• Navid Nassiri

How to deploy DataOps: Step 1 - Identify KPIs

In our last post, we set out the four steps to realizing the benefits of using DataOps (identify KPIs, normalize raw data, transform foundational data, and use the right tools to automate the process).


So, as a data-driven organization, how do you identify KPIs that will help you measure and improve performance?


A DataOps approach aims to provide value by enabling data collaboration across an organization. Unfortunately, in their rush to set up metrics, business teams will often base their choice of measurements on whatever data is available or is easiest to present. This inevitably results in arduous, manual work when it comes to providing higher-level insights. Developing a DataOps strategy starts not with technology, but with organizational communication.


Your initial goal should be to recognize, at a high-level, which specific metrics you want to measure. It may sound obvious, but it‘s essential to start the process by targeting questions that will yield the insights you need to drive key decisions.


For example, these are just some of the questions a digital company might ask:


  • Where are we likely to have sell-out or unfilled inventory over the next 30 days?

  • How are monetization rates changing across our verticals over time?

  • Which programmatic deals bring us the most revenue and highest eCPM?

  • How can we ensure the most valuable inventory with the highest viewability is sold first?

  • What are our most valuable audiences in terms of viewability and sell-through?

  • How can we best optimize our sales between programmatic versus direct?


Once you’ve established which questions to ask, you can identify the core KPIs that can be enabled by DataOps to answer those questions. For example:

  • How much of the available inventory has been monetized (ideally, broken down by specific verticals or ad units)?

  • Have impressions been delivered to campaigns as required, in the necessary numbers and from the appropriate inventory sources, such as Google Ad Manager (GAM), YouTube, or Programmatic?

  • What is your eCPM for individual groups of ad products or verticals?

  • Which programmatic sources are producing the most revenue and at the highest efficiency? What is the current and historical performance of different programmatic deal types and partners?

  • How much ad revenue are you generating per visitor? What is their LTV (Lifetime Value)?

  • What is the total addressable revenue opportunity across all inventory and demand sources?

  • Where are the biggest opportunities to increase revenue and profit?

  • How much inventory will you have to sell looking forward (ideally, broken down by vertical, ad unit, audience, viewability)?


While it‘s possible to generate these metrics with manual techniques, you risk burning valuable talent on repetitive and error-prone tasks: logging in to numerous tools and user interfaces, as well as downloading, scrubbing and wrestling unstructured CSV files into rudimentary spreadsheets.


Data automation solves these tedious processes, so questions that used to take weeks to answer can now be handled in minutes.


In our next post, we’ll tackle Step 2 of successful DataOps - normalizing raw data into foundational data. In the meantime, head to our blog page to check out our other posts in the DataOps series, or read the full guide to DataOps.