Daily Nudge System — BASELINE

This workflow sends automated daily nudges to youth entrepreneurs at 7PM and 8PM via SMS or WhatsApp, encouraging them to track their business sales and profits. It operates as a baseline measurement system, logging all nudge decisions for analysis while not actually sending messages to users.

Purpose

No business context provided yet — add a context.md to enrich this documentation.

How It Works

The workflow operates on two daily schedules (7PM and 8PM) and follows this process:

1. Schedule Triggers: At 7PM and 8PM daily, the workflow checks if it's Sunday (skips execution on Sundays)
2. Baseline Logging: Creates a baseline run record via webhook for tracking purposes
3. User Fetching: Retrieves active users from the database, including both regular users and internal team members
4. Channel Detection: For each user, determines whether they prefer WhatsApp or SMS communication
5. Chat History Analysis: For WhatsApp users, checks recent chat activity to avoid duplicate messaging
6. Message Generation: Creates personalized nudge messages asking about daily sales or profits
7. Decision Logging: Records all nudge decisions (channel choice, message content, eligibility) via webhook
8. Database Storage: Saves nudge messages to chat history for tracking purposes

The workflow processes three user groups: - 7PM Regular Users: Active entrepreneurs who onboarded before today - 7PM Internal Team: Internal team members for testing - 8PM All Users: Combined processing with enhanced chat history filtering

Mermaid Diagram

graph TD
    A[7PM Nudge Trigger] --> B[Create Baseline Run 7PM]
    C[8PM Nudge Trigger] --> D[Create Baseline Run 8PM]

    B --> E{Is Sunday? 7PM}
    D --> F{Is Sunday? 8PM}

    E -->|No| G[Flag Evening Session Started]
    F -->|No| H[Fetch Active Users 8PM]

    G --> I[Fetch Active Users 7PM]
    G --> J[Fetch Internal Team Users]

    I --> K[Dedup Users 7PM]
    J --> L[Dedup Users Internal]
    H --> M[Dedup Users 8PM]

    K --> N{Is WhatsApp User? 7PM}
    L --> O{Is WhatsApp User? Internal}
    M --> P[Loop Users 8PM]

    N -->|Yes| Q[Loop WhatsApp Users 7PM]
    N -->|No| R[Loop SMS Users 7PM]

    O -->|Yes| S[Loop WhatsApp Users Internal]
    O -->|No| T[Loop SMS Users Internal]

    P --> U[Fetch Chat History 8PM]
    U --> V{Has Chat Data? 8PM}
    V -->|No| W{Is WhatsApp User? 8PM}

    Q --> X[Fetch Last WhatsApp ChatLog 7PM]
    R --> Y[Set SMS Nudge - Sales Question 7PM]
    S --> Z[Fetch Last WhatsApp ChatLog Internal]
    T --> AA[Set SMS Nudge - Sales Question Internal]

    W -->|Yes| BB[Fetch Last WhatsApp ChatLog 8PM]
    W -->|No| CC[Set SMS Nudge - Sales Question 8PM]

    X --> DD{Has WhatsApp ChatLog? 7PM}
    Z --> EE{Has WhatsApp ChatLog? Internal}
    BB --> FF{Has WhatsApp ChatLog? 8PM}

    DD -->|Yes| GG{ChatLog Within 24hrs? 7PM}
    EE -->|Yes| HH{ChatLog Within 24hrs? Internal}
    FF -->|Yes| II{ChatLog Within 24hrs? 8PM}

    GG -->|No| JJ[Set SMS Nudge - Profit Question 7PM]
    HH -->|No| KK[Set SMS Nudge - Profit Question Internal]
    II -->|No| LL[Set SMS Nudge - Profit Question 8PM]

    Y --> MM[POST Decision 7PM SMS]
    AA --> NN[POST Decision Internal SMS]
    CC --> OO[POST Decision 8PM SMS]
    JJ --> PP[POST Decision 7PM WhatsApp]
    KK --> QQ[POST Decision Internal WhatsApp]
    LL --> RR[POST Decision 8PM WhatsApp]

    MM --> SS[Save Nudge to DB 7PM SMS]
    PP --> TT[Save Nudge to DB 7PM WhatsApp]
    NN --> UU[Save Nudge to DB Internal SMS]
    QQ --> VV[Save Nudge to DB Internal WhatsApp]
    OO --> WW[Save Nudge to DB 8PM SMS]
    RR --> XX[Save Nudge to DB 8PM WhatsApp]

