How to build Tomba deliverability dashboards in Metabase

{
  "mcpServers": {
    "tomba": {
      "url": "https://mcp.pipedream.net/v2"
    }
  }
}

# Install the Metabase CLI
npm install -g @metabase/cli

# Log in (opens your browser; requires Metabase v62+)
mb auth login --url https://your-metabase.example.com

# Load a Tomba CSV export — creates a table AND a model
mb upload csv --file tomba-emails.csv --collection root

# Refresh that same table later from a new export
mb upload replace <table-id> --file tomba-emails.csv

Create a polished Metabase dashboard for Tomba lead-finding analytics.
Work end to end: get the data into Metabase if it isn't there yet, then build.

Goal: Help RevOps and deliverability teams understand match rate, confidence,
verification quality, credit usage, and list quality from Tomba data.

Step 1 — Find or load the data:
- First, check what already exists in Metabase (search for Tomba tables and
  models). If durable Tomba data is already present — synced from a warehouse or
  uploaded earlier — use it and skip to Step 2.
- If nothing is there, pull it with the Tomba MCP server: searches (finder
  requests), found emails (with confidence score and sources), verifications
  (deliverable/risky/undeliverable, MX/SMTP checks, catch-all), and credit/usage
  events. Write each result to a CSV, then load it with the Metabase CLI — run
  "mb upload csv --file <export>.csv" so each upload creates a table and a
  ready-to-query model. Use "mb upload replace <table-id> --file <export>.csv" to
  refresh an existing table instead of creating duplicates.

Step 2 — Inspect before querying:
Tomba CSV exports are usually flat (one row per found email, with a confidence
score, source count, and verification status). Warehouse tables are rawer, often
with a separate searches table and a per-verification row. Inspect the actual
tables and column names first; do not assume exact names or that a searches table
exists.

Important:
- Build on whatever data is present; don't claim Metabase connects natively to
  Tomba — it reads a database or CLI-uploaded tables.
- Define match rate as found ÷ searched, and deliverable share as deliverable ÷
  verified — keep them distinct.
- Use the confidence score as a quality band, not a guarantee, and treat
  catch-all separately from deliverable.
- Report credits with a denominator (per usable contact) visible.
- Only build a card if its underlying column/metric exists in the data.
- A single CSV is a point-in-time snapshot: only build trend cards if there is a
  usable date column or multiple periods have been uploaded.

Dashboard title: Tomba Deliverability Overview

Sections:
1. Executive summary (KPI cards): Match rate; Deliverable share; High-confidence
   share; Credits used; Credits per usable contact; Usable contacts added.
2. Match rate: By search type; found vs. not found; confidence distribution.
3. Verification: Deliverable vs. risky vs. undeliverable; MX/SMTP pass; catch-all.
4. Credits: By week; per usable contact; wasted credits.
5. List quality: Usable contacts added; duplicates vs. CRM; high-confidence share.

Filters: Search type, Confidence band, Date range.

Reuse the models Metabase auto-created from uploaded CSVs, or (for a warehouse)
create reusable models: modeled_tomba_searches, modeled_tomba_emails,
modeled_tomba_verifications, and modeled_tomba_credit_events.

Output: Build the dashboard if you have permission; otherwise provide the exact
questions, SQL, model definitions, and layout. Include caveats for any metric
that cannot be calculated from the available data. Reconcile totals against
Tomba's own usage views. Keep it practical, dense, and executive-readable.

Found as a share of searches over the last 90 days.

-- Match rate by search type over the last 90 days
SELECT
  s.search_type,
  COUNT(*)                                             AS searches,
  COUNT(*) FILTER (WHERE s.found)                      AS found,
  ROUND(100.0 * COUNT(*) FILTER (WHERE s.found)
    / NULLIF(COUNT(*), 0), 1)                          AS match_rate_pct
FROM modeled_tomba_searches s
WHERE s.searched_at >= CURRENT_DATE - INTERVAL '90 days'
GROUP BY s.search_type
ORDER BY searches DESC;

Deliverable share across high, medium, and low confidence finds.

-- Deliverability by confidence band
SELECT
  CASE
    WHEN e.confidence >= 90 THEN 'high (90+)'
    WHEN e.confidence >= 70 THEN 'medium (70-89)'
    ELSE 'low (<70)'
  END AS confidence_band,
  COUNT(*)                                               AS emails,
  ROUND(100.0 * COUNT(*) FILTER (WHERE v.status = 'deliverable')
    / NULLIF(COUNT(*), 0), 1)                            AS deliverable_pct
FROM modeled_tomba_emails e
JOIN modeled_tomba_verifications v ON v.email_id = e.id
GROUP BY confidence_band
ORDER BY confidence_band;

Cost efficiency by week, using deliverable emails as the usable base.

-- Credits per usable contact by week
SELECT
  date_trunc('week', c.used_at) AS week,
  SUM(c.credits)                AS credits_used,
  COUNT(DISTINCT e.id) FILTER (WHERE v.status = 'deliverable') AS usable_contacts,
  ROUND(SUM(c.credits)::numeric
    / NULLIF(COUNT(DISTINCT e.id) FILTER (WHERE v.status = 'deliverable'), 0), 2)
    AS credits_per_usable_contact
FROM modeled_tomba_credit_events c
LEFT JOIN modeled_tomba_emails e        ON e.id = c.email_id
LEFT JOIN modeled_tomba_verifications v ON v.email_id = e.id
GROUP BY 1
ORDER BY 1;