Better business decisions require better data. These guidelines provide the roadmap to get you there.

Starting off with business requirements

Implementing Adobe Analytics, Adobe Experience Platform, or any other digital marketing tool must begin with a deep dive into a company’s business requirements. Doing so ensures the implementation will directly address the specific questions stakeholders have regarding their digital marketing effectiveness. By defining these goals upfront—such as boosting online sales or visitor retention—you can immediately map them to measurable KPIs like purchases, return visitor rates, and campaign-specific use cases.

However, formalizing these KPIs requires significant additional effort, particularly in defining the specific data points needed for reporting. Relying solely on just the high-level, generic objectives often leaves developers and analysts without the detailed guidance necessary to determine exactly what to track. To ensure high-quality reporting, the implementation process must bridge the gap between a conceptual business requirement and the actual mechanics of collecting the right data to meet the requirement.

Getting into the detail with data requirements

And so, I propose that in addition to discussing business requirements, KPIs, and use cases, the company should also put as much effort into discussing data requirements at the start of an engagement. Business requirements must be accompanied by a clear definition of the data required to fulfill them.  When the company completely maps out the precise dimensions and metrics needed to support the desired outcomes, those who work on the technical side of the implementation can create a solid blueprint for a robust, scalable data collection solution.

Discussing data requirements shifts the conversation from the vague question of "what do we want?" to the actionable question of "what specific events on the site or app do we need to track?" That distinction is massive: it allows technical consultants to build a solution based on a concrete dataset rather than guessing at high-level goals. It also streamlines the implementation, saving everyone time by focusing only on the data points that drive the reports that actually matter.

The questions to ask

If you are unsure how to begin discussing data requirements, I recommend applying the '5W1H' framework—Who, What, When, Where, Why, and How—to every granular interaction, dimension, and metric within your data collection solution.

In my list of questions below, I start off with the "What" questions. Answering the "What" questions first when putting together a list of data requirements will help guide—or could even "automatically fill in"—the answers for the remainder of the questions.

What

• What website or app interactions do we need to track?

• What website or app interactions do we not need to track?

• What are the friendly names for the dimensions or metrics tied to each interaction?

• What overarching tracking category does each dimension or metric belong to (e.g., Site Content, Product Information, External Marketing, Search)?

• What is the specific purpose of each dimension or metric, and what is an example of a valid value at the time of collection?

• What is each dimension or metric not meant to track, and what is an example of an invalid value at the time of collection?

• What data type is associated with each dimension or metric (e.g., string, integer, double, boolean)?

• What is the data source that will populate the values for each dimension or metric (e.g., data layer, query string parameter, cookie)?

Where

• Where on the website or app will we track each interaction?

• Where on the website or app will we not track each interaction?

When

• When is the exact moment we will track each interaction?

  • On page load?

  • Upon data layer readiness/rendering?

  • On visitor interaction (e.g., click or tap)?

  • On a specific application state change?

Who

• Who are the stakeholders that require data from these dimensions or metrics?

• Who is responsible for deploying the data collection solution for each interaction?

Why

• Why is tracking each interaction and its associated metrics or dimensions important?

How

• How does each interaction, dimension, and metric directly map back to the defined business requirements, KPIs, or use cases?

• How will we store the dimension and metric data in Adobe Experience Platform or Adobe Analytics?

An example of what the answers provide

As an example, a boilerplate business requirement I frequently see is "Track the most popular links that visitors interact with on each page". That's a fine requirement by itself, but it’s also generic in nature and could easily be misinterpreted.  Truth be told, visitors do a lot more than just "interact with links" when they view a page.  A visitor could interact with images, videos, buttons, tabs, and other similar objects. Going through the questions above will help determine the level of data you want to collect from any (or all) of these objects.

Assuming a company wants to know the details behind each of these interactions, they could come up with the following set of data requirements:

• Friendly name: Page Interaction Detail

• Tracking Category: Site Content

• Description: The details behind all the interactions that could occur after a page finishes loading.

• Type: String

• Format: The value should include the following:

  • The type of interaction that took place (either user-initiated or otherwise)

  • The object that the user interacted with (when applicable)

  • Any other details that distinguish the interaction from all other interactions that could occur on the page.

  • Delimit each portion of the value with a space-pipe-space (e.g., " | ")

  • Use all lower-case characters

• Example values:

  • click | button | add to cart

  • click | link | http://www.abc123.com

  • click | tab | patio furniture

• Other notes:

  • Capture the page name—at the same time and in a separate dimension—so that we can know the pages on which each interaction took place

  • Also capture a "Page Interaction" metric to get the total number of page interactions across the whole site.

Conclusion

Going through this exercise will take time and careful consideration, but ultimately, hashing out the fine details of your data requirements and getting them documented will save everyone a lot of headaches down the road. And when you get the data right, your KPIs and reports will be genuinely reliable—which is the whole point! High-quality, accurate data is what actually drives better business value.