The Agency Data Pipeline

Fully managed data analysis for digital agencies

* If you’re here to grab a copy of the Agency Data Pipeline Tracking Plan, head straight here to grab it from the CIFL Template Vault.

Every successful digital agency sings a particular song on behalf of their clients.

You sing a song that helps your clients grow, but not in a halting and haphazard way - through a diligent, organized and repeatable process.

The Agency Data Pipeline allows you to lay that song down on wax, to be replayed at will, by codifying your agency’s framework for working with client data.

Think of the CIFL team and the Agency Data Pipeline service as your agency’s data analysis rhythm section - we hold the pocket so that your team can be free to lead.

You can almost hear your team rejoicing (shhh listen for it), as they can fire themselves from their least favorite job: wrangling data for research and reporting.

Now you can also build your own Agency Data Pipeline, with a self-paced course that includes a 1-hour office hours session with David, the Commissioner of CIFL, to put your data pipeline tracking plan into action.

Who the ADP serves

  • PPC agencies: the classic pipeline involves joining all campaign data together to pluck out anomalies, dial performance, and balance your spend between channels.
  • SEO agencies: moving the voluminous keyword, content, and technical research out of spreadsheets and into the big leagues.
  • eCommerce agencies: mining product sales data to model top performing products, purchase paths, customer + product segments.

THE AGENCY ANALYTICS TRACKING PLAN

When building out an analytics pipeline, your tracking plan is the sun, moon and stars all at once.


The Tracking Plan allows you to stay organized with the key datapoints needed to manage your pipeline:

  1. The questions your analysis is seeking to answer.
  2. The data sources (APIs) and data models (SQL queries) required to answer those questions.
  3. Site or client-specific background info: their domains, conversion goals, and channel mappings
  4. Statuses of the finer implementation points, like any good project management tool.

Best of all, the Tracking Plan talks directly to BigQuery (being able to push data up and pull it down), so it’s able to act as the nerve center for your data pipeline.

Curious to dive into your own Tracking Plan?  Grab a free copy from the CIFL Template Vault.

What’s the end output?

Watch for a quick walkthrough of the ADP process end-to-end, including all of the glorious components and tools in use:


  • Data models: SQL transformations of your raw data into metrics, sleeping cozily in a Github repo.

alt text

  • A Google BigQuery database: containing all of your agency’s client data, pre-formatted to be ingested by reports and research tools.

alt text

  • Data Studio report templates: to share that data internally amongst your team to dial strategy, and externally with clients to report on results.

alt text

Data is loaded, modeled and QC’d on a regular basis, so all of the above are always up to date.

How is it built?

CIFL sends blasts of love to the manufacturers of parts used in the Agency Data Pipeline:

  • DBT: the open-source data modeling framework.
  • Sinter: a helper tool built by the DBT team to run SQL models on a schedule.
  • Stitch: the crispiest ETL platform out there.
  • Supermetrics: the handy Google Sheets Add-on for ETL and data QC.
  • Google Sheets, Scripts and Data Studio: ‘nuff said.

We stand on the shoulders of giants, without whom such a glorious data analysis process would not be possible.

If flow charts are more your thing, here’s how a completed Agency Data Pipeline flows:

alt text

Ready to dive in?

Warning: embarking on building your own Agency Data Pipeline is a major undertaking, suitable only if you’re serious about standardizing your agency’s data analysis process.

Pipeline construction and maintenance costs range for $4,000 - $8,000 per month, with a minimum 6-month commitment.

If you’re curious to learn more about having CIFL build out your pipeline, drop a note to help@codingisforlosers.com with answers to a few questions:

  1. Background on your team - what’s your current data analysis process, and who runs it?
  2. The song you sing for your clients - what’s your go-to strategy?
  3. Why is data analysis on your radar?

If you’re looking to build out your agency’s own analytics pipeline, head on over to the Build your own Agency Data Pipeline course.

With love,

David Krevitt

Commissioner, Coding is for Losers