Trigger

• Schedule Trigger (7PM): Runs daily at 7:00 PM
• Schedule Trigger (8PM): Runs daily at 8:00 PM
• Sunday Skip: Both triggers skip execution on Sundays (weekday = 7)

Nodes Used

Node Type	Count	Purpose
Schedule Trigger	2	Daily execution at 7PM and 8PM
Postgres	15	Database queries for users, chat logs, and data storage
HTTP Request	10	Webhook calls for baseline run creation and decision logging
If (Conditional)	12	Logic branching for user types, chat history, and day checking
Set	6	Message content preparation for different nudge variants
Split in Batches	5	User processing loops
Remove Duplicates	3	User deduplication by phone number
Filter	1	Recent chat filtering (within 60 minutes)
Aggregate	1	Chat data consolidation
No Operation	1	Data passthrough

External Services & Credentials Required

Database Credentials

• Postgres account 2: Database connection for user data, chat logs, and nudge storage
  • Required tables: youthEntrepreneursReal, youthProgressReport, chatLog, n8n_chat_histories

Webhook Endpoints

• Baseline Run Creation: https://sifa.alpha-test.kriftx.app/webhooks/baseline/nudge-run/[token]
• Decision Logging: https://sifa.alpha-test.kriftx.app/webhooks/baseline/nudge-decision/[token]

Environment Variables

No explicit environment variables are used. Configuration is embedded in node parameters: - Webhook URLs with authentication tokens - Database connection details via credentials - WhatsApp contact URL: https://wa.me/254203892316

Data Flow

Input

• User Data: Phone numbers, names, business info, onboarding status from database
• Chat History: Recent WhatsApp interactions and SMS chat logs
• Trigger Time: 7PM or 8PM execution context

Processing

• User eligibility filtering (onboarded before today, not 'new' status)
• Channel preference detection (WhatsApp vs SMS)
• Recent activity analysis (24-hour and 60-minute windows)
• Personalized message generation with user names

Output

• Baseline Run Records: Execution tracking via webhook
• Decision Logs: Channel choices, message content, user eligibility via webhook
• Chat History: Nudge messages stored in n8n_chat_histories table
• Database Updates: Evening session flags in user records

Error Handling

The workflow includes several error handling mechanisms:

• Continue on Error: Most database operations use continueRegularOutput or continueErrorOutput
• Never Error Responses: HTTP requests configured with neverError: true for webhook calls
• Always Output Data: Critical nodes marked to ensure workflow continuation
• Graceful Degradation: Missing chat logs or user data don't stop processing
• Fallback Paths: Alternative message variants when primary conditions aren't met

Known Limitations

Based on the workflow structure: - Sunday Exclusion: No nudges sent on Sundays - Timezone Dependency: Uses Africa/Nairobi timezone for day calculations - No Actual Sending: This is a baseline measurement system - messages are logged but not delivered - Internal Team Filtering: Some queries exclude internal team members while others include them - Chat History Dependency: WhatsApp logic relies on chatLog table availability

Related Workflows

No related workflows mentioned in the provided context.

Setup Instructions

1. Import Workflow: Import the JSON into your n8n instance

2. Configure Database Credentials:

   • Create PostgreSQL connection named "Postgres account 2"
   • Ensure access to required tables: youthEntrepreneursReal, youthProgressReport, chatLog, n8n_chat_histories

3. Set Up Webhook Endpoints:

   • Configure baseline run creation webhook URL
   • Configure decision logging webhook URL
   • Ensure authentication tokens are properly embedded

4. Verify Database Schema:

   • youthEntrepreneursReal: phoneNumber, firstName, channel, businessOwned, preferredTime, eveningSessionStarted
   • youthProgressReport: phoneNumber, onboardingStatus, onboardedAt
   • chatLog: userPhone, channel, created_at
   • n8n_chat_histories: session_id, message, created_time

5. Test Execution:

   • Manually trigger to verify database connections
   • Check webhook endpoints respond correctly
   • Validate timezone settings for Africa/Nairobi

6. Activate Workflow: Enable the workflow for daily execution at 7PM and 8PM

7. Monitor Logs: Check execution history and webhook responses for proper baseline data collection