MadKudu Docs

Home

Introduction to MadKudu

MadKudu analyzes data from your existing customer to predict who is most likely to convert and when is the perfect time to engage.

MadKudu uses several types of customer data:

  • Product usage: All the data captured when people are using your product. It usually comes from your web analytics platform (e.g. Mixpanel, Kissmetrics) or your own internal database. It includes every user profiles and the events associated to them. For example, “team member added”, or “report created”.

  • CRM information: All the data that gets into your CRM (Leads, Contacts, Accounts, Opportunities).

  • Support interactions: “Tickets” created by your customers in your support system (e.g Zendesk).

  • Marketing interactions: Sent by your marketing automation solution (e.g. Marketo) or your communication app (e.g. Intercom). It includes your contact profiles and the history of marketing interactions including events like “received a message”, “opened a message”, “clicked a message”, etc.

The most important data source is product usage. It is often a great start and other data sources can be provided later when it makes sense.

Predictive Models

MadKudu ingests as much data as possible in order to predict future behaviors. We have several models that fit different needs:

  • Customer Fit: Looks at firmographics and demographics data to identify who is most likely to convert.
  • Likelihood to Buy: Looks at product usage data to identify patterns of behaviors that lead to conversions.
  • Lead Grade: We use Customer Fit and Likelihood to Buy to grade your leads (usually ranging from A to E).
  • Potential Account Value (PAV): Predicts how much an account will be worth to your business so that you can focus on your best deals.

We configure these models independently so if you’re not sure which ones are enabled in your account, reach out to your Customer Success Manager.

How do I send data?

There are several ways to start sending data (click on each one to get more details):

  • One-click integrations: We have dozens of integrations with best-in class systems like Segment, Salesforce or Marketo. You can enable these by the click of a button.

In the left navigation menu, click on the Integrations icon. The next screen will show you a list of all available integrations. To enable one, click on it and follow the instructions.

  • Flat files: Send your raw data as flat files by uploading them to Amazon S3.

  • REST API: By sending data directly to MadKudu.

You should also learn more about how do we sync scores to your integrations.

Glossary

  • Recall: Percentage of conversions that we accurately predicted.
  • False Negatives: Leads who scored low and that ended up converting.
  • False Positives: Leads who scored high but that the sales team disqualified.