Time-based cohort analysis

Time-based cohorts are helpful for understanding a pattern or change that happens weekly, monthly, yearly, or quarterly. For example:

• Studying month-wise deal closure pattern: You can run a cohort analysis to see the deal closure pattern for the past few months. The total number of deals that were closed in the first, second, third, and fourth month from the month they were created in will help you understand how the deal closure pattern varies by mont. The cohort below shows that the deals created in February and March were most likely to be closed the first and second months from deal creation and that there was a considerable drop inclosure rates after the second month from deal creation.
   
• Day-wise product purchase frequency: In retail businesses, it is useful to see how often a particular product is purchased by customers. You can run a cohort analysis for the product category with a timeline from sale start to sale end date (purchase end date) and see how many times a product was bought from the first to fifth day of a sale period. The data below shows a considerable increase in the purchase on the third day. The data also indicates that customers made more purchases in the electronics category than the other categories.
  

Business size-wise purchase pattern: Let's look at another example where the marketing strategies can be tweaked based on product's purchase frequency. A cohort analysis of different business sizes versus their purchase frequency can help you understand if a certain product interests a certain type of business or not and then market the products based on the audience's interest and requirements. 

• Basic: A basic cohort lets you analyze by date. For example, you can choose the date a record was created or modified, the date of last activity or any other custom date/time field. 
• Standard: A standard cohort lets you run an analysis using dimensions or pick-list fields. You can use standard or custom picklist fields such as industry type, product category, lead source, lead status, and record owner.
  
• Advanced: This is a combination of basic and standard cohorts. You can analyze using either dates or dimensions as the parameter and add additional metrics such as the total annual revenue, total number of leads, number of purchases, or average unit price. Any number field can be used for additional metrics.
   

To configure cohort analysis

1. Click Add Component in the Analytics tab and select Cohort.
   
2. Select your desired cohort type: basic, standard, or advanced.
3. Enter the following details in the Cohort Builder page
   • Component Name: Enter the component name, such as "purchase frequency after end-of-season sale" or "sales cohort for lead sources".
   • Cohort Analysis For: Select a module and grouping from the dropdown list. Note: You can group by either date or dimension.
   • Cohort Data: Select the number fields you want to run the cohort for, e.g., record count, total purchases, product count versus the number of records.
   • Criteria Filter: Specify a criteria if you only want specific records to be considered for analysis.
   • Duration from: Select the duration of your analysis, e.g.,sales start to end date, purchase start to end date, created time to modified time. Note: Only date and date/time fields will be listed here.
     
   • Duration interval: Choose Day, Week, Month, Quarter, or Year.
     
   • Duration Range: Select the range you want for the results, e.g. 1-4 or 1-6. The maximum duration range is 1-10.
     
   • Additional Metric: Click Add Metrics to add more values from any number field to the cohort. You can also define criteria to consider specific records only.
     
4. Click Done to save the additional metrics.
5. Click Add & Reorder to change the configuration.
6. Click Save.