Sign inStart free

Menu

For data teams

Stop reverse-engineering campaign names.

Today your warehouse has UTM strings and campaign names, and every analysis starts by guessing what marketing meant. With Kleardata the kd_ ID in each URL and platform name joins to a registry row that already knows the goal, stage, audience, campaign and test.

Start freeAsk a question

One row in the registry

kd_id
kd_4a2w917
program_name
Spring teams promo
touch_name
Upgrade nudge email
audience
Free workspaces
lifecycle_stage
Prospect
planned_start
2026-11-09
Your sessions, conversions and spend join on kd_id.

How the pieces fit.

  1. The registry

    One row per kd_ ID, with every ancestor and resolved dimension: program, campaign, unit, touch, version, test and role, wave and planned dates, market, objective and target, audience and lifecycle stage, partner, strategy goal and marketing stage.

  2. The nightly export

    A manifest and four tables as Parquet, readable at a tokenized URL without an API key. CSV is available if your loader prefers it. The columns are a documented contract, and tests fail if the registry and the package drift apart.

  3. The dbt package

    Point it at your own sessions, conversions and spend tables. It finds the ID in the columns you name and builds results by strategy goal, marketing stage, audience and test variant.

What the package builds

  • dim_kd_registry the dictionary of intent
  • fct_sessions, fct_conversions, fct_spend
  • fct_kd_daily everything on one key, by day
  • results_by_strategy_goal
  • results_by_marketing_stage
  • results_by_audience
  • results_by_test_variant
  • data_coverage how much of your data carries an ID

What stays where

  • Kleardata holds planned intent. Targets, budgets and goals as the team meant them.
  • Your warehouse holds actuals. Sessions, conversions and spend never leave it.
  • Kleardata never needs access to your warehouse. You load our export; we do not connect to yours.
  • Download works on every plan, forever. Automated exports and the package start on Starter.

Requires dbt 1.10 or later. The SQL is written for Snowflake, BigQuery and Databricks, and there is a ready loader for DuckDB and MotherDuck.

Proof of value, from the same join.

When intent and actuals sit on one key, marketers can report results by strategy goal, stage, audience and campaign in the BI tools they already use. Budget conversations stop being arguments about attribution and start from what the plan said it would do.

See how strategy and coverage work

Give your analysts the key.

Start free See pricing