Setting up your AI scoring

Welcome to MadKudu. Below you will find the steps you will go through during your onboarding. Each step has detailed documentation linked to it. Some steps are marked with a 💡 which means they require a decision related to your use case. Steps marked with a ⚙️ contain set-up instructions.

Basic Setup (also covered by MadKudu Copilot Setup)

  1. ⚙️ Create your Account (you will receive an invite to do so)

  2. ⚙️ Connect your systems (also covered by MadKudu Copilot Setup)

    1. Connect Salesforce

    2. Connect MAP (Integrations overview)

    3. [Optional] Connect other signals (ex. LinkedIn)

  3. ⚙️ Installing package in Salesforce

Data Mapping

  1. ⚙️ Set up attribute mapping and computations

    1. 💡 Determine which fields you want to try using in scoring

    2. ⚙️ Pull fields into MadKudu

    3. ⚙️ Map fields as attributes in app

    4. ⚙️ Go to MadKudu Data Studio

    5. ⚙️ Create computations in data studio

  2. ⚙️ Set up event mapping

    1. 💡 Determine which behaviors (ex. page visits, product actions) you want to try using in scoring

    2. ⚙️ Map behaviors as events in app

  3. ⚙️ Create conversion mapping

    1. 💡 Define the conversion metric to train your models

    2. Use definition to set up conversion types

  4. ⚙️ Create audience mapping

    1. Define the people and accounts to train your models

    2. Use definition to set up audiences

Model Setup

  1. ⚙️ Go to MadKudu Data Studio

  2. ⚙️ Create person-level fit and engagement (aka “likelihood to buy”) models

    1. Set up fit model

    2. Set up engagement model

  3. 💡 Add business overrides into fit model

  4. 💡 Review model performance

  5. ⚙️ Deploy models

    1. Deploy fit model as “standard fit” when prompted

  6. ⚙️ Set up push configuration

  7. 💡 Review data in CRM or MAP

Workflow Setup

  1. ⚙️ Update lead or account-level workflow to use MadKudu scores

    1. If no workflow exists, build workflow in CRM or MAP

  2. ⚙️ Build reports to determine scored lead or account volume

  3. 💡 Determine score thresholds

Roll Out

  1. Run enablement session with sales team

  2. Monitor scored lead or account volume on a daily basis

    1. Update score threshold in workflow to meter scored lead or account volume

  3. Set up sales feedback loop to make iterations over time

  4. Review scored lead or account performance at least monthly

  5. Iterate on scoring models every 3-6 months, using sales and marketing feedback