predicted deals for the buckets. 

Scrolling down the page further there is a list of example leads for each group so you can get a feel for the type of lead which comes under each section, including the Salesforce ID, company, score, whether the lead was converted and the amount of the opportunity. You can change the group shown by clicking on the tabs to the right of the screen

There are also other tabs across the top of the model view page, let's take a look at the first one called key signals. This tab gives you a chart which shows the signals that the model is using to determine the score of the lead. If you hover over one of the dots on the chart it will highlight and show you the name of the signal and the information on the right hand side of the screen will change to show you the details for that signal. This includes the average deal size for leads in which this signal is present and the average conversion rate.

Next we have the scoring matrix. On this tab you can see how fit and behavioral models interact together if you have both available, and how likely to convert each entity is.. You can compare the groups on conversion multiplier, conversion rate or coverage. 

The final tab on the page is score csv. Using this tab you can upload a csv file and let the model run and give each entry on the file a score according to the algorithm. This can be useful if you have a group of new leads outside your CRM  and you want to see if they are worth following up on. You can then prioritize them by the score they are given as to which ones are worked first. To do this, just select the file you want to score and click upload. Once the file is processed it will show in the list below and you can edit the mapping used, delete the file, download the original data, and download the scored file.

I hope you found this video on fit models useful, please take a look at our other videos and articles for more information